SCROLL NEWS / DISCOVERY
Search headlines
Find the stories shaping the conversation.
Results for " magnitude " 38 found 🔀 AI-powered Shuffle
How Much Of Cisco Stock's Rise Did The Business Earn?
Cisco Systems (CSCO) supplies businesses and the hyperscalers building AI infrastructure with networking and security products, along with services. The company's stock climbed 92% between late July 2021 and mid-October 2026, before dividends. A rally of that magnitude makes it natural to credit a growing business. So take a guess: how much of that rise came from Cisco selling more.
The Big Holes in Wearable Heart Rate Variability And Readiness Scores
On September 9th, Apple announced it was revamping its Apple Watch Health Sensing System, rolling out a Readiness score (0 to 10), and increasing the frequency of heart rate variability (HRV) outputs 24-fold. This can be viewed as upping its competition with various other consumer wearable sensors. These “readiness scores” as a composite of multiple metrics, with heart rate variability (HRV) being front and center for most. The majority of Americans are now using wearable sensors, which equates to well over 100 million adults. HRV and Readiness scores are increasingly being marketed as a measurement of autonomic nervous system health, a digital marker for future disease, a clock for biological age, and a holistic metric to promote healthspan and even longevity. (Apple also introduced a new longevity tab and “Health Age”.) None of this has been proven. In this edition of Ground Truths I am going to review what we know about heart rate variability and readiness scores. Heart Rate Variability HRV is the variation in normal heart cycle timing. The variability of the heart rate, the barely perceptible millisecond changes in time between consecutive heart beats (see R-R intervals in the Figure below, left panel), is due to interplay between the sympathetic and parasympathetic (vagal nerve) inputs. Distinct from heart rate, individuals with the same heart rate can have very different HRVs. It is a rough reflection of the autonomic nervous system (ANS) activity, inadequate to say whether a person’s ANS function is abnormal. For more than three decades, heart rate variability (HRV) has been measured and several studies have found an association of low HRV and clinical outcomes, particularly a link with higher all-cause and cardiovascular mortality. There have also been less well established links of low HRV to risk of early cognitive impairment, dementia, mental illness, Type 2 diabetes and substance abuse. An important reminder is that HRV is a surrogate marker without any established cause-and-effect relationship. If you increase your HRV, that doesn’t mean it will improve health outcomes. In fact, there is no hard evidence for that. All that work linking to health outcomes was done with electrocardiogram (ECG) derived HRV. Now, in the era of consumer wearables, this is getting assessed differently, by optical pulse (yes, the lights you see) plethysmography (PPG) or what is called pulse rate variability (PRV). They are not the same, as shown below (right panel) and only concordant when the delay between the ECG and pulse is kept constant, which basically means at rest. I should mention there’s also what I will call MPV, a mattress mechanical movement sensor, a derived heart rate variability, that companies like Eight Sleep use, even further away from directly measuring HRV. HRV has not one uniform measurement but many different types of quantification, such as RMSSD, the magnitude of difference between successive R-R intervals of normal sinus beats (N-N) or SDNN, the standard deviation of NN intervals, both in milliseconds. SDNN is one of the so-called frequency domain HRVs (others are LF, HF, LF/HF). Different wearable sensors use different metics; Apple has relied on SDNN and nearly all of the others use RMSSD, which is generally considered the more accurate metric. There’s also the different length of time measured, such as for a matter of minutes, all day, or an overnight’s sleep. Short measurements are especially problematic since they don’t capture enough of respiratory modulation and other factors that influence HRV. Share Ground Truths How well does HRV correlate with PRV? There are very limited studies, especially independently done. One that is commonly cited was conducted by Air Force researchers in only 13 healthy adults assessing Oura ring 3 and 4, Whoop 4.0, and Garmin Fenix 6 and showed a correlation coefficient of 0.88 to 0.97 and a mean absolute percentage error from 6 to 10%. The correlation is not a perfect 1.0, but there’s at least a fairly high level of correlation. HRV is supposed to increase during the night due to takeover of the parasympathetic nervous system, and higher during deep sleep. A recent example of my 1 week, all day “HRV,” and one during sleep is shown below. As you can see, the N of 1 data are inconsistent for the same days from different sensors (Oura, AppleWatch, Fitbit Air, and Eight Sleep) by patterns, absolute numbers, and comparison with prior days and weeks. The largest study in over 8 million Fitbit users (the old version, not Google Fitbit Air, introduced in May 2026) gives you a sense of the effect of age, sex, and the 2 different main HRV (here PRV) metrics, with RMSSD on the left and SDRR (=SDNN) on the right below. That study, from data collected in 2018, is a major outlier, since all the more recent ones are tiny with respect to sample size. Many of the companies have not had independent evaluation of their HRV, such as Eight Sleep, but have published a low standard error on their website. There are some other published studies on the correlation between HRV and PRV, but they are all small and only in healthy adults. A scoping review emphasized the lack of study in underrepresented individuals, including the aged, people of color (which affects the PPG signal), and individuals who are underweight or obese. Add the typical adult age 60 plus with one or more chronic diseases. For example, one study in over 900 adults found poor correlation of HRV and PRV, non-uniformly underestimated across many chronic diseases (cardiovascular, endocrine, neurological, respiratory, and others), concluding PRV is “an invalid surrogate for HRV.” A recent systematic review of 43 studies comparing HRV and PRV found reasonable pooled absolute standardized error (HRV as gold standard) but only 10 of the studies provided quantitative synthesis in ideal conditions. Their main conclusion was similarly cautious: “PPG-derived HRV [PRV] should not be regarded as universally interchangeable with ECG-derived HRV across all devices, populations, and recording contexts.” Factors Affecting HRV and PRV That gets me to the long list of factors that affect HRV (and PRV) besides the device, the type of measurement (RMSSD, SDNN or others), the person’s signal, the sensor site, the duration of data capture, if weighting by sleep stage is used, how artifact is processed and corrected. And this list is not complete!: Oura puts out data from their community of users (who input data) on what affects their overnight HRV. The factors currently provided are: no alcohol (increase 8%), melatonin (increase 2%, float tank (increase 2%), wine (decrease 4%), and party (decrease 14%) in overnight HRV. Must be some big parties! Subscribe What is a PRV measurement good for? It has been falsely characterized as an index of “autonomic balance” and a specific indicator of stress. A 2018 review of the studies available for HRV and its relationship to stress, not using any of the current wearables, found that stress can lower HRV. But so can many other factors. The non-specificity of the signal, indexed to the table above, is striking. Evidence from a UK Biobank study of over 46,000 participants with actual HRV looked at genetically predicted HRV, a genetic risk score, that failed to show the expected HRV-mortality link, indicating that _HRV is likely not causa_l, but rather a reflection of person’s physiologic state. A review of consumer wearable HRV data from 5 longitudinal studies showed that nighttime PRV was not associated with perceived stress, and surprisingly higher HRV, in the largest cohort (N=717 participants), was correlated with higher stress. An Oura ring cohort of 525 first-year college students found a link between overnight PRV and perceived stress, but that was also seen with resting heart rate, sleep, and respiratory rate. Several very small studies have examined the relationship of HRV and athletic injuries or guiding training with mixed, and predominantly negative results. HRV biofeedback training with paced breathing had no significant effect on reducing stress or raising HRV, as demonstrated with sham controlled trials. When HRV for multiple days showed a decline in conjunction with body temperature, the Oura ring published data for prediction of Covid. The WHOOP company sponsored an observational study, published in 2026, of 30,000 users for 72 weeks, without a control group, that reported reduced alcohol intake (5.8 % points) by self-report. That doesn’t tell us much, and particularly about the merits of HRV for behavioral change. If you use the same device and conditions as longitudinal trends for multiple (at least 2-3) weeks that may be the one way to get something useful from the measurements. Data for overnight sleep with minimal motion and using RMSSD is the best proxy for real HRV. The reason to look at trends rather than any given night is that it more likely represents something, even though you won’t know with certainty what the “it” is. Keep in mind there are no data, no peer-reviewed evidence, to show that HRV fluctuation in-person has any correlation with health outcomes. Share Readiness Scores These are proprietary scores that integrate different metrics for each of the wearables: no algorithms have been disclosed. They are unvalidated against health outcomes. In a review of 14 composite health scores of readiness and recovery, HRV contributed 86% to the scores, followed by reading heart rate (79%), physical activity and sleep duration (both at 71%). That review noted the substantial variability in measurement protools and lack of standardization. Sleep staging is notoriously inconsistent and inaccurate by these sensors, which adds further to the HRV uncertainties for what the scores, which use sleep stage data, mean. Only resting heart rate has been shown consistently across devices to be extremely accurate. I’ve made a Table to summarize what we know about which metrics are included, the scores, any peer-reviewed studies that compared the readiness score with health outcomes, and the corresponding (if any, NA-not available) citation. You will note that some companies do not use the term “readiness," such as WHOOP for recovery, and Garmin, which has 2 different scores, one of which is Body Battery. Eight Sleep uses the term “Fitness Score.” They all include HRV; Apple includes a new metric they call “Recovery HRV” which among other components uses 7-days of sleep, but it is unclear what this means or how it is differentiated from other scores (there are clearly no data for outcomes). We have no knowledge of how the different components are weighted or whether any of these scores are better than resting heart rate, HRV alone, physical activity, or any other single metric. Since none of these are standardized, they are not interchangeable, so if you get a 90 for Oura that has no relationship to a 90 on a Google Fitbit Air. Notably, the company can update its algorithm for readiness score at any point without notification to device users. Without any useful evidence of actionability for these scores or established relationship with health outcomes, it is hard to make a case for their value. At the Apple recent announcement they showed their Readiness score (0-10) on the watch (Figure below) but there are no published data on this score, not even on their website. It’s available only on their new Watch Series 12 or Ultra 4 [of course, ;-)]. That exemplifies the problems with these scores, lack of data and evidence for being meaningful to promote health. Perhaps the best study (which isn’t saying much) is the WHOOP Recovery for golfer performance, because it did correlate with an objective outcome, even though there was no control group and the authors were all from the company. Among the 389 pro golfers, an absolute 10-per cent point increase in Recovery score was associated with about 0.5 fewer strokes per round. But that’s hardly a health outcome! WHOOP is also conducting a study in over 2,700 runners to see if their recovery score will be linked to less injuries and improved performance, but that is not yet published and has no control group or randomization. Putting This in Context For two decades I’ve been enthusiastic about the potential for digital health and particularly wearable biosensors. Over the years, we’ve seen some great progress for their ability to promote physical activity and accurately detect atrial fibrillation (the first FDA cleared deep learning AI for consumers). That work was the subject of rigorous research. But there are holes in the data and evidence for other metrics. One notable one is the “VO2 max” story that I wrote about earlier this year. At that time many subscribers asked me to cover heart rate variability, which I finally got to here. When I dived into the research and publication for HRV and readiness scores, I expected to find at least some that were of high quality and demonstrated their utility by linkage to health outcomes. To my surprise, I found none. The wearable sensor measurements for HRV (PRV) are, for the most part, accurate, but that validation work has only been done in small studies of healthy adults and does not take into account the long list of factors, from the device, software side, and the user side, that affect HRV measurements. Moreover, this metric chiefly relies on optical sensing and, as we have learned for heart rate PPG sensing, may be less accurate in people of color. Keep in mind that all of the health outcome association evidence comes from ECG-derived HRV; none are from wearable sensor data. I will repeat the key point: there's no peer-reviewed evidence to show that in-person HRV fluctuation—or efforts to raise your HRV— has any correlation with health outcomes. For those of you who look at your HRV on awakening, or even 2+ week trends of it being low, I hope this context helps to relieve any anxiety. Yes, low HRV (not PRV) has been shown to increase risk of some diseases as summarized above. But efforts to raise your HRV—a surrogate metric— has not been established for improving any health outcomes. HRV does not have any evidence of causality (the genetic evidence actually goes against this possibility). In this summary, I have not included data for other wearable sensors such as Polar, Samsung, Withings, Suunto, Amazfit, Coros, Ultrahuman, or additional mattress sensors. These are beyond my first-hand experience, and as far as I know from my in-depth review none have any peer-reviewed published data that differ from the 6 sensors I’ve reviewed here. From the points I’ve gone over above, we’re not ready for readiness scores. Besides being proprietary, they are predominantly based on metrics that have their own issues. It’s compounding the problem, like building a house without a solid foundation that has never undergone a rigorous inspection, and then selling it. Like I mentioned for PRV, you can look at trends over weeks rather than any single day, to get a handle, but even that may not be helpful. There’s simply no evidence that these scores meaningfully relate to health outcomes. I’d emphasize they might, but that requires doing prospective or randomized studies to prove it. None exist. There’s great promise for HRV/PRV utility_. For example**,**_ Prof Maiken Nedergaard, who discovered the brain glymphatics that are essential in eliminating metabolic waste products from the brain during sleep, has posited that HRV could be a non-invasive marker for neuromodulator oscillations, brain-body regulatory circuits, and brain clearance. That would be extremely useful, but like everything else on HRV and readiness scores it requires solid research and validation. The lay media isn’t helping much to get the story straight. Earlier this year The Economist published a piece entitled “The most useful indicator of your overall health” which ordained HRV as an “accumulated stress score.” That’s akin to the false assertion about VO2max: “V02 max is the singular most powerful marker for longevity.” As I’ve summarized here, that is not established. The fact is that so many things can lower HRV, including physical exercise (especially an intense workout), reduced sleep quality, stress, the list above, no less the device, signal, and software. Whatever fluctuations observed have not been correlated with any health outcome. Sadly, “datamaxxers” are widely using HRV and readiness scores that have never been validated to mean anything. We already know that for some people using the sensors for sleep metrics, it can induce “orthosomnia,” an obsession to get high sleep quality, with associated high levels of anxiety. In an experiment done by a company to promote sleep quality for its employees, “For those employees who did use the trackers, many reported feeling perfectly rested until their tracker told them they had had a terrible night. Others were told that they had slept like a baby when they had actually been lying awake worrying about the quality of their sleep. “ The same problem can result from preoccupation with HRV or readiness scores, with anxiety that would lead to further reduction in both. It you are using a wearable like >100 million American adults, it’s OK to look at these data, but contextualized with the major caveats reviewed here. If you are one to require evidence that HRV or readiness scores are linked to health outcomes, you may not even want to look. The companies make it hard to turn them off! Let me end with the companies that make and sell wearables. Apple’s doubling down on HRV (24-fold more reporting and heart rate very 5 seconds) and introduction of a Readiness score tells us that consumers have bought into these metrics and they are joining the club. However, all of this is occurring with a backdrop of tens millions of users, claims about the data that are not backed up by adequate evidence, marketing way out in front of whatever limited data exists, and not being transparent about their readiness score algorithms. The companies can well afford to do the research that is needed to connect these metrics with health outcomes show, once and for all, that increasing HRV or using readiness scores promotes our health. If they believed and invested in the products they are selling, we’d not be in this position of not knowing. That’s essentially where we are with HRV and readiness scores. Perhaps someday this will change and we’ll have good reason to embrace them. NB: I wrote this post. No AI. I have no conflicts of interest with any of its content. Loading... Ground Truths has 215,000 subscribers from every US state and 214 countries. There are over 300,000 followers of Ground Truths so more than 90,000 folks who can easily convert to be free subscribers. Your subscription to these free essays and podcasts makes my work in putting them together worthwhile. If you’re not a subscriber, please join! If you found this interesting PLEASE share it! Share Ground Truths The proceeds from all voluntary paid subscriptions go to support our summer internship program. It enabled us to accept and support a record number of 62 summer interns that joined us in 2026! These are high school, college and medical students selected from thousands of applicants. We couldn’t do this expanded program without the funds coming in through Ground Truths. Thank you!
What Does the Science of Climate Change Really Look Like?
Editor’s Note: This is the second in a three-part series on how the Right should think about environmental and climate policy. Read Chris Barnard on Reclaiming Environmental Policy from the Left. More than a decade ago, the world’s governments negotiated and signed the Paris Climate Agreement, committing to hold global warming well below 2 degrees Celsius. In the following years, climate had immense cultural power. Greta Thunberg emerged. Fortune 500 companies established net-zero plans. The U.S. got the Inflation Reduction Act. Then the pendulum swung back. President Donald Trump returned to the Oval Office, speaking of green energy as a “scam” and carbon footprints as a “hoax” and on all matters of international policy emphasizing bilateral engagement and power over globalism and cooperation. The technology companies that had been at the forefront of net-zero pledges and green leadership discovered the importance of all-of-the-above power to their visions for artificial intelligence and began quietly postponing or removing their targets. On the left side of the political spectrum, rising concerns about inflation and cost pushed climate lower down in the stated priorities of voters and stump speeches of politicians. The moment is ripe for a reset on the politics of climate change and for conservatives especially to chart a course that acknowledges and addresses the real challenges in plausible ways. So what should we think about climate change as we enter a post-peak climate era? If it is neither an apocalypse nor a hoax, what will hold up as a durable view of climate change? And how does it project onto the concerns and priorities of the right-of-center’s emerging coalition? New to Commonplace_? Subscribe below to get the magazine in your inbox._ Subscribe A Solid Foundation As the politics and economics of climate careened over the past ten years, the physics carried on regardless. In 2015, the year the Paris Climate Agreement was adopted, the world emitted 41.2 gigatonnes of carbon dioxide. By 2025, emissions had risen only 2.4%, to 42.2 gigatonnes, because falling emissions from land use offset increasing emissions from fossil fuels. But slow growth in annual emissions is not rapid decline, let alone net-zero, so carbon dioxide kept accumulating in the atmosphere, going from 399 to 426 parts per million. Global temperature kept rising too, from about 1.1 degrees Celsius above the 1850–1900 average in 2015, to 1.4 degrees above it in 2025. None of this should surprise anyone. Even most prominent skeptics, like the authors of the 2025 report commissioned by the Trump administration’s Department of Energy, accept the basic physics. Carbon dioxide added to the atmosphere at industrial volumes accumulates and traps heat that would otherwise escape to space. That heat warms the upper ocean and increases temperatures at the surface. Increased ocean heat and melting of land ice cause sea levels to rise. Changes in the climate become noticeable to us as subtle shifts in temperature and precipitation, seasons arriving early or late, changes in the surrounding ecosystem, and weather extremes. The much more meaningful debate is not about whether these things are happening, but about what it all means for human societies. (While it is a proxy measure of overall risk, no one experiences global surface temperature.) This debate invokes at least three interesting questions: how much today’s warming is showing up in regional climate trends, how much more warming we should anticipate, and how extreme weather events can be understood in the context of climate change. Recent scientific advances give us new insights into each of these, but there is still much to learn. This is the context in which the next generation of conservative leaders will be the first to deal with significant climate change as a fact, not a forecast. It will be with them for their entire careers. Our evolving scientific picture of climate change will come from weather and climate records stretching further into the past, new kinds of observations in the present, better modeling tools for the future, and simply more time to observe the emergence of climate signals and their tangible effects on ecosystems and communities in the United States and around the world. The policy landscape will be shaped by not only the impacts of changes in the climate, but also by the ways in which societies respond and by the emergence of new technologies for both altering the trajectory of emissions and adapting to new climate realities. Global climate records now show the fingerprints of warming in different phenomena around much of the world. The IPCC, in its most recent assessment report, documents warming trends over all land regions, even as natural variability adds substantial variance to local trends. And the highest temperature extremes have become more intense and frequent, almost everywhere, since the 1950s. Scientists are highly confident that these changes are attributable to human influence. But confidence in detecting trends and attributing them to human influence degrades across heavy precipitation events; drought trends are heterogeneous and not attributable to human activity with great confidence. For hurricanes, tornadoes, and other severe storms there is even less confidence in trends or their relationship to climate change. In general, a fair summary of the evidence is that for well-observed phenomena, like surface temperature or heavy rainfall, with a clear relationship to warming, our multidecadal observations are consistent with human influence overcoming natural variability over long time periods and extremes increasing. For regions that are sampled more sparsely, or phenomena with a higher noise-to-signal ratio, we will need to observe them longer to understand the magnitude of the climate signal or improve dynamical understanding using models and observations together. But for Climate For those who can’t wait, we also have new techniques that attempt to understand the role climate change has in influencing particular extreme events. When the climate is changing and disaster strikes, it is natural to ask whether that disaster was somehow “caused” by climate change. For scientists and policymakers acting in good faith, answering that question helps develop a more accurate picture of the problem, how it may be getting worse, and what preparations we should be making to respond to it. It may, at some point, give some weight to how liability is assigned by courts or adaptation funding is distributed by society. But how we ask this question is extremely important. It is easy to get the analysis wrong by discounting the role that meteorology plays. As meteorologist Theodore Shepherd explained in a clear 2016 review of climate attribution methods, “if a weather or climate event is truly extreme in the present climate, then perforce it requires unusual meteorological conditions, which means that climate change is at most a contributing factor.” Share Scientists have two ways to probe climate as a contributing factor, both of which have entered media coverage of extreme weather events. One asks a probabilistic question: How much more likely is a particular event (e.g., a temperature record over a particular area) amid global warming? The answer involves using historical records and computer simulations to estimate the likelihood of such an event in both a changed and a preindustrial climate. The findings are less about the specific event, and more about events of that nature. The other question is: How has the changed climate affected the specific event in question? Here, scientists try to understand what a similar event would have looked like in the preindustrial climate. This is sometimes called the storyline approach, where the story is the specific meteorological details of the event. These methods offer ways to test the intuition of scientists about real weather extremes. In late June and early July of 2021, a persistent high-pressure ridge, or heat dome, set up over the Pacific Northwest and an extraordinary heat wave affected the area from Oregon to British Columbia. For six days, it shattered temperature records across the region and hundreds died from heat-related causes, in an area where such high, and persistently high, temperatures were well outside of experience and many live without air conditioning. Scientists have studied its connection to climate change using multiple approaches. This specific event was created by a rare combination of meteorological factors, a strong high-pressure ridge created the conditions for extreme heat, which occurred on top of higher average temperatures in the region from global warming. Probabilistic analyses showed that climate change increased the likelihood of such an event by at least 8-fold to more than 100-fold. One standout example found that such an event had effectively zero probability of occurring in the preindustrial climate. The enormous range in these assessments reflects how hard it is to estimate probabilities of events at the tail of the historical record. Estimates of the effect on the temperature of the event are more clustered. A recent review paper documents how multiple methodologies have found a positive influence, roughly 1-2 degrees Celsius of an anomaly that exceeded 15 degrees, of climate change on the temperature magnitude of the event. Multiple studies have investigated and found some positive influence of climate change in other extreme events. Attribution studies found climate likely increased the heavy rainfall that accompanied Hurricane Helene in North Carolina and surrounding regions in 2024. The extensive fires that struck Los Angeles in 2025 illustrate how the causal chain can become messy, though. Studies do detect a positive influence of climate on the event’s likelihood, but while climate change likely contributed to underlying aridity, it would not have played a role in the heavy winds or land practices that preceded disaster. These event attribution studies will become more common for extreme, or damaging, weather events. They can be produced quickly, and often are reported before peer review. As the Pacific Northwest heatwave example shows, when multiple methods converge on a positive attribution, the finding should probably carry some weight even if you have to be careful about accepting the results from a single study. I expect that as climate change proceeds and the climate thus departs further from a preindustrial counterfactual, the influence will become more detectable across a variety of extreme events and more easily identified. This will be used to cast blame, but can also be used to inform how communities and society adapt to ongoing change. We May Still Be Surprised One positive development, insofar as less climate change is better, has been that our central estimates for future climate change should probably be revised downward. Mostly, that is because the high-end warmings that scientists regularly analyzed about ten to 20 years ago, driven by high emissions throughout the twenty-first century, now appear to be somewhere between unlikely and impossible. At one time, the upper end of mainstream climate projections extended well into 4 to 6 degrees Celsius of warming by the end of this century, which, models suggest, would have wrought enormous real-world damage. Current energy and policy trends now point toward 2.5 or 3 degrees. Damages and risks are commonly modeled as increasing steeply with more warming, so this is already a better-than-previously-expected outcome for the climate (which is independent of the physical response to emissions). But we should maintain a wide range for plausible outcomes and prepare for the possibility of being surprised. Emissions trajectories are subject to deep uncertainty, the physical climate response is still developing, and whatever change occurs in the climate will then be mediated through unpredictable economic, social, and political institutions. Over the past decade, the pace of warming surprised some keen observers and appeared to accelerate, though not yet outside the range of expectation provided by climate models. The reasons for that apparent acceleration are being actively studied (as was the apparent pause in warming from 1998 to 2012), but no single driver has emerged. Some blame it on the El Niño, variability which would have no bearing on climate. Some think it is a result of factories in East Asia and global shipping fleets cutting aerosol emissions, which in the strongest version of the argument would indicate higher climate sensitivity to carbon-dioxide emissions. Others are more measured, not yet ready to draw solid conclusions as to whether climate projections require revision. We will have to see. Long-Term Thinking Climate change asks us to think over long time scales. But the pace of human-driven warming is compressing a large global change into a century. Changing weather patterns and extreme events are already causing adjustment costs and damages and will do more. Adaptation will help, but it is not free. For some communities, these costs will erode livelihoods and well-being, and may force migration. Are we prepared for large-scale managed retreats in the United States? Thankfully, American wealth, geographical diversity, and moderate climates may leave us better prepared than many around the world. But wealthy countries can still suffer serious disruption from narrowly concentrated costs, even if they appear entirely manageable in aggregate. Readers of Commonplace are familiar with the challenges of rapidly adapting to economic forces. The displacement that came from the China Shock is barely perceptible in aggregate GDP and employment data, but it affected millions of people and those people were far less mobile than economic models tended to assume. Deindustrialization rippled through the nation with enormous effects for not only our economic vitality, but also our national security. The forces surrounding climate change may not be so different; the world gets richer and must accept some diffuse costs. Those may be modest overall, but cause all manner of unpredictable effects with which policymakers must cope. In the case of the China Shock, our faith in a growing pie served us poorly. We’ll need to do better on climate. Leave a comment
A rare “boomerang” earthquake shocks scientists with its bizarre U-turn in Burma - Futura-Sciences
A March Morning That Changed Everything On March 28, 2025, a violent earthquake struck Myanmar (formerly Burma) with an estimated magnitude between 7.7 and 7.9. The epicenter was close to Mandalay, the country’s second-largest city. The quake’s power rippled far—so much so that tremors were even felt in neighboring Thailand,...
UC Riverside scientists develop tool to predict future earthquake sites
Researchers used GPS data to measure where energy was building up in faults, correctly predicting where an 8.8 magnitude earthquake later occurred in Russia last year.
UC Riverside scientists develop tool to predict future earthquake sites
Researchers used GPS data to measure where energy was building up in faults, correctly predicting where an 8.8 magnitude earthquake later occurred in Russia last year.
UC Riverside scientists develop tool to predict future earthquake sites
Researchers used GPS data to measure where energy was building up in faults, correctly predicting where an 8.8 magnitude earthquake later occurred in Russia last year.
How geopolitical rifts cut scientific collaboration
Researchers have measured how worsening geopolitical relations hobble scientific collaboration, at last quantifying the impact of rifts between governments on cooperation between labs. Over the last decade, a tenser world has thrown up new barriers to cross-border science. Collaboration between the US and China has dimmed as the two countries have sunk into an ever-deeper rivalry. Links between Russia and Europe have collapsed following Moscow’s full-scale invasion of Ukraine. But while a concern, the magnitude of the damage is less than might be expected. For all the geopolitical turmoil of the last decade, international scientific collaborations are near a record high. “Some part of the story is positive,” said Shreyas Gadgin Matha, a social scientist at Harvard University and an author of the new study. “Research collaborations are largely resilient.” Matha and colleagues looked at data on cross-border research papers and patents in the period 2000-23, and analysed how they were impacted by countries’ political relations, as measured by how often states voted together in the United Nations. They found that when states drift apart at the UN by the median difference in votes between all countries, joint scientific papers decline by 3.1% over the next three-year period. To put that in context, over the last decade, the US and China have diverged by around a third of that difference in votes every three years. In other words, on average, it takes quite a significant diplomatic rift even to trigger a small single-digit percentage drop in joint papers. “In the short term, even something like a 3% reduction [in joint papers] is, in some ways, not that big of a number,” said Matha. Research remains more international than ever, according to an analysis extending to 2023 released by the Organisation for Economic Cooperation and Development (OECD) in 2025. We’re still living in an era of unprecedented scientific globalisation, the analysis says, with 27% of papers in OECD countries written across borders. In 1970, just 2% of articles had authors from different countries. Longer-term damage?However, this trend towards increasing international collaboration “has lost momentum and might be partly breaking down,” the OECD analysis says, largely due to a fall in the intensity of US-China joint papers since 2018. Matha and his colleagues stressed they are not complacent about the global science system. For now, “it's pretty resilient,” said co-author Muhammed Yildirim, director of academic research at Harvard’s Growth Lab. But if links are weakened, through the denial of study visas, for example, then this could have a much longer-term effect, cascading through generations of scientists, he added. Related articlesChina becomes Russia’s biggest collaborator after war decimates science ties with the west US scientists prefer German to Chinese collaborators, study finds Horizon Europe is now truly global. But what comes next? Another finding of the study, Burning bridges or bridging divides: Geopolitics and collaboration in research and innovation, is that trust appears easier to destroy than create.The hit to scientific collaboration when countries drift apart is around three times as big as the growth in joint papers when they move closer, the paper finds. “It's far easier to break these ties than to mend them, and that that also perhaps has some larger policy lesson,” said Matha.PatentsThe results around patents were less clear cut. Some methods of analysis found a statistically significant link between a diplomatic rift and a drop in co-patenting, but others did not. According to Yildirim, patents with cross-border inventors are often developed by multinational companies, which may have invested heavily in global R&D facilities, so are less likely to change how they operate simply due to a change in the geopolitical weather. “That's why we might not be seeing the effect immediately in technology,” he said.By contrast, researchers might be more sensitive to geopolitics, because visa policies can quickly affect who meets who at academic conferences, he went on. International collaboration on patents is much lower than for academic articles, the study shows. It peaked at around 10% in 2010, and has since fallen back to around 8%. This could be because China dramatically ramped up domestic patents in the 2010s, said Yildirim, squeezing the international share.
How geopolitical rifts cut scientific collaboration
Researchers have measured how worsening geopolitical relations hobble scientific collaboration, at last quantifying the impact of rifts between governments on cooperation between labs. Over the last decade, a tenser world has thrown up new barriers to cross-border science. Collaboration between the US and China has dimmed as the two countries have sunk into an ever-deeper rivalry. Links between Russia and Europe have collapsed following Moscow’s full-scale invasion of Ukraine. But while a concern, the magnitude of the damage is less than might be expected. For all the geopolitical turmoil of the last decade, international scientific collaborations are near a record high. “Some part of the story is positive,” said Shreyas Gadgin Matha, a social scientist at Harvard University and an author of the new study. “Research collaborations are largely resilient.” Matha and colleagues looked at data on cross-border research papers and patents in the period 2000-23, and analysed how they were impacted by countries’ political relations, as measured by how often states voted together in the United Nations. They found that when states drift apart at the UN by the median difference in votes between all countries, joint scientific papers decline by 3.1% over the next three-year period. To put that in context, over the last decade, the US and China have diverged by around a third of that difference in votes every three years. In other words, on average, it takes quite a significant diplomatic rift even to trigger a small single-digit percentage drop in joint papers. “In the short term, even something like a 3% reduction [in joint papers] is, in some ways, not that big of a number,” said Matha. Research remains more international than ever, according to an analysis extending to 2023 released by the Organisation for Economic Cooperation and Development (OECD) in 2025. We’re still living in an era of unprecedented scientific globalisation, the analysis says, with 27% of papers in OECD countries written across borders. In 1970, just 2% of articles had authors from different countries. Longer-term damage?However, this trend towards increasing international collaboration “has lost momentum and might be partly breaking down,” the OECD analysis says, largely due to a fall in the intensity of US-China joint papers since 2018. Matha and his colleagues stressed they are not complacent about the global science system. For now, “it's pretty resilient,” said co-author Muhammed Yildirim, director of academic research at Harvard’s Growth Lab. But if links are weakened, through the denial of study visas, for example, then this could have a much longer-term effect, cascading through generations of scientists, he added. Related articlesChina becomes Russia’s biggest collaborator after war decimates science ties with the west US scientists prefer German to Chinese collaborators, study finds Horizon Europe is now truly global. But what comes next? Another finding of the study, Burning bridges or bridging divides: Geopolitics and collaboration in research and innovation, is that trust appears easier to destroy than create.The hit to scientific collaboration when countries drift apart is around three times as big as the growth in joint papers when they move closer, the paper finds. “It's far easier to break these ties than to mend them, and that that also perhaps has some larger policy lesson,” said Matha.PatentsThe results around patents were less clear cut. Some methods of analysis found a statistically significant link between a diplomatic rift and a drop in co-patenting, but others did not. According to Yildirim, patents with cross-border inventors are often developed by multinational companies, which may have invested heavily in global R&D facilities, so are less likely to change how they operate simply due to a change in the geopolitical weather. “That's why we might not be seeing the effect immediately in technology,” he said.By contrast, researchers might be more sensitive to geopolitics, because visa policies can quickly affect who meets who at academic conferences, he went on. International collaboration on patents is much lower than for academic articles, the study shows. It peaked at around 10% in 2010, and has since fallen back to around 8%. This could be because China dramatically ramped up domestic patents in the 2010s, said Yildirim, squeezing the international share.
Extreme inequality is not a given. It is a political choice.
An International Panel on Inequality would produce rigorous assessments of the magnitudes and changes in inequality, writes Joseph Stiglitz.
Scientists find a way to slash computer memory energy use by orders of magnitude
Scientists have devised a new way to switch magnetic computer memory while using far less energy than today's leading technologies. By mathematically optimizing the pulses used to flip digital bits, the method could reduce energy consumption by several orders of magnitude. Simulations suggest it may bring future memory devices surprisingly close to the fundamental physical limit for processing information. The same idea could eventually work with electrical currents or ultrafast lasers.
Scientists find a way to slash computer memory energy use by orders of magnitude
Scientists have devised a new way to switch magnetic computer memory while using far less energy than today's leading technologies. By mathematically optimizing the pulses used to flip digital bits, the method could reduce energy consumption by several orders of magnitude. Simulations suggest it may bring future memory devices surprisingly close to the fundamental physical limit for processing information. The same idea could eventually work with electrical currents or ultrafast lasers.
Scientists Expose Unexpected Driver of the May 2024 Superstorm
Even though the solar wind was unusually dense during the May 2024 superstorm, ions from the solar wind contributed very little to the ring current, marking the first time both were observed simultaneously during a storm of this magnitude.
Lane Kiffin Feeling ‘Magnitude’ of LSU Job Entering Crucial Season
As the new college football season gets going soon, all eyes will be on the LSU Tigers and head coach Lane Kiffin.After a controversial exit from Ole Miss last year, Kiffin’s new role with the Tigers has been discussed more than ever. There is more pressure on Kiffin entering this season than he’s ever faced […] The post Lane Kiffin Feeling ‘Magnitude’ of LSU Job Entering Crucial Season appeared first on HEAVY.
Colombia Earthquake: Direct Relief Mobilizing Medical Aid for Immediate Health Needs
A 7.4-magnitude earthquake reverberated through Colombia today, the largest earthquake to impact the country in a decade, according to the Associated Press. The earthquake shook the country at 7:34…
Experts share the science behind massive earthquake in Texas Panhandle
We don't often think of earthquakes here in Texas, but the U.S. Geological Survey reports a 5.0 magnitude quake hit last week in the Texas Panhandle.
Hospital, local officials support WVU Health System takeover of Independence Health
Danielle Yackovich had a simple message for state regulators who will have the final say over the proposed takeover of Independence Health System by West Virginia University Health System. “When we talk about an investment of this magnitude, we need to ask the simple question: What will it mean for
Magnitude Returns To Work Tab Ahead Of Whitney Bid
Winchell Thoroughbreds' Magnitude returned to the work tab on Sunday morning for trainer Steven Asmussen for his first workout since winning the June 27 Stephen Foster (G1), breezing a half-mile in :50.40 on the Oklahoma training track at Saratoga, Daily Racing Form reports. The Gun Runner colt ...
Business Monday: Big Island companies continue to recover from Kona low storms and earthquake
With the completion of property damage assessments from South Kona’s magnitude 6 earthquake that struck on May 22, Hawai‘i County officials now are waiting for Gov. Josh Green’s decision about whether he will file for a presidential disaster declaration that will help individuals and business owners secure funding for repairs.
iPhone 18 Pro and Pro Max May See $200 Price Increase
Apple's iPhone 18 Pro models could be up to $200 more expensive, according to a prediction from analytics firm IDC. IDC expected Apple to raise iPhone 18 prices, but prior to yesterday's Mac and iPad price hike, the prediction was a $100 increase for the 18 Pro and Pro Max, and a $50 increase for the base models. IDC Senior Director of Data & Analytics Nabila Popal says the magnitude of the Mac and iPad price increase points to even higher iPhone 18 prices.
PAHO mobilizes emergency health response following earthquakes in Venezuela
Caracas, Venezuela/Washington, D.C., 25 June 2026 (PAHO) – The Pan American Health Organization (PAHO) has activated its emergency response mechanisms to support the health response in Venezuela following two magnitude 7.5 earthquakes that struck the country on June 24.
A NASA satellite caught a giant tsunami doing something no one expected
A Pacific-wide tsunami triggered by a magnitude 8.8 Kamchatka earthquake gave scientists their first detailed satellite view of a major tsunami in motion. The observations revealed unexpected wave behavior and helped uncover a larger earthquake rupture than earlier models predicted.
Is SMIC N+3’s Metal Pitch Smaller than Intel 18A’s?
Almost four years ago, we published that SMIC had started shipping 7 nm (N+1) chips. Now, SMIC is shipping its third-generation 7 nm (N+3) process in Huawei’s Kirin 9030, with a minimum metal pitch of 32.5 nm, about 10% tighter than the 36 nm minimum metal pitch shipping in Intel’s latest Panther Lake CPUs on 18A. The headline is true, but incomplete cherry picked metric. N+3 reaches the density of TSMC N6 through aggressive DUV multi-patterning and design-technology co-optimization (DTCO), but it pays for that in complexity, efficiency and process control. We found this and more in our reverse engineering and teardown where we cover SMIC’s N3 process technology, Huawei’s packaging, memory, architecture, and more. SemiAnalysis has been building a state-of-the-art teardown lab in Oregon capable of analyzing the world’s most advanced and important chips over the last year and half. We have already generated revenue on advanced datacenter chip teardowns including our recent reverse engineering of a major TSMC customer’s COUPE CPO optical engine + EIC 3D stack. This is the first public report from the SemiAnalysis Teardown Engineering & Evaluation Lab, or STEEL for short. The lab is aggressively scaling up and out and we’re excited to announce it publicly. This is a bit of inconvenient timing for TechInsights as they are private equity owned and currently being sold while having enjoyed virtually no credible competition for decades. This has led to TechInsights underinvesting in CAPEX. SemiAnalysis exceeds TechInsights in revenue despite no venture or private equity ownership and being founded only 6 years ago. Because we have no external investors and are founder led, we move faster, build faster, and we can release client chip teardowns for free regularly, while focusing on datacenter for our major clients. Here’s the first public image from our lab, the HiSilicon Kirin 9030 Pro SoC: This report will detail our teardown of the Kirin 9030 and our findings on SMIC’s N+3 process, the most advanced in China. For comparison, we’ll show our teardown of the MediaTek Helio G99, made on TSMC N6. Through this comparison, we can look at the effect of export controls – SMIC N+3 and TSMC N6 are comparable nodes, but one is heavily export-controlled, the other free to use the West’s most advanced equipment. Here we see both China’s progress and constraints. SMIC N+3 reaches TSMC N6-class logic density, but it requires far more aggressive DUV multi-patterning, so it does not match N6 on process maturity or cost. The Kirin 9030 Pro performs similarly to three-year-old Android flagships, and trails far behind the current flagship SoCs from Apple, Qualcomm, MediaTek, and Samsung. The efficiency gap is even wider. Export controls have not stopped Huawei and SMIC from shipping advanced silicon, but they have forced a different path. Without EUV, SMIC is leaning harder on DUV multi-patterning, DTCO, and increasingly complex integration. The roadmap continues forward through tighter design rules and backside power, but each step adds cost and process risk. Huawei’s τ scaling and LogicFolding show another path: stacking active logic and recovering density through advanced packaging and system-technology co-optimization (STCO). To understand the Kirin 9030, we must first understand Huawei’s SoC history. HiSilicon is Huawei’s chip design arm, responsible for the Kirin smartphone SoCs, Kunpeng server CPUs, Ascend AI accelerators, and switch/router networking silicon. Before export controls, Huawei was TSMC’s largest customer – the only customer on TSMC’s first EUV node, N7+, and among the first on N5, alongside Apple. That ended in late 2020. Huawei switched to Qualcomm SoCs in its flagship smartphones, though export controls limited them to 4G-only variants. In late 2023, Huawei returned to in-house silicon with the Kirin 9000s, a successor to the Kirin 9000, fabricated on SMIC N+2 instead of TSMC N5. In the following years, they released the Kirin 9010 and 9020 on the same N+2 process. These chips used Huawei’s in-house TaiShan CPU cores and Maleoon GPU. We have not torn down a Kirin 9020 ourselves, so the predecessor die shot is from Kurnal. The die shots show how Huawei spent its silicon budget: which functional blocks are where, and how their areas compare to the predecessor. First, a quick guide to the major blocks on the die. The total die area is nearly identical, but the 9030 uses that area more aggressively. A denser process lets Huawei fit an extra middle CPU core, more GPU and NPU cores, and larger caches into the same footprint. In contrast, the Helio G99 is a much smaller, low-cost SoC, built for budget smartphones rather than a flagship device. While the Kirin 9030 is ~140 mm², the G99 is only ~29 mm², roughly one-fifth the area. The underlying TSMC process technology, however, is directly comparable as a baseline for analyzing SMIC’s. The Kirin 9030 is an evolutionary refresh, not a clean-sheet design. Its CPU, GPU and NPU cores carry over the 9020’s families, and the gains come from three levers: the SMIC N+2-to-N+3 process step, DTCO and floorplan work, and incremental microarchitecture. Area is where the first two show up, and the 9030 scales well here. Performance and efficiency are the harder test. Huawei’s design holds up better than its node would suggest, but the chip still trails, both because N+3 sits behind the leading-edge nodes and because its cores, while competent, remain a few generations behind the newest designs. The new prime core is an incremental update. The main changes are a 10% frequency increase from 2.5 GHz to 2.75 GHz and a doubling in the L2 cache from 1 MiB to 2 MiB. Despite the increased cache, the core size decreased by 7.6%. Excluding the private L2 cache, the core size decreased by 21%. This is a large reduction for an incremental node. Compared with the TaiShan New V120 core in the Kirin 9020, the Kirin 9030’s middle core is almost unchanged architecturally, yet each core shrinks by ~22%. Most of that comes from the move from N+2 to N+3, with layout likely accounting for the rest. Visually, the most noticeable change is the increase from 3 middle cores to 4. There is also a 20% increase in the shared L3 cache of the big cluster. This helps to improve multi-core performance without sacrificing much in terms of area. Even with each core shrinking, the big CPU cluster’s total area is essentially unchanged. The per-core savings went back into an additional middle core and larger caches. The tiny cores shrank less than the prime core (excluding its L2 cache) and less than the middle cores. This is likely because fixed overhead is a larger share on a small core. We cannot resolve any architectural changes from the die shot alone, but the per-clock and efficiency gains shown below point to more than pure process and layout scaling. The area reduction was offset by a doubling of the shared L2 cache from 2 MiB to 4 MiB, leaving the total tiny CPU cluster area slightly larger. Area is the easiest improvement to see from a die shot, but it is only one part of PPA (power, performance, area). For modern logic, power and performance matter just as much, and often more. Since Dennard scaling broke down in the mid-2000s, voltage and frequency have not scaled in step with transistor dimensions, so each node has had to fight harder for gains in performance and efficiency. The starkest comparison is not Kirin 9020 versus Kirin 9030 Pro. Apple’s efficiency cores run circles around Huawei’s prime core. Apple’s low-power core delivers 20% higher integer performance while drawing only 1 W, compared with 4.5 W for Huawei’s prime core. N+3 matches TSMC N6, but N6 is several generations old. Apple and Qualcomm build on N4 and N3P, which are denser and sit on a better voltage-frequency curve, giving them a larger transistor budget and more performance per watt. The 9030’s own cores did improve. The middle and tiny cores gained 17% and 14% in per-clock integer performance over the 9020, with floating-point flat on the middle core and up 11% on the tiny. The tiny core improves cleanly, with performance rising while power falls and efficiency increases by 45% in integer and 24% in floating point. The middle core is mixed: integer performance rises but power rises faster, cutting integer efficiency by 7%, while lower power lifts floating-point efficiency 16%. Per-clock gains at the same or lower frequency are microarchitectural, so the cores are tuned, not just shrunk. Both also failed to hold their rated maximum frequencies, pointing to thermal, power, or stability limits. Per clock, the middle core sits around Arm Cortex-A720 and the tiny core near the Cortex-A520; absolute performance trails because Huawei clocks them much lower. The prime core is roughly Cortex-X2 class per clock, a 2021 design. Apple’s 2020 M1 Firestorm core is still 35% higher per clock and 57% faster in absolute integer performance at a similar 4.5 W. The current leading edge is further ahead again: the Apple M5 P-core is 60% higher per clock and 2.7× faster, the Arm C1 Ultra 45% higher and 2× faster. Matching older high-end cores per clock is a genuine design achievement. What Huawei cannot match is the voltage-frequency curve and transistor budget of leading-edge nodes, which let Apple, Qualcomm and others spend more transistors in the same area on wider cores, larger caches and deeper buffers while running at lower voltage. Huawei’s LogicFolding roadmap is one answer, stacking active logic to recover density and shorten signal paths. We return to it later. The GPU compute units (CUs) changed more visibly than the CPU cores, moving to a more rectangular layout for both the arithmetic logic unit (ALU) clusters and the CU overall. Even with ray-tracing support added, a CU shrank ~28%. However, that shrink is offset by the increase from 4 to 6 CUs and the area outside the CUs grew 33%. Overall, the GPU cluster is larger by ~10%. The GPU is where Huawei makes its biggest gains. The Maleoon 935 is not competitive with current flagships, but it is a large step up from the 920 and reaches older-flagship territory. In 3DMark it is 70% faster in Wild Life Extreme (WLE) and 79% faster in Steel Nomad Light (SNL) than the 920; with 11% higher clocks and 50% more CUs, the ~67% theoretical uplift roughly matches WLE and is beaten by SNL. It edges ahead of the Snapdragon 8+ Gen 1 in WLE and SNL, and the Dimensity 9200 and Apple A16 in WLE, but stays far behind newer parts: the Snapdragon 8 Elite Gen 5 and Dimensity 9500 are ~2.4–2.6× faster in WLE and ~3.2× faster in SNL. The Maleoon 935 is Huawei’s first GPU with hardware-accelerated ray tracing; there it lands slightly ahead of the Exynos 2200, and on par with the Apple A16, with current flagships up to 3.7× faster. The Neural Processing Unit (NPU) saw the largest structural changes of any block, moving from a Lite and a Tiny core in the Kirin 9020 to a Lite and two Tiny cores in the Kirin 9030. Both core types also show significant layout changes. This is a reversal in Huawei’s NPU design. The Kirin 9000 5G, its last flagship chip on TSMC N5, used two Lite and one Tiny core. The series of SoCs on SMIC N+2 moved to one Lite and one Tiny core, likely for area savings. With the Kirin 9030, Huawei has shifted back toward a larger multi-core NPU cluster, but with the additional area going to a Tiny core rather than a Lite core. We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline or to commission a custom teardown, contact sales@semianalysis.com. Interested in joining us on this ride and think you can be a difference maker? Check out our Careers page. Before diving into the process stack, the package and memory are worth separating from the SoC itself. The Pro variant of the Kirin 9030 carries 12 GB of Samsung DRAM, with two stacks of four dies each. The dies were identified as the K4L2E165YD, a 12 Gb LPDDR5X-9600 device fabricated on Samsung’s 1a node, the fourth generation of its 10 nm-class DRAM after 1x, 1y and 1z. 1a has shipped in volume since 2022, so this is current memory rather than older-node inventory. The 16 GB Pro Max variants we obtained were found with both CXMT and Samsung packages. The CXMT package is marked CXDD7JEDM, with two stacks of four dies, packaged in week 45 of 2025. The inferred die dimensions from X-ray computed tomography (CT) are consistent with a known density of ~0.3 Gib/mm² for the CXMT G4 process, roughly equivalent to other manufacturers’ 1z processes. The Kirin 9030 uses a typical integrated package-on-package (iPoP) stack: multiple DRAM dies in a memory package sit above an organic redistribution layer (RDL) interposer, which sits above the SoC and package substrate. The full package is then mounted to the printed circuit board (PCB) through ball-grid array (BGA) solder bumps. The memory package substrate is a thin bismaleimide-triazine (BT) laminate carrying the LPDDR5X stack. The organic RDL interposer over the SoC routes the PoP signals around the die and carries possible dummy thermal copper pillars. The package substrate, a thicker Ajinomoto Build-up Film (ABF) build-up over a BT core, fans the flip-chip bumps out to BGA pitch and embeds the power planes. The whole stack is organic. The only silicon is the SoC and the LPDDR5X dies; there is no silicon interposer. Keeping it all-organic brings the package’s coefficient of thermal expansion (CTE) close to the PCB’s, reducing board-level warpage, and avoids the cost of a silicon interposer the SoC’s bandwidth does not need. In an iPoP stack, the memory package connects to the organic RDL interposer through an array of solder bumps. Underfill fills the gap around those bumps, adding stiffness and protecting the joints from mechanical stress. The Pro and Pro Max variants differ here, which we cover behind the paywall. The die shot and architecture tell us how Huawei allocated its silicon budget. The process tells us what SMIC can manufacture. We use the Helio G99 as the process reference for TSMC N6. Both SMIC N+3 and TSMC N6 are evolutions of previous 7 nm-class nodes. We used targeted TEM cross-sections through logic and memory regions, imaged in both fin-cut and gate-cut directions. Each cross-section caption gives its horizontal field width (HFW), the real width of the imaged area. We start at the transistor fins, then move up through standard cells, local interconnect, and SRAM. SMIC has not overtaken Intel or TSMC. It uses aggressive DUV scaling and DTCO to reach N6-class density, but that density doesn’t translate into comparable performance and efficiency, for two reasons: the node gap to leading-edge nodes, and Huawei’s core designs. Fin Profile One of the most important knobs in a FinFET process is the fin profile: the shape of an individual fin and the channel where current passes from source to drain. The ideal fin is tall, narrow, and nearly vertical. A taller fin increases effective channel width, while a narrower fin improves electrostatic control by thinning the body the gate must control. Push either too far, and the process pays for it: weaker drive current, fragile fins, taper, footing and line-edge variation that hit yield and device variability. The Intel 22 nm, 14 nm, and 10 nm fin cross-sections show how FinFET nodes have improved over time. 22 nm fins were a first-generation structure, relatively short, wide and strongly tapered. The shape limits current density and reduces gate control uniformity across the height of the fin. At 14 nm and 10 nm, Intel pushed the fins taller and narrower while also making the sidewalls more vertical. Rather than shrinking the device, these changes increase the effective channel width per fin and improve electrostatic control. The trade-off is that taller fins at tighter pitches are much more difficult to manufacture. Now, let’s compare the Helio G99 on TSMC N6 with the Kirin 9030 on SMIC N+3. Both processes are in the same class, with fin pitch of 30-32 nm on N+3 and 34 nm in our N6 cross-section. The pitch for N6 is especially interesting as N7’s HD library is generally listed with a 33 nm fin pitch, and N6 did not shrink pitches directly. Its density gains came from DTCO instead of tighter pitches. The 34 nm pitch was stable across our sampled region and serves more as a comparison against the SMIC N+3 we have not investigated further. Pinning down N+3’s fin patterning scheme takes more than one core unit. The CPU cores show a dense ~32 nm pitch, with the pitch between N-P fin pairs alternating between 78 and 88 nm. Logic alone may be consistent with dual-pitch mandrels of 120 and 110 nm, but this is a complex and unusual approach. Combining the pitch from the the 8T SRAM, which has more complex repeat unit, with the CPU core sequence allows us to reverse engineer the patterning steps with more confidence. As both the logic and SRAM should share the same base grid, a single CD mandrel lithography pattern with 128 nm pitch undergoing SAQP produces a die-wide ~32 nm grid (128 nm/4), which supports the pitch sequencing seen in both logic and SRAM cells. In the sampled cross-sections, N+3 shows a taller, narrower, higher-aspect-ratio fin than N6. The measured fin aspect ratio is ~9.5:1 on N+3 versus 7.8:1 on N6. N+3 also shows less top rounding, with an estimated radius of ~2 nm, compared with 2.8 nm on N6. Even though the fin widths differ, the ratio of top rounding to fin width tells the same story, with N+3 at 0.37 and N6 at 0.44. In a geometric sense, lower is better; a perfectly rectangular fin would have no top-rounding penalty. These are single-digit-nanometer features measured from a handful of cuts, so treat the absolute numbers as approximate. The important result is the relative gap: N+3’s fins are consistently taller, narrower and less rounded than N6’s. We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline or to commission a custom teardown, contact sales@semianalysis.com. Interested in joining us on this ride and think you can be a difference maker? Check out our Careers page. Standard Cell A standard cell is the basic building block of chip layout: a fixed-height row pairing one NMOS and one PMOS transistor that share a gate, tiled in a grid to build logic blocks. The key dimensions are contacted gate pitch (CGP), cell height (CH), fin count, and the lower-metal routing grid. To measure density, we use the Bohr metric: a weighted average of NAND2 gate area (60%) and scan flip-flop area (40%). This represents a realistic mix of combinational and sequential logic. This metric has its limitations, especially for complex cell layouts like TSMC’s FinFLEX, which alternates cells with different fin counts. Even so, it is the best metric for a pure process-level comparison. Another important measurement is the fin pitch; it refers to the distance between two fins of the same transistor. In a FinFET process, multiple fins are used in each transistor to increase the drive current and thus performance. TSMC N6 ships both a high-density (HD) library with 2 PMOS and 2 NMOS fins per cell, and a high-performance (HP) library with 3 of each. More fins under the shared gate mean more effective channel width. HP cells switch harder at the cost of area. Designers mix the two on a die, primarily spending HP cells on timing-critical paths, and matching their PPA targets. In the Cortex-A55 core of the Helio G99, we found a cell height of 240 nm for the HD cell. MediaTek has used HD cells in the G99 to minimize die size and thus cost. As an SoC for budget smartphones at ~$100, this is essential. By contrast, we found only one library in the Kirin 9030, with 2 NMOS and 2 PMOS fins. This suggests a narrower library strategy than TSMC N6, where both HD and HP libraries are widely used. This likely reflects the smaller customer base and the more constrained domestic design and electronic design automation (EDA) ecosystem. In all three CPU cores of the Kirin 9030, we found cell heights of 228 nm, 5% smaller than on N6. This is also a reduction of 9.5% over SMIC N+2’s cell height of 252 nm. SMIC N+3 and TSMC N6’s HD library both feature a CGP of 57 nm. For SMIC, this is a 9.5% shrink over N+2. In the past, CGP and cell height alone may have been enough to compare transistor density. Now, however, we must consider scaling boosters and DTCO as well. SMIC’s density gain does not come from EUV. It comes from using every available DTCO booster aggressively. First is fin depopulation: reducing the number of NMOS and PMOS fins in each cell. The first FinFET nodes started with 3 or 4 fins for each transistor. SMIC N+3 and TSMC N6 HD both use only 2 fins per transistor, trading drive strength for density. Next is contact over active gate (COAG). By landing the gate contact directly over the active gate, instead of out over the isolation region, the cell height drops. N+3 integrates COAG while N6 does not. Our N+3 gate-cut cross-sections indicate COAG, with the gate contact sitting over the active region, while N6 shows an off-gate contact. Last is single diffusion break (SDB). Diffusion breaks are inserted between cells in the same row to provide electrical isolation, but they also introduce local layout effects (LLE), layout-dependent shifts in electrical characteristics. In the past, a double diffusion break was used, consuming the space of two CGPs. SMIC N+3 and TSMC N6 instead use SDB, saving area but increasing LLE sensitivity. This must be controlled at the process level and accurately modeled in the process design kit (PDK) so that EDA tools can account for it. Overall, SMIC N+3 has a transistor density of 113.4 MTr/mm², slightly above TSMC N6 at 107.7 MTr/mm². Even without EUV, SMIC has achieved density beyond TSMC’s mature N6 node which utilizes EUV. Metal Stack The smallest critical dimension in the teardown is M0; SMIC N+3 uses a 32.5 nm local metal pitch. That is smaller than the 36 nm M0 pitch on Intel 18A in Panther Lake. However, this does not mean that SMIC has a better process than Intel 18A or TSMC N3P. M0 is a local intra-cell routing layer. Its usefulness depends on the full interconnect stack: M1 and M2 pitch, track count, via and line resistance, design rules, mask count, overlay control, and routing flexibility. The 32.5 nm M0 is consistent with self-aligned quadruple patterning (SAQP), whose four-population line-width loading we read coarsely as alternating widths of 21.5 to 24 nm; M1 and M2, at 38 and 40 nm, are consistent with self-aligned double patterning (SADP), a single A/B split. On TSMC N6, M0, M2, and M3 sit at a relaxed ~40 nm and are consistent with SADP-class double patterning, with no need for quadruple patterning. That said, we measure M2 for example at ~43 nm, likely inflated by sparse routing. We do not assign any specific layer to EUV from our cross-sections; the distinction we can draw is double versus quadruple patterning, not lithography wavelength. Transistor-level density in the front-end-of-line (FEOL) sets an upper bound, but the design is ultimately limited by what the interconnect stack can route. The lowest metals are the most important for standard-cell density, but the semi-global and global layers determine how usable that density is at the block and chip level. Two axes are commonly used for chip cross-sections: the fin-cut and the gate-cut. The micrograph above is a fin-cut and shows metals 0 through 3. This axis lets us see the even-numbered metals, with M0 right above the fins. There are two kinds of M0 lines. The first are the power rails; these are wide wires for the VDD and VSS running horizontally at the top and bottom edges of each standard cell. The wide wires measure 55 nm across, more than double the other M0 lines. Their width minimizes resistance and reduces IR drop. The second kind are intra-cell wires, short segments within the cell that connect terminals to M1. These have alternating widths between 21.5 and 24 nm. The M0 pitch is 32.5 nm, a 19% reduction versus N+2 and N6. At this pitch, DUV patterning requires more aggressive multi-patterning, increasing mask count, overlay sensitivity, process complexity and cost. M0 is below what a single DUV-defined spacer (SADP) can resolve, so SMIC cascades a second spacer step (SAQP). The cross-section reflects the cost: the M0 trenches are visibly more re-entrant (narrower at the bottom than the top) than M1 or M2 on the same chip and carry a bright barrier-rich foot where the trench meets the etch-stop layer. That shape is partly the intended damascene profile, as a slightly narrow bottom helps void-free copper fill, but its magnitude at M0 is driven by the tight pitch and the higher trench aspect ratio. Intel 18A supports an M0 pitch of 32 nm, although Panther Lake has only shipped with a looser 36 nm pitch. This is due to Intel’s heavy usage of HP libraries. Among leading-edge nodes, 18A has the loosest M0 pitches due to PowerVia. With power routing moving to the backside, congestion is reduced, and the entire front-side metal stack can be used for signal routing. M2 is the first true inter-cell routing layer. It runs horizontally like M0 but spans across multiple cells to carry block-level signals. The M2 pitch sets the cell’s track height – the number of M2 tracks that fit between the VDD and VSS rails, defining what the library calls a 6-track or 7.5-track cell. This layer is the most important, limiting the routing of entire blocks. SMIC N+3 features a 5.7-track cell. The M2 pitch is 40 nm, a 5% decrease over N+2 and the same as N6. This shrink keeps the pitch at the edge of what is possible with double patterning. Future nodes will need to increase the number of masks for M2 as reducing the number of tracks is much harder due to the limitations in routing. The micrograph above is in the perpendicular direction, the gate-cut, and shows metals 0 through 4. This allows us to see and measure the odd-numbered vertical metal layers. The M1 pitch is 38 nm, 9.5% less than N+2 and 33% less than N6. The M1-to-gate ratio matters because it sets local routing flexibility. N+2 and N+3 use a 3:2 ratio, while N6 uses a 1:1 ratio, explaining the huge differences in M1 pitch. The more M1 lines there are compared with the gates, the more flexibility there is for power and signal crossing within the cell. Routing flexibility enables more complex and better cells. Clean fractional ratios are also preferred as a grid is periodic and improves layouts. The 3:2 ratio gives SMIC more local routing flexibility than a strict 1:1 grid, but it also complicates layout and patterning. This is a DTCO choice, with SMIC increasing process complexity to recover density and routability without EUV. This 3:2 ratio is not very popular in the leading-edge nodes. TSMC has only used it on N7+, the N5 family, and the short-lived N3(B). They have switched back to a 1:1 ratio for N3E. Intel has only used it on the 10 nm/Intel 7 family, with Intel 4, 3 and 18A all using a 1:1 ratio. Samsung is the only one still using a 3:2 ratio at the leading edge, using it in the SF4 and SF3 families. It remains to be seen if SMIC will remain at a 3:2 ratio or move to a 1:1 ratio with its future nodes. The industry is still actively exploring these local-routing ratios. At VLSI 2026, imec will be presenting work on even higher ratios, including a 2:1 scheme that can reduce area by up to 14%. We will be covering the conference in a future newsletter article. Subscribe We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline or to commission a custom teardown, contact sales@semianalysis.com. Interested in joining us on this ride and think you can be a difference maker? Check out our Careers page. The final local interconnect layer for N+3 is M3, with a pitch of 44 nm. The M3 pitch is the same as on N+2 and 10% larger than on N6. The semi-global layers carry the majority of block-level signal routing. They have a coarser pitch than the lower local layers. On leading-edge nodes, they are designed to sit at the limit of DUV single patterning. M4 through M11 pitches were found divided between 80–82 nm (M4–M6), 128 nm (M7–M10), and 148 nm (M11). Given limited sampling, it is possible these are divided further in dense routing areas. At the top are two giant metal layers, M12 and M13. These have kept the same pitches as N+2 at 1920 nm and 4600 nm respectively. While the lower layers’ pitches are generally fixed by the process and library, the upper layers vary much more in pitch and count, depending on the design. Even two smartphone SoCs on the same process can have wildly different metal stacks. The Helio G99 carries fewer routing layers, reaching coarse metal pitches of 850 nm by M9, while the larger and higher-performance Kirin 9030 keeps fine pitches until M11. SRAM At the leading edge, SRAM is much more difficult to scale than logic. TSMC’s latest nodes have seen little to no bitcell scaling, while logic still has more DTCO levers to pull. While looking for other logic libraries in the GPU compute units, we stumbled upon the SRAM. The most common type of SRAM has 6 transistors (6T), but this cell had 8 transistors (8T) instead. 8T SRAM adds two transistors to form a dedicated read port. Unlike a 6T cell, where reading disturbs the storage, the decoupled read port removes read-disturb, improving read stability and letting the cell be pushed harder for performance. At first glance, the cut looked like an unusual logic library, with each cell row having 3 fins of one polarity and 5 fins of another. The rows also alternated in orientation. Energy-dispersive X-ray spectroscopy (EDS) resolved our confusion. The cut had not landed on the GPU logic, but on the SRAM macro beside it. The unusual fin pattern was due to the SRAM library. We return to EDS in the process flow analysis behind the paywall. SRAM libraries are not like traditional logic libraries. Due to the unequal number of PMOS and NMOS transistors, they require specialized rules and layout libraries. They do not need the flexibility of logic libraries, so they are hyper-optimized for one purpose: dense, reliable memory. The SRAM cell we found is a 1:2:2-2:2 cell. This means there is 1 fin per pull-up (PU) PMOS transistor, and 2 fins per pull-down (PD) and pass-gate (PG) NMOS transistor. These 2 PU, 2 PD and 2 PG transistors would usually form a single 6T high-current cell (HCC). An 8T HCC adds a read-pull-down (RPD) and a read-pass-gate (RPG) NMOS transistor, each with two fins. We measured a cell height of 406 nm, which brings the bitcell size to 0.0463 µm². That is a theoretical peak density of 21.6 Mib/mm². We estimate that a 6T HCC would have a cell height of 292 nm and a size of 0.0337 µm². This is ~12% larger than a 6T HCC on Intel 3 and 4. We also estimate the 6T high-density cell (HDC) to have a cell height of 228 nm and a size of 0.0260 µm². This is coincidentally the same as the logic standard-cell height measured earlier. The estimate puts the cell near Samsung 7LPP/5LPP and slightly below TSMC N7/N6. That is a theoretical peak density of 38.5 Mib/mm². 6T HDC is arguably the most important cell as it is used for the largest caches in a chip, the L3 caches and system-level cache (SLC). Both the Kirin 9020 and 9030 have split the SLC into 4 banks to raise total SLC bandwidth. In the Kirin 9030, the SLC increased from 2 MiB to 3 MiB per bank. Correspondingly, the number of arrays within the bank also increased by 50%, from 16 to 24. Each array can store 128 KiB and forms an orderly pattern on the die shot. From the Kirin 9020 to the Kirin 9030, the area of a 128 KiB SLC array decreased from 0.0477 mm² to 0.0392 mm², an 18% shrink. The achieved density is 25.5 Mib/mm², 66% of the theoretical maximum. While the SLC was quite similar across both chips, the L3 has seen some major changes, particularly in terms of its layout. The total capacity also went up from 10 MiB to 12 MiB. Much like the SLC, the L3 is also split into 4 banks. In the Kirin 9020, an L3 bank consisted of 16× 128 KiB arrays and 16× 32 KiB arrays. However, an L3 bank in the Kirin 9030 instead consists of 48× 64 KiB arrays. In the Kirin 9020 L3, a 128 KiB array was 0.0513 mm² and a 32 KiB array was 0.0154 mm². The size of the 128 KiB array is different on the L3 and SLC as the assist circuitry for the two arrays differs depending on their purpose. In the Kirin 9030 L3, a 64 KiB array is 0.0210 mm². Although not a like-for-like comparison, normalized for capacity, it is 18% smaller than the 9020’s 128 KiB L3 array and 31% smaller than its 32 KiB L3 array. The achieved density is slightly lower than the SLC, at 23.8 Mib/mm², 62% of the theoretical maximum. Unlike the L3 and SLC, the prime cores’ private L2 cache uses a 2-bank design. As the prime cores’ L2 is latency-critical, it likely uses 6T HCC instead of 6T HDC. The 9020 has 16 arrays in each bank while the 9030 has 32. Each array has a capacity of 32 KiB. A 32 KiB array in the L2 shrank from 0.0171 mm² to 0.0142 mm², ~17% smaller. The density is 17.6 Mib/mm², ~59% of the theoretical maximum for 6T HCC. SRAM scaled well from N+2 to N+3, shrinking by ~19%, close to the theoretical logic shrink. The caveat is that N+2’s bitcells were unusually large, bigger than comparable 7 nm-class nodes, so part of the gain is catch-up rather than true scaling. With the insights from STEEL’s teardowns, we will be doing a deep dive into SRAM in a future newsletter article. Subscribe Everything above came out of a single STEEL teardown: die annotation, block-level area analysis and TEM cross-sections through logic and SRAM. We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline or to commission a custom teardown, contact sales@semianalysis.com. Future Roadmap The same cross-sections that pin down N+3 also show where SMIC can go next. Although N+3 is already close to the practical limits of DUV multi-patterning in several layers, SMIC still has a few scaling levers left. A theoretical N+4 would likely start with cell height. N+3 uses 5 M0 tracks between its power rails. Moving to 4 M0 tracks, as on SMIC N+2 and TSMC N6, could reduce cell height by roughly 15%. The routing grid is only one side of the shrink; the front end also must fit into the smaller cell. One possible FEOL lever is reducing the p-to-n isolation spacing from two diffusion grid units to one. Intel used this scaling booster on Intel 4, and TSMC did so on its N3 family. This path trades layout flexibility for density. Fewer M0 tracks reduce local routing resources, while tighter p-to-n spacing raises integration and design-rule difficulty. M2 is also constrained by the cell height shrink. For SMIC to maintain a ~5.7-track cell, M2 would need to move toward ~35 nm. That would move another layer into SAQP territory. SMIC could also reduce the CGP from 57 nm to 54 nm. Intel reached a similar CGP on Intel 10 nm/Intel 7 without EUV. The local interconnect is also tougher. If SMIC keeps the 3:2 M1-to-gate ratio, M1 would need to shrink to 36 nm and would likely require SAQP as well. If SMIC moves to a 1:1 ratio, M1 could relax to 54 nm, but it would give up routing flexibility. Under this theoretical path, we estimate that SMIC N+4 could reach a cell height of 198 nm and a CGP of 54 nm, implying a Bohr density of 137.8 MTr/mm², on par with TSMC N5 or Samsung SF4. However, the difficulty is cumulative. Each step is individually plausible, but together they make N+4 harder than the transition from N+2 to N+3. It will likely take longer, cost more, and carry less process margin. A theoretical N+5 would require a larger integration shift. One possible path is backside contacts (BSCon), moving power routing and source/drain contacts to the backside, which would reduce front-side routing pressure and enable another reduction in cell height. The front-side metal pitches could relax to reduce the process complexity. M0 would likely relax slightly to ~34 nm, and M2 and M4 pitches could relax further. CGP is unlikely to shrink much further. Even with EUV, 48 nm has been the practical limit for yield and process control. This approach would allow N+5’s cell height to fall to 170 nm and its CGP to 53 nm. This implies a Bohr density of 163.6 MTr/mm², on par with Intel 18A’s HP library. However, this would not make N+5 cost-competitive with the leading edge. It would reach a similar density through a much more expensive route. The integration difficulty rises sharply, with new process flows for backside alignment, wafer thinning, contact reveal, and backside metallization. Past this point, standard density and interconnect scaling become increasingly unattractive. That is where Huawei’s roadmap stops looking like a normal foundry roadmap and starts looking like a packaging roadmap. At ISCAS 2026, Huawei unveiled its tau (τ) scaling law, reframing process scaling in the time domain. τ is the time cost of data movement and processing: switching delays in transistors, RC signal propagation delays in circuits, compute, memory, and networking latency. Outside Huawei’s terminology, this is called system-technology co-optimization. This is Huawei’s answer to its lack of EUV lithography. Without EUV, planar density cannot keep pace with TSMC, Intel, or Samsung. If transistor density cannot shrink further, Huawei’s alternative is to shorten wires, reduce buffering, and stack logic vertically. “LogicFolding”, Huawei’s implementation of this new scaling idea, is, in practice, an aggressive 3D stacking approach. AMD V-Cache places SRAM above or below a CPU die. AMD’s MI350X places active interposer dies (AIDs) underneath accelerator and compute dies (XCDs), with the AIDs handling cache, IO interfaces, the network-on-chip (NoC) and embedded metal-insulator-metal (MIM) capacitors. With LogicFolding, parts of the same logic block are split across multiple active dies bonded face-to-face at ultra-fine pitch. This allows Huawei to shorten some critical paths and reduce buffer overhead, not merely add cache capacity or offload the IO and interconnect. Shortening wires is where the higher clocks come from. A large share of a modern core’s delay and energy budget goes into driving long interconnects and the repeater buffers along them. LogicFolding distributes a block’s critical-path gates across multiple stacked tiers bonded at a very fine pitch, so the bond interface behaves like an additional metal layer and the longest paths get shorter. That is how Huawei expects to recover frequency and efficiency it cannot get from the process alone. Huawei’s roadmap shows its intent. Prime core frequency is targeted to rise from 2.75 GHz in the Kirin 9030 to roughly 5 GHz by 2031, far beyond what planar scaling alone could deliver. Prime cores with 3.1 and 3.39 GHz clocks are being tested in its labs, although their power consumption is unknown. Beyond that, chips are in the design, simulation or pathfinding phase, meaning the frequencies are targets. However, the direction matters more: LogicFolding also helps with performance, not just density. The catch is that Huawei’s density claim is not directly comparable to foundry densities. A stacked design can report more transistors per package footprint by adding active layers, even if each patterned die remains well behind TSMC or Intel in front-end density. This is how Huawei can claim to reach foundry 14A-equivalent density by 2031. This is not a like-for-like foundry comparison, with Huawei using stacked logic and measuring density per package footprint. On a normalized Bohr-density basis, SMIC N+3 is ~114 MTr/mm², 38% less than Intel 18A’s HD library. Huawei’s 3D roadmap closes the gap by stacking active logic, reaching 215 MTr/mm² by 2030. In 2031, the roadmap density jumps to 295 MTr/mm², implying either a third active layer, partial EUV insertion or aggressive planar DUV scaling. Huawei’s methodology makes other foundries look much denser as well. Applying it to AMD’s MI450X with an N2 top die on an N3P base die yields a theoretical density of 460.2 MTr/mm² in 2026, compared with Huawei’s 295 MTr/mm² in 2031. This Kirin 9030 does not use LogicFolding, remaining in a conventional mobile SoC package. Instead, it forms the baseline for how far Huawei and SMIC can push planar scaling. Future teardowns of Kirin and Ascend chips will show both planar logic density and Huawei’s hybrid bonding solutions. Export controls changed China’s optimization problem rather than ending it. EUV restrictions raised the cost and complexity of leading-edge manufacturing without freezing it. SMIC reaches N6-class logic density through DUV immersion, SAQP and DTCO, while Huawei shifts more of the burden onto architecture, packaging and system-level integration. Future nodes will be tougher. N+3 still had room to tighten local metals and reduce cell height and CGP. Further scaling without EUV leaves fewer levers. More aggressive multi-patterning adds masks and overlay error. SMIC can keep pushing DUV, but each step will get more expensive and less forgiving. The design side is just as critical. Huawei had domestic EDA tools and flows before the Kirin 9030, with the Kirin 9000s, 9010 and 9020 making that clear. Huawei was able to ship multiple consumer SoCs on SMIC N+2 and N+3 while cut off from the Western EDA stack. US export controls restricted EDA tools for advanced chips in 2022 but did not target tools for more mature chips. In 2025, the US government briefly placed much broader restrictions on EDA software from Synopsys, Cadence, and others, before lifting them less than two months later as part of a trade deal tied to rare earths. Huawei has been unable to access those tools because it remains on a US trade blacklist. That forced Huawei, SMIC and Chinese academic institutions to build their own tools and flows. Researchers at Peking University recently announced a prototype EDA tool for Huawei’s LogicFolding architecture, which requires a new flow to handle the multilayer layout and floorplan. This is not the same as replacing the full Synopsys or Cadence stack, but it shows where domestic EDA is headed: toward tighter co-optimization between architecture, process and packaging. These advances are also diffusing into the Chinese ecosystem. SMIC is licensing its N+2 and N+3 processes to HLMC/Hua Hong at the government’s direction rather than by choice. If the same process learning feeds into Ascend accelerators for AI training and inference, the choke point shifts from one named fab to an ecosystem. Alibaba’s T-Head silicon arm and Cambricon, a Chinese AI chip designer that is expected to supply ByteDance, could also be major beneficiaries. Sanctions aimed at SMIC alone become less effective once the manufacturing knowledge has spread to other fabs and design houses. China is not closing the gap with Intel, Samsung and TSMC. The teardown shows the opposite in several places: no EUV, no backside power, higher process complexity, and visible trade-offs. But China is still advancing. If domestic chips become good enough for phones, inference, networking and security-sensitive workloads, they can matter strategically without matching TSMC at the leading edge. Behind the paywall, we show what else STEEL can do, with material and process flow analysis of SMIC N+3, and analysis of the Kirin 9030 package. We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline, access the full Kirin 9030 and SMIC N+3 analysis or to commission a custom teardown, contact sales@semianalysis.com.
The chance for history awaits Arsenal
Arsenal are back in the Champions League final for the first time in 20 years and are on the cusp of a historic double. After lifting the Premier League title after a 22-year wait last week there had been a suggestion that this match against Paris St-Germain (17:00 BST kick-off) is something of a free hit. Arsenal fans have travelled to Budapest in their thousands such is the magnitude of this match.
Bonus 228: The $1.776B Political Question
Every so often a story comes along that is so brazen, so cleanly emblematic of the moment we are living through, that staying quiet about it feels like a kind of complicity unto itself. Monday’s announcement by the Department of Justice that the Acting Attorney General has unilaterally created a $1.776 billion “Anti-Weaponization Fund” to compensate so-called “victims of lawfare” is one of those stories. And the reason I want to write about it here, in a newsletter ostensibly devoted to the Supreme Court and the law and politics surrounding it, is that I’m increasingly worried that we’re asking the wrong institution to save us from one of the most offensive, alarming, and anti-constitutional things we’ve seen from any president in American history. As I explain below the fold, the Anti-Weaponization Fund is not principally a legal problem awaiting a judicial fix. It is a shockingly egregious political abuse—the President suing his own government, purporting to settle with himself, and routing nearly $2 billion from the Treasury to his own supporters as a not-so-subtle (and not-so-formal) reward, if not an incentive, for political violence. And the Constitution’s answer to political abuses of this kind and magnitude is a political remedy, not a legal one. In his majority opinion in NFIB v. Sebelius (the Affordable Care Act case from 2012), Chief Justice Roberts defended the Court’s endorsement of Congress’s power to adopt the individual mandate by noting that it is “not our job to protect the people from the consequences of their political choices.” As much of a judicial supremacist as I am, that sentiment is, for better or worse, an entirely correct (and, in the long term, normatively desirable) institutional assessment. One can believe in substantial judicial power without believing that literally every political dispute in our country can and should be resolved by unelected judges. The real upshot of this latest headline (among so, so many other similar headlines) is that it is past time to reinvigorate impeachment—not because conviction is realistic in this Senate, or even because impeachment itself is feasible in this House, but because forcing every member of Congress to vote on the record whether this brazen, corrosive, and affirmatively dangerous corruption is impeachable is itself a point worth fighting for (and fighting with our friends over). Assuming the courts will save us absolves our political leaders of both responsibility and accountability for allowing this kind of mischief to continue, and it absolves all of us of our responsibility to refuse to tolerate those who would just as quickly shrug their shoulders and move on. For those who aren’t paid subscribers, we’ll be back (no later than) Monday with our continuing coverage of the Supreme Court. For those who are, please read on.
New MatterChat Model Helps AI to ‘See’ the Language of Science
From writing emails to generating computer code, much of the artificial intelligence prevalent in our daily lives has succeeded by mastering one domain: text. However, this leaves a major blind spot in the physical sciences, where models depend on the high-resolution, three-dimensional data of the physical world, like the intricate lattice of atoms in a crystal. Delivering on the promise of using AI for science requires teaching these data-driven text models to seamlessly “talk to” physics-based models. Now, a new AI framework from Lawrence Berkeley National Laboratory (Berkeley Lab), called MatterChat, solves this problem by creating a specialized “bridge.” It connects the conversational power of a Large Language Model (LLM) with a physics-based AI that models “interatomic potentials”: the complex physical forces between atoms. The resulting system already significantly outperforms general-purpose AI tools like GPT-4 at predicting material properties, and the team hopes it can accelerate scientific discovery by serving as a robust research partner that provides grounded insights and generates step-by-step instructions for synthesizing novel materials. A paper describing this work was recently published in Nature Machine Intelligence. “Traditional simulations can provide the physical rigor required for materials science, yet their computational cost remains prohibitive for high-throughput screening. Conversely, while LLMs excel at rapid knowledge synthesis, they inherently lack the ‘structural vision’ to interpret materials directly from their underlying atomic coordinates,” said Yingheng Tang, a postdoctoral researcher in Berkeley Lab’s Applied Math and Computational Research Division (AMCR) and lead author on the paper. “MatterChat was built to solve this dilemma, empowering LLMs with a structural ‘vision’ that allows researchers to leverage their full potential for solving complex, real-world materials challenges.” Empowering language models to solve complex challenges in materials science To build MatterChat, the Berkeley Lab team drew inspiration from technologies like Vision Question Answering (VQA) and Text-to-Image (T2I) generation. In these tasks, AI must translate high-level text concepts into visual images, or vice versa. Doing so requires developers to build tools that “bridge” two fundamentally different forms of data. The researchers adapted this concept to the physical sciences. With MatterChat, they created a “bridge model” that successfully connects an LLM’s general knowledge with the deep understanding of the atomic-scale world encoded in scientific interatomic potentials. Until now, researchers using LLMs to solve materials problems usually had to feed them raw data files as if they were just strings of text. It’s like asking an AI to understand a complex 3D engine based only on a parts list: the LLM can read the names, but it can’t “see” how the atoms fit together in space. MatterChat solves this by training a specialized AI bridge model, pre-trained on millions of crystal structures and an LLM, to align the LLM’s representation of the world with the interatomic potential’s representation of the world. The bridge model can translate physical insights into a format the LLM can actually understand. By giving the LLM these “scientific eyes”— a scientific “inductive bias” in the terminology of AI — the Berkeley Lab team has transformed it into a robust research tool capable of providing grounded scientific insights into complex materials challenges, such as predicting thermal stability or analyzing electronic band gaps. “We think of atoms as living in a physical space, but from a machine learning perspective, they are just vectors living in some very non-trivially structured manifold in a high-dimensional Euclidean space; and, of course, the same is true for the sentences and paragraphs that we use to express our ideas about those atoms,” said co-author Michael Mahoney, Berkeley Lab’s Scientific Data Division (SDD) AI Initiative Research Lead. “The bridge model basically gets those two structures to ‘talk with’ each other.” As a proof-of-concept of this general approach, the team trained their bridge model on a dataset curated by pairing nearly 143,000 stable atomic structures from the Materials Project with their corresponding physical properties. This training data was automatically assembled using the Materials Project’s API and deliberately enriched with properties fundamental to microelectronics design — like formation energy and bandgap — allowing MatterChat to learn the complex patterns connecting a material’s atomic blueprint to its functional performance. To validate their model, the researchers benchmarked MatterChat against a suite of other AI systems, from general-purpose LLMs to other specialized scientific AI methods. The results show that MatterChat consistently outperformed its competitors across a range of tasks. The model was more accurate in classifying material types and demonstrated superior precision in predicting numerical properties. For example, it excelled at predicting a material’s bandgap, a property critical for designing new electronics from high-capacity energy storage to next-generation computer chips. “Our design is significantly more efficient because we don’t have to build a massive AI model from the ground up,” said co-author Zhi (Jackie) Yao, a research scientist in Berkeley Lab’s AMCR. “Instead, we take two powerful, pre-trained models — a structural encoder for materials physics and an open-source LLM — and use them off-the-shelf. The only component we actually train is the lightweight ‘bridge model’ that translates between them. It’s the difference between building an entire car factory and simply designing a smart adapter that connects a world-class engine to a world-class navigation system. This approach is not only computationally efficient, but also makes the system modular, so we can easily upgrade components or adapt the bridge for other scientific domains in the future.” Crucially, this modular design highlights exactly how institutions like Berkeley Lab and the Department of Energy are carving out a highly valuable niche in the booming AI landscape. Rather than competing with Silicon Valley tech giants to build ever-larger language models from scratch, the lab is focusing on the specialized connective tissue that makes commercial AI useful for hardcore science. Because the bridge model approach underlying MatterChat is forward-compatible, it is perfectly positioned to leverage these parallel tracks of innovation. As Mahoney pointed out, “We expect that industry will continue to develop improved LLMs, and we expect domain scientists and facilities will continue to generate new data. An important part of scientific machine learning is not simply to solve problems on today’s data, but instead to develop general methods that will be forward-compatible with orders of magnitude more data, whether from scientific domains or from LLMs.” According to Yao, the MatterChat project, which was initially developed and enhanced with funding from a Berkeley Lab Laboratory Directed Research & Development (LDRD) Program, will now expand its capabilities. In a collaboration with Fermilab, MatterChat is already contributing to a U.S. Department of Energy Genesis Mission project — called Accelerating eXtreme Environment Specs-to-Silicon (AXESS) — that aims to speed up the development of next-generation, high-speed, radiation-hardened detectors for challenging particle physics experiments by using advanced 3D integrated circuits (chiplets) and AI-driven data analysis. In addition to the LDRD support, the team also credits supercomputing resources at the National Energy Research Scientific Computing Center (NERSC), located at Berkeley Lab, with MatterChat’s success. “We are incredibly grateful to NERSC; this research simply would not have happened without access to the Perlmutter supercomputer through their AI for Science program,” said Tang. Wenbin Xu, a NERSC postdoctoral fellow at the time, was also a major co-author of the work, as was Benjamin Erichson, a research scientist in Berkeley Lab’s SDD, highlighting the benefits of AMCR-SDD collaboration on AI for science. ### Lawrence Berkeley National Laboratory (Berkeley Lab) is committed to groundbreaking research focused on discovery science and solutions for abundant and reliable energy supplies. The lab’s expertise spans materials, chemistry, physics, biology, earth and environmental science, mathematics, and computing. Researchers from around the world rely on the lab’s world-class scientific facilities for their own pioneering research. Founded in 1931 on the belief that the biggest problems are best addressed by teams, Berkeley Lab and its scientists have been recognized with 17 Nobel Prizes. Berkeley Lab is a multiprogram national laboratory managed by the University of California for the U.S. Department of Energy’s Office of Science. DOE’s Office of Science is the single largest supporter of basic research in the physical sciences in the United States, and is working to address some of the most pressing challenges of our time. For more information, please visit energy.gov/science.
Scientists discover hidden “brakes” that stop massive earthquakes
A mysterious underwater fault near Ecuador has been producing nearly identical magnitude 6 earthquakes every five to six years, baffling scientists for decades. Researchers now believe the fault contains hidden “brake zones” where seawater and unusual rock structures work together to stop quakes from becoming even larger. The discovery came from ultra-detailed seafloor recordings that captured how the fault behaves before and after major earthquakes.
Australian business sentiment crashes in March on worries about fallout from Iran war
Australian business confidence plunged last month by a magnitude only seen during financial crises as the Iran war gave rise to a global oil shock, a survey showed on Tuesday.
Science of seismic proportions
Arizona’s deserts might not be situated on a tectonic plate boundary where geographical activity can cause earthquakes, but that doesn’t make the state immune from the disturbance of tremors.It has been 139 years since Arizona’s strongest recorded earthquake on the Pitaycachi fault struck in 1887. The 7.5 magnitude event, known as the Sonoran Earthquake, took place near the southeastern Arizona-Mexico border.
Jorge Martin comeback win earns praise from fellow world champions
Both Francesco Bagnaia and Fabio Quartararo underlined the magnitude of what Martin achieved at COTA on Saturday
Glyphosate: A story of science, risk, and nuance
In the summer of 2014, a California groundskeeper named Dewayne Johnson started noticing patches of raw, weeping skin spreading across his body. He had spent years applying Roundup, Monsanto’s glyphosate-based herbicide, for a school district, sometimes drenched in it when equipment malfunctioned. By 2014, he had a diagnosis: non-Hodgkin’s lymphoma. In 2018, a jury awarded him $289 million in damages, later reduced on appeal. Tens of thousands of lawsuits followed. The story hasn’t quieted down. It’s all over the internet, swinging between “this will kill you, buy organic” and “relax, you’re fear-mongering.” Some of that recent chatter stems from an Executive Order protecting glyphosate and its production, which set off intense infighting within the MAHA movement. Some of it reflects a deeper truth: Americans genuinely care about harmful chemicals in their environment (see poll below). And the conversation isn’t going anywhere, with a Supreme Court case involving Monsanto set for this April. Glyphosate (the active ingredient in Roundup) is one of the most widely used herbicides in the world. But does glyphosate cause cancer? And how much exposure is actually risky? With so much mistrust in institutions and corporations (and some genuine conflicts of interest), making sense of the actual scientific evidence is genuinely hard. Here is a translation of what the scientific evidence actually says. Does glyphosate cause harm? Glyphosate is an herbicide that kills most plants it contacts rather than targeting specific species. It works by blocking an enzyme pathway (EPSPS) that plants and many microorganisms need to produce essential amino acids. Without those amino acids, the plant can’t grow and dies. Glyphosate was first developed by Monsanto in the 1970s and became ubiquitous after the company introduced “Roundup Ready” crops in the 1990s. In 2024, Bayer removed glyphosate from residential Roundup-branded products due to litigation pressure. But today it is still the most widely used herbicide in the world. Farmers also discovered they could spray it directly on crops like oats, lentils, beans, and peas just before harvest to dry them out faster and make harvesting easier. Because of its widespread use, it’s obviously really important to understand if it’s harmful to humans. The key principle: The dose makes the poison The most important concept in toxicology is that “the dose makes the poison.” Sixteenth-century physician Paracelsus articulated what remains the foundational principle of toxicology: sola dosis facit venenum. Water will kill you if you drink enough of it. Aspirin saves lives at 81 milligrams and causes internal bleeding at higher chronic doses. So, at what dose is glyphosate poisonous? The LD50 in rats (the dose that kills half the test animals) is around 5,600 mg/kg body weight. Table salt is roughly 3,000 mg/kg. By that classic measure, glyphosate is less acutely toxic than salt. But some people are exposed to a lot of glyphosate through their work. And among the general public, there are many questions about low-dose, cumulative exposure from multiple sources over long periods. Does glyphosate cause health problems? Cancer Correlation maps are striking, with one analysis finding that 60% of counties that spray the most glyphosate have non-Hodgkin’s lymphoma rates above the national average. But correlation alone doesn’t establish cause, which is why animal and human studies matter. In animals: Rat studies are typically what most fear is rooted in, but most people don’t know the study details on the dosage given to rats. A pooled analysis of the 13 most rigorous rat studies found evidence that very high doses, around 55,000 times greater than typical food residue exposure, can cause certain cancers. In humans: One of the strongest studies followed 55,000 herbicide users over 20 years and found no overall increased cancer risk. There was one signal: among those with the highest use, rates of acute myeloid leukemia (AML) appeared elevated. A broader meta-analysis of 18 human studies also found no overall cancer risk, though a separate meta-analysis focused specifically on non-Hodgkin lymphoma found that people with high cumulative glyphosate exposure faced a 41% higher relative risk compared to those with low or no exposure. These human studies are observational studies (randomized controlled trials would be unethical), and meta-analyses are only as reliable as the studies they pool. But the story seems consistent: at very high exposure levels, there may be some cancer risk. But that’s about as far as the science currently takes us. Emerging research: gut and reproductive health Gut: Because glyphosate targets a pathway that exists in some gut bacteria, scientists have questioned whether it could affect the microbiome. In mice, a 2023 University of Iowa study found that glyphosate altered gut bacterial composition and raised markers of intestinal inflammation. A 2024 systematic review found similar patterns: disrupted gut bacteria and problems with the mucus layer that protects the intestinal wall in mice. Whether this translates to humans is an open question. The human gut is highly variable and resistant, and we don't yet have strong studies confirming the same effects in people. Reproductive health: Animal studies have found that glyphosate-based herbicides can interfere with reproductive hormones, including disrupting egg cell development in mice and reducing the number of eggs available in sheep’s ovaries. However, the doses used in these experiments were high, so it’s not yet clear whether the same effects would occur in humans. On the human side, a 2021 pilot study led by Mount Sinai researchers tested urine samples from pregnant women across the U.S. and found glyphosate in 95% of them. The study also found a possible connection between higher glyphosate exposure during pregnancy and a subtle physical difference measured in female newborns. That said, some connections weren’t statistically reliable and this was a small study. The researchers themselves said more and larger studies are needed before drawing firm conclusions. What do regulators say about glyphosate safety? In 2015, the International Agency for Research on Cancer (IARC), part of the World Health Organization, classified glyphosate as "probably carcinogenic to humans" (Group 2A). The classification is worth taking seriously. But, there’s an important distinction: IARC classifies hazard, meaning the potential to cause harm under some conditions. It does not assess risk, which accounts for actual exposure levels and the real-world probability of harm. The U.S. EPA, the European Food Safety Authority, Health Canada, and regulatory agencies in Australia and Japan have all concluded that glyphosate is “unlikely to be carcinogenic to humans” at realistic exposure levels. They’re asking a different question: given how much of this people actually encounter, what is the probability of harm? Your actual exposure is what determines your actual risk. What does glyphosate mean for your daily life? At the grocery store: Food residues are small. The EPA’s acceptable daily intake is 1.75 mg/kg of body weight per day. To hit that threshold from oatmeal alone, a 150-pound adult would need to eat roughly 50 pounds of oats every day. But this calculation addresses only one pathway (food), and we don’t yet have robust tools to measure people’s actual total exposure. If reducing exposure matters to you, organic certification prohibits synthetic herbicides, and organic oat products generally test at lower or undetectable levels. Either way, a diet rich in whole grains, conventional or organic, is far better for your health than one that omits them out of fear of residues. On the farm: Occupational exposure is orders of magnitude higher than consumer exposure. Farmers and farm workers who mix, load, and apply glyphosate-based herbicides, especially without protective equipment, face meaningfully different risk profiles. Proper PPE isn’t optional. Integrated weed management strategies that rotate chemical modes of action are also worth taking seriously, both for health and for long-term effectiveness. If you live near a farm: Spray drift is real. Concentrations near field edges can spike during and after application, particularly on windy days. This is an area where I think public health could really step up and empower rural communities with buffer zones, rather than placing the onus on individuals. But for now, knowing your local agricultural calendar and keeping windows closed during application are the first steps. Bottom line Dewayne Johnson was exposed to glyphosate at levels most people will never encounter, and a jury decided he deserved compensation. That outcome is not evidence that the oatmeal in your pantry is dangerous. But it is a reminder that the dose, the context, and the person doing the exposing all matter enormously. Glyphosate is a chemical tool with benefits, trade-offs, and genuine open scientific questions. Learning to hold that complexity without collapsing into panic or dismissal is, honestly, one of the most useful and hardest things to do in navigating modern health science. Love, YLE Your Local Epidemiologist (YLE) is founded and operated by Dr. Katelyn Jetelina, MPH PhD—an epidemiologist, wife, and mom of two little girls. Hannah Totte, MPH, is an epidemiologist and YLE Community Manager. YLE reaches more than 425,000 people in over 132 countries with one goal: “Translate” the ever-evolving public health science so that people will be well-equipped to make evidence-based decisions. This newsletter is free to everyone, thanks to the generous support of fellow YLE community members. To support the effort, subscribe or upgrade below: Subscribe
Optical Vortex Phase Masks for the Detection of Habitable Worlds
A team of NASA researchers is developing new types of optical masks that could help enable the many orders of magnitude of starlight suppression needed for future space observatories to pick out very faint habitable exoplanets from the far brighter glare of their stellar hosts.
Uniting to respond to mental health impacts of digital violence
Technology-facilitated gender-based violence (TFGBV), from doxxing and cyberstalking to cyberbullying and online harassment, is reshaping how women and girls navigate digital spaces. Globally, an estimated 23% of women have experienced digital violence. In Indonesia, reported cases of TFGBV reached 1,791 in 2024, a 40.8% increase from the previous year. Yet very few survivors seek help, leaving lasting effects on their mental health and wellbeing. As part of the multi-UN agencies UNiTE campaign for the 16 Days of Activism against Gender-Based Violence, the World Health Organization (WHO) and the Ministry of Health (MoH) co-hosted a talk show on 6 December 2025 to spotlight this urgent issue. The event brought together voices from Meta, the Indonesian Clinical Psychologist Association and Dr Vivi Syarif, a WHO Fides content creator, creating a cross-sectoral dialogue on the magnitude and mental health impact of digital violence. The talk show introduced psychological first aid as a vital initial intervention that communities can provide to survivors. This intervention aligns with the Psychological First Aid: Guide for Field Workers, a framework that has been standardized for over a decade. The "look, listen, link" approach focuses on ensuring safety and urgent needs, reassuring survivors and connecting them to appropriate services. In addition to this public health approach, Meta also outlined mechanisms to respond to digital violence and reaffirmed its commitment to building safer digital spaces by continuously updating community guidelines based on user feedback. Such collaborative discussions help connect community engagement initiatives with daily digital life. They empower audiences not only to understand how the health sector and industry are responding, but also to educate others, promote first-line support and encourage referrals to appropriate help. By fostering awareness and engagement, these dialogues strengthen collective resilience against digital violence. Written by Jessie O. Yunus, National Professional Officer (Gender, Equity, Rights and Disability Inclusion), WHO Indonesia
The biggest accelerant of our political tribalism
In years past, my go-to word in describing our political condition was “polarization.” I used it endlessly and still do on occasion. But as time has moved on, and the divide has only grown in depth and toxicity, polarization no longer seems to fully capture the magnitude of what has overtaken us. In its place, […]
Astronomers aim to take ‘revolutionary’ moving image of black hole
Newly appointed Cambridge professor says feat would accelerate scientific knowledge by an order of magnitude
Something hidden deep underground supercharged this Chile earthquake
A powerful 7.4-magnitude earthquake struck northern Chile in July 2024—and it wasn’t supposed to be that strong. Unlike Chile’s infamous shallow “megathrust” quakes, this one ruptured deep inside the Earth, where shaking is usually weaker at the surface. Researchers discovered that the quake broke long-held assumptions by tearing through hotter rock layers than expected, fueled by a rare chain reaction that accelerated the rupture.
Beneath the icy face of Alaska, a long theorized ‘connector fault’ may be revealed in the wake of major earthquake
After a magnitude 7 earthquake ripped under Hubbard Glacier earlier this month, scientists believe a “connector fault” may exist between established cuts in the thick planetary crust.
3.0 magnitude earthquake shakes East Bay
A 3.0 magnitude earthquake struck near San Ramon, according to data from the U.S. Geological Survey.