A bunch of AI-related news has popped up this week, so let’s do a roundup.
Some AI companies are complaining that TSMC is killing the AI boom by not expanding rapidly enough:
Asianometry notes that TSMC’s caution at expanding is amply justified by the boom-and-bust nature of the semiconductor industry:
“I’m hearing many similar views in the Silicon Valley Borg that TSMC is the break or limiter on the AI boom, as if they’re the reason why we don’t have AGI yet. Because they didn’t and still don’t believe.”
“If we can ever say that a company that spent $41 billion on capital expenditure in 2025, with another $53 to $56 billion in 2026 planned, is sitting on its hands, doing nothing.”
“TSMC having 90% share of the AI chip market looks pretty unhealthy. That should go down and it will. Samsung seems to be doing well so far.”
“The cold, hard reality is that shortages are a fact of life in semiconductors, as are horrific gluts.”
“What we are flippantly labeling as TSMC we really mean is the AI supply chain. And that supply chain is as complicated as you can possibly imagine. Like an iceberg, it looks big enough on the surface of the water, but goes way far deeper underneath. TSMC has thousands of suppliers in two categories: Equipment like the famed ASML lithography tools and materials like photoresist, silicon wafers, acid etch gases and so on. These are not generalized tools and materials. They are not fungeible like AWS compute units.”
“And then there are the memory guys. You cannot ship an AI system without memory. DRAM and NAND. Nvidia’s AI chips use a special form of DRAM called high bandwidth memory, and they use quite a lot of it. The memory industry is just as consolidated as the logic industry, with the major players being Samsung, SK Hynix and Micron.”
“The chip guys are last to know when the party is getting started, but first they get batoned in the face when the police shut things down.”
He points out that semiconductor manufacturers have log supply chains. He uses a different metaphor (the beer distribution game, or a bullwhip), but back when I was working at Applied Materials, it was described as trains linked together with slinkys. First software takes off, then hardware gets yanked along, then the chip manufacturers get yanked, and then, finally, semiconductor equipment manufacturers get yanked into motion, and shortly after that happens, the bust hits the front of the train, and the trailing cars all crash into each other. It’s a regular boom/bust cycle.
“From 1961 to 2006, electronics consumption in the United States grew positively but with wild volatility swings between 0 to 20%. But for the semiconductor makers, that translates to swings anywhere from 20% to 40%. And for the equipment makers, it is amplified even more, plus or minus 60%. The whip hits particularly hard in the semiconductor industry because of the industry’s long lead times. It takes 4.5 months to fabricate and package a chip. It takes 18 months to 2 years to build a fab. Meaning from shovels down to producing chips, and it takes 12 to 18 months to produce and install something like an EUV machine into the fab. Another 6 months before that machine actually starts patterning wafers.”
“Long lead times mean having to make very long demand forecasts, which leads to extreme volatility swings during up and downturns even if those up or downturns are relatively small.” People forget that in 1998, during the time we now think of as the DotCom Boom, there was a small semiconductor downturn that had Applied Materials forcing employees to take unpaid leave.
“ASML just reported 2025 earnings, and we see the bullwhip in full effect. TSMC raised capital expenditure 35% but ASML announced €13.2 billion of net new bookings. Analysts had expected just €6.32 billion. This is because ASML collected orders not just from TSMC, but also Samsung, Intel and the memory guys. When it rains it pours, right? Again, this is why I fear that another AI foundry would not mean our compute shortage is solved, because ultimately, when those foundries start scaling their capacity, they all go to the same suppliers.”
He goes over how car manufacturers cancelled orders during Flu Manchu, and then scrambled when the economy took off afterwards. “TSMC was trying to discern between double booked orders and real demand, which is not an uncommon experience for them. Customers lie about their own demand all the time, or at least we can say that they are eternally optimistic. TSMC tried to respond in 2022. The Taiwanese giant poured $36 billion into capital expenditure. They went to their suppliers and pushed like no tomorrow.”
“It turned out those customers really were double booking orders and artificially inflating demand. When the macro environment turned in 2022, the automotive, smartphone, and PC chips that were so hot during the COVID era fell out of vogue and customers started cutting orders.”
“Meanwhile, deeper down in the supply chain, TSMC and the rest of the semiconductor industry were getting bullwhipped by COVID hangover. Utilization at TSMC’s multi-billion dollar N7 fabs crashed, Semi analysis wrote in April 2023. Now, Semi analysis data indicates that the 7nm utilization rates were below 70% in Q1. Furthermore, Q2 gets even worse with 7nm utilization rates falling to below 60%. This is primarily due to weakness in both smartphones and PCs, but there is a broader weakness in most segments. A fab’s break even utilization rates are about 60% to 70%. So those N7 Taichung fabs were taking financial losses potentially on the order of hundreds of millions, maybe even billions. The financial burdens of low utilization are another reason why I’m skeptical another AI foundry could have rushed into the AI chip fray to save the day.”
He says that Intel incurred losses during this period due to an unnecessary fab expansion, which is probably true, but that was a secondary factor next to their longer running problem of getting their process wrong.
“ChatGPT was released in November 2022, and that kicked off a massive increase in capex amongst the hyperscalers in particular, but it sure seems like TSMC didn’t buy the hype. That lack of increased investment earlier this decade is why there is a shortage today and is why TSMC has been a de facto break on the AI buildout/bubble.”
“I recall news in mid 2024 of TSMC struggling with CoWoS capacity bottlenecks and yield problems, including one design issue that caused cracks in the Nvidia chips packaging.” CoWoS is Chip on Wafer on Substrate, which involves fabbing an interposer as a substrate for faster connections between your processing chips and memory.
“I also recall news in late 2024 noting how the vendors in charge of making the server racks for Nvidia’s Blackwell servers struggled with overheating, liquid cooling leaks, software bugs, and connectivity issues. Such technical difficulties delayed server deployment until early to mid 2025, creating a weird situation for several months where TSMC was pumping out chips that just went into storage. So that gated things, because you don’t scale until you first fix the technical problems.”
Then there’s the power-scaling issue, which is a whole ‘nuther can of worms.
There’s a lot of talk about a SaaSpocalypse going on thanks to a new AI tool. (SaaS is “Software as a Service.” Instead of hosting your own payroll or sales-tracking or whatever servers, you hire a company that already has cloud software setup to do it and you just tie into that, which can considerably reduce startup costs. A whole lot of successful new tech companies over the last decade plus have been SaaS companies.)
The software sector was jolted overnight with what analysts are calling a “SaaSpocalypse” — a sudden and severe selloff triggered by new artificial intelligence tools unveiled by US AI startup Anthropic. The episode has sharpened investor fears that AI is no longer merely helping software companies but may now begin replacing them.
Anthropic has expanded its enterprise AI platform, Claude Cowork, by launching 11 new plugins aimed at automating a wide range of professional tasks. Claude Cowork is an agentic, no-code AI assistant built for corporate users, allowing companies to automate workflows without writing software. The new plugins are designed to handle tasks across legal, sales, marketing and data analysis functions. The most recent addition is Anthropic’s Claude Legal agent, which can perform routine legal work such as document and contract review, and compliance checks.
Anthropic has said that the tool does not provide legal advice and that all AI-generated outputs must be reviewed by licensed attorneys. Even so, the breadth of automation signals a step change in how much white-collar work AI systems can now perform.
Productivity — Manage tasks, calendars, daily workflows, and personal context
Enterprise search — Find information across your company’s tools and docs
Plugin Create/Customize — Create and customize new plugins from scratch
Sales — Research prospects, prep deals, and follow your sales process
Finance — Analyze financials, build models, and track key metrics
Data — Query, visualize, and interpret datasets
Legal — Review documents, flag risks, and track compliance
Marketing — Draft content, plan campaigns, and manage launches
Customer support — Triage issues, draft responses, and surface solutions
Product management — Write specs, prioritize roadmaps, and track progress
Biology research — Search literature, analyze results, and plan experiments
A lot of those are already automated elsewhere, but I suspect a lot accountants and paralegals just felt a goose strut across their grave. On the other hand, who is really going to turn over, say, Accounts Payable to an AI? One glitch, and your entire bank account is drained…
If it works (a big if, give so many AIs are prone to hallucinations), this is potentially good news for Anthropic and the companies using their tools, and bad for SaaS companies and the employees currently doing those jobs.
I note there’s no plugin for technical writing…yet.
And Google Cloud ended 2025 at an annual run rate of over $70 billion, representing a wide breadth of customers, driven by demand for AI products.
We’re seeing our AI investments and infrastructure drive revenue and growth across the board. To meet customer demand and capitalize on the growing opportunities we have ahead of us, our 2026 CapEx investments are anticipated to be in the range of $175 to $185 billion.”
Remember how Nvidia was going to invest $100 billion in OpenAI? Yeah, not so much.
In September 2025, Nvidia and OpenAI announced a letter of intent for Nvidia to invest up to $100 billion in OpenAI’s AI infrastructure. At the time, the companies said they expected to finalize details “in the coming weeks.” Five months later, no deal has closed, Nvidia’s CEO now says the $100 billion figure was “never a commitment,” and Reuters reports that OpenAI has been quietly seeking alternatives to Nvidia chips since last year.
Reuters also wrote that OpenAI is unsatisfied with the speed of some Nvidia chips for inference tasks, citing eight sources familiar with the matter. Inference is the process by which a trained AI model generates responses to user queries. According to the report, the issue became apparent in OpenAI’s Codex, an AI code-generation tool. OpenAI staff reportedly attributed some of Codex’s performance limitations to Nvidia’s GPU-based hardware.
After the Reuters story published and Nvidia’s stock price took a dive, Nvidia and OpenAI have tried to smooth things over publicly. OpenAI CEO Sam Altman posted on X: “We love working with NVIDIA and they make the best AI chips in the world. We hope to be a gigantic customer for a very long time. I don’t get where all this insanity is coming from.”
Microsoft’s Copilot chatbot has become central to its artificial-intelligence strategy as the company’s close partnership with OpenAI diminishes. But the effort to build it up as a ChatGPT alternative has been tough going.
Confusing brand positioning and interoperability problems have frustrated users, current and former employees who have worked on Microsoft’s AI products said.
Interoperability problems? With a Microsoft product?
Only a small proportion of subscribers to Microsoft’s enterprise suite use Copilot, and the percentage who favor it over Google’s Gemini or other tools has decreased in recent months, according to data reviewed by the Journal.
The stakes are high for Microsoft because Copilot is core to a push by Chief Executive Satya Nadella to transform Microsoft into an AI-first company, much as he transformed it into a cloud-first company around a decade ago. Copilot is one of Nadella’s top priorities, current and former executives said.
Microsoft shares tumbled after its earnings report last week sparked investor concern that growth in its most important unit, the Azure cloud-computing business, is slowing, and that its AI business is reliant on OpenAI while Copilot remains unproven. Shares fell nearly 3% Tuesday amid a slide in software stocks prompted by fresh concerns that AI tools will make enterprise subscriptions less necessary.
For other AI companies, we merely suspect they’re evil. For Microsoft (and Google), we already know they’re evil…
Asianometry has an interesting video up about East Germany expensive, strenuous efforts to catch up to the west in semiconductor manufacturing technology.
Spoiler: They didn’t.
Some takeaways:
“In the late 1980s, the German Democratic Republic, or East Germany, went all in on the monumental task of domestic semiconductor production. This semiconductor obsession failed, and the billions of marks spent on it eventually bankrupted the country’s failing economy.” I think he oversells the role the semiconductor push had on bankrupting the economy; everything in late commie East Germany was failing (just like the rest of the Warsaw Pact), they suffered a credit crunch for investment due to tightened western restrictions, couldn’t export Soviet oil as profitably due to the Reagan/Saudi created oil glut, and also were running into hard currency shortages to but the components their manufacturing sector needed to keep exporting.
The East German Uprising of 1953 kicked off what would be a persistent, and ultimately existential problem, for the GDR: Emigration. Throughout its history, its best and smartest people consistently sought a way out to the West. To convince its people to stay, the SED [Sozialistische Einheitspartei Deutschlands, AKA Socialist Unity Party of Germany] promised a better future through the use of technology. More than the Soviets, East Germany leaned on information technology as a pathway towards economic vitality and a glorious socialist future. The Party’s elites saw themselves locked in a technology race with the capitalists to see who can build a better society. Leader Walter Ulbricht called for an “industrial transformation” with the ultimate aim of “catching up with and surpassing capitalism in terms of technology.” A thriving computer industry was crucial towards making this ideology work. And in order to produce these superior computers, East Germany needed to learn and master microelectronics technology.
“Less than four years after the Americans invented the germanium transistor, East Germany moved quickly to build their own line of first generation semiconductors. In 1952, development work began at the VEB Works for Electrical Components for Communications Technology, or WBN, in the town of Teltow near the city of Berlin. This put them about even with West Germany. The FRG’s first semiconductor factory came about in 1952, built by Siemens.” Indeed, this is very early to get into the semiconductor game. It wasn’t until 1957 that Fairchild Semiconductor, widely considered as Company Zero for America’s semiconductor industry, was founded.
“WBN suffered from a lack of cooperation between its industrial and academic sides. The production teams lacked discipline, hands-on experience, and did not appreciate the scale and difficulty of the task they were facing. In one incident, the team dumped hot ashes right outside a factory window where they were producing a pilot run of semiconductors.” Ouch! A very uncleanroom…
“The state failed to give their young semiconductor team the resources it should have gotten. Administration – their chief accountant, in particular – seemed to care very little for semiconductors. When the team asked for money to purchase felt slippers to prevent static charge buildup in the clean room, their chief accountant denied the request.”
The Soviets didn’t help. “Despite being the GDR’s primary political backer, the Soviets were strangely wary. In 1958, two WBN staff members traveled to the Soviet Union to do technical exchanges. A year later, they came back complaining of limited cooperation. Much of what the Soviets had developed was created for military use. Thusly, the Soviets were concerned that transferring that to the East Germans would leak via scientists defecting to the West.”
They tried to get information from the U.S., but Cold War tech transfer policies were already falling into place. They had better luck in the UK. “Through the contacts of Arthur Lewis, a British Labour Party politician, the delegation saw plants owned by British Philips, Siemens-Edison, and British-Thompson-Houston. The latter is a descendant of the Vickers Company that sold oil equipment to the Soviets in the early 1900s. Just thought that was a nice connection. This visit was very successful. The East Germans learned a whole lot about industrial level semiconductor manufacturing. They even managed to purchase equipment for low-frequency transistors, a trailing edge technology.”
Despite that, the gap grew wider: “In 1958, WBN produced 100,000 germanium diodes, transistors, and rectifiers. Worse yet, some 98% of what they produced eventually needed to be discarded throughout their entire working lives.” Classic commie quality. “That same year in 1958, the United States alone produced 27.8 million transistors. Two years later in 1960, the US grew that production capacity five times over to 131 million.”
“Erich Apel, head of the Economic Commission of the Central Committee Politburo and an economic reformer – wrote in late April 1959: ‘Compared to … the American, Japanese, and West German industry, we lie in a state of backwardness that can scarcely be estimated … this backwardness will not decrease through 1961 at least, but will instead grow. Another inspection in 1960 identified more items of backwardness in semiconductor production. Workers tended to use rules of thumb rather than their instruments to measure. The various factory lines did not cooperate with one another.”
“Interestingly, when reporting these results to the Economic Commission of the Central Committee Politburo, that inspector softened his results. In his notes to state authorities, he said the GDR was 5 to 6 years behind. But in his analysis to the more politically charged Economic Commission, he cut it in half, 3 to 4 years.” Commies always institute thermoclines of truth to avoid being purged.
The brain drain to the west continued. The solution: The Berlin Wall. “For semiconductors however, the Wall pinched off what little technology the GDR had imported from the West.” The solution was to suck up even more to the Soviets, and to spy harder.
In 1963, the aging Walter Ulbricht launched a new initiative – called the New Economy System of Planning – to bring more market elements to the GDR economy. Now industrial groups, not bureaucrats, can actually decide how money can be spent. The reform also elevated the status of technology sectors like semiconductor manufacturing in the economy. R&D spending increased by over a third from 1959 to 1963. In 1965, nearly 40% of the electronics that the GDR produced by value were semiconductors – 82 million marks out of 223 million marks in total. Four years later in 1969, that number grew four-fold. Many of these transistors went into new consumer technical goods like radios, TVs and fridges. In 1971, semiconductor production reached 535 million marks by value. That year, East Germany began producing their first integrated circuits, some 10 years after Texas Instruments did it.
“Strange inequalities in policy planning meant that color televisions were widely available, but consumer items like toothbrushes and toilet paper were in short supply.” Communist planning at its finest!
One day in 1967, the Minister of Electrical Engineering and Electronics showed up to an East German electronics firm with a suitcase full of integrated circuits from TI. He told them to copy them exactly. The Ministry for State Security – better known as the Stasi – had been engaged in scientific and technology espionage since the 1950s – mostly related to atomic engineering and other sciences. Then in 1969, the Stasi’s Scientific and Technical Sector was reorganized and expanded with the goal of acquiring military technologies. After Honecker came into power in 1971, the Stasi’s job shifted from acquiring scientific knowledge to specific technologies – mostly via informants in the West who found and handed the goods over to East Germany. One such informant was Hans Rehder, a physicist working for the West German firms Telefunken and AEG. He handed over technical secrets for over 28 years and was never caught.
“Western companies knew about this copying of course. In one famous example, a GDR chip analyst looking at a stolen chip from the US firm Digital Corporation saw a message n Russian, roughly translating to: ‘When do you want to stop to swipe. Own design is better.'”
Stasi intellectual theft kept them from falling further behind, but couldn’t close the gap. “Because the Stasi were spymasters not technical experts, they frequently asked for the wrong item. Their methods of laundering the technology before passing it on made it harder to understand how to use it. Tightening embargoes from the West also interfered with industrial development. Stolen Western products got progressively older and more expensive to acquire. The embargoes gave other countries the chance to scam the Stasi, adding mark-ups frequently in the range of 30% to 80% to even 100%. This drained the East Germans’ already limited R&D budgets.”
“The wholesale copying also undercut the country’s ability to export its goods abroad. The Stasi did not want other countries to see what they had managed to acquire. And had they tried anyway, sales would have been blocked on patent infringement grounds. And finally, semiconductors were getting to the point that East German technicians struggled to replicate them. As early as 1976, an IC’s physical form no longer yielded secrets on how to produce them.”
“In 1981, with the GDR still about 7-10 years behind the West in microelectronics development, Erich Honecker announced a ten-point program to produce the majority of its semiconductors domestically by 1985. The 1970s were rough years for the GDR. Tighter export bans. The Oil Crises of the 1970s. Heavy borrowing from the West. Declining productivity and worsening competitiveness. It was all weighing heavily – grinding the country’s economy to a halt. Gerhard Schürer, head of the State Planning Commission, convinced Honecker that investing in semiconductors would bring the country out of its economic morass.”
They even struck a deal with Toshiba.
In exchange for 25 million marks, Toshiba – a long running technology partner with the GDR – would furnish the GDR with designs for their 256 kilobyte memory chips along with instructions on how to produce them. At the time, 256-kilobyte was leading edge stuff. The GDR was still struggling to produce 64 kilobyte memory. This would have been a game-changer. But in 1987, Toshiba got caught selling submarine propeller equipment to the Soviet Union. Huge scandal back then. Afraid of getting caught again, Toshiba offered the Stasi a 95% refund to destroy the evidence. [Spy Gerhardt] Ronneberger agreed. So in July 1988, he got the money back and dissolved the chip designs in a vat of acid in front of Toshiba’s people. But never trust a spy! Those were just copies, produced for exactly that purpose.
Finally in September 1988, Zeiss General Director Wolfgang Biermann triumphantly presented Erich Honecker with the first samples of that 1 megabit chip – the U61000. Honecker said that the chips were “convincing proof that the GDR is maintaining its position as a developed industrial country.” This technical “triumph” was the bitterest of them all. In semiconductors, prototypes mean nothing. Production means everything. Dresden produced just 35,000 chips throughout the entirety of 1988 and 1989 with a yield of 20%.
To say this was “piss poor” would be an understatement. Those are ruinous, “fire everyone” numbers for actual semiconductor manufacturers.
They planned to scale up to 100,000 1 megabit chips each year. Toshiba alone produced that many in a single day. Two months later in November 1988, the leading edge moved once more. Toshiba began shipping its 4-megabit DRAM in high volume, seeking to produce a million chips a month by March 1989.
Then history happened. “By then, the East German economy was in shambles, scheduled to default on its debts by early 1990. It never even got there. In May 1989, Hungary opened its borders with Austria and East Germans swarmed through there en route to West Germany. Later in November 1989, a year after its one megabit technical triumph, the Berlin Wall fell.”
East Germany stole as many designs as they possibly could, but they couldn’t steal the intellectual expertise behind the numerous process tweaks, nor the furious swarm of technological innovation drive by Silicon Valley’s capitalist high risk/high reword startup culture that drove Moore’s Law for decades.
Top-down communist command economies never had a chance to keep up.
