Hello world, its Thursday, November 19th, 2026
Intel Foundry spent years promising outside customers, and on Wednesday it named one and signed another. Alibaba admitted Washington’s new rule will cut its access to overseas compute, Micron said the memory squeeze tightens from here, and Anthropic raised $107 billion to pay its chip bills. Meanwhile, OpenAI’s models are quietly editing AMD’s GPU blueprints.
Let’s get into it. --Austin & Vik
Latest episode: Google Built Two New TPUs Without New Silicon: The 8i Max and 8TX — listen on YouTube, Apple Podcasts, or Spotify. Full transcript: daily.semidoped.com.
Presented by Crusoe
Fine-tune and serve open models without managing a single GPU.
When we sat down with Crusoe’s product team, Austin asked: what will you do with my data? Their answer: They don’t touch or see it.
Your training data is encrypted so even Crusoe’s own admins can’t see it, and the fine-tuned model is yours. Deploy on Crusoe in one click, or download and run it anywhere. Upload a dataset, pick an open model, and pay per token rather than per GPU hour. Run a few versions, compare them, and download the one that works best.
Accounts are free, and get $5 in credit when you sign up.
Intel Foundry lands Microsoft and an 18A customer
Microsoft will package its future Azure chips at Intel. The two said on Wednesday that Microsoft is the first company to announce use of Intel Foundry’s EMIB packaging for its first-party Azure silicon, under a long-term deal that also locks in multiyear Xeon supply, including next-generation Diamond Rapids.
Separately, Lip-Bu Tan told the TomorrowX Summit in Austin that Intel “just signed the first customer” for 18A, without naming it. CTO Pushkar Ranade said the customer is on 18A-P, the improved version that entered risk production in June.
EMIB gives hyperscalers a second route for packaging big AI chips while TSMC’s CoWoS capacity stays stretched. A leading-edge wafer customer is the harder prize, and Intel has chased one for five years. Socionext already announced 18A-P design work in August, and Intel didn’t say how the new deal differs.
Source: Intel · Microsoft · CNBC · The Oregonian · The Next Web
China’s cloud giants count the cost of the remote-access rule
Washington’s remote-access rule took effect on Thursday with its publication in the Federal Register, and China’s biggest clouds are already writing it into their results. Alibaba said the rule “will impact our company’s ability to access compute from certain providers outside China,” and that it can’t yet estimate the financial hit.
Alibaba is still spending. September-quarter capex rose 60% to 50.5 billion yuan, partly on higher chip-component prices, and AI cloud revenue grew 53%. Tencent said on Wednesday the curbs will likely slow its next model, though its inference-heavy cloud can run on domestic chips.
The pain is spreading. Japan’s Datasection, which press reports have tied to Tencent, cut its revenue forecast 89% this week after a large AI infrastructure order was cancelled. Chinese GPU designers including Hygon and Cambricon slid on Thursday on rumors that overseas foundries will review Chinese orders. Trump, for his part, told Fox News that Blackwell sales to China could come “in two years or three years.”
Source: Alibaba results (HKEX) · Federal Register · Reuters via Investing.com · Datasection filing · Investing.com · Yonhap Infomax · The Straits Times
Micron says the memory squeeze tightens from here
Micron CEO Sanjay Mehrotra told the TomorrowX Summit on Wednesday that 75% of the company’s 2027 production is already sold, and customers are now negotiating 2028. He expects 2027 and 2028 to be even tighter than 2026.
The contracts are the story. Micron has signed 26 supply and collaboration agreements, many running five years or longer, backed by $32 billion of upfront cash and financial commitments, take-or-pay terms and pricing bands. Mehrotra argued customers won’t walk away because memory now sets the economics of AI factories.
Next comes stacked 3D DRAM. Micron says every major customer is asking to partner on it, with market penetration starting in 2028 to 2029.
Source: Micron at TomorrowX (Investing.com) · Transcript (StockAnalysis)
Anthropic raises $107 billion to pay for compute
Anthropic priced its IPO at $83 a share on Wednesday, above its $70 to $78 range, in a 980 million-share offering worth $81.3 billion, the largest stock-market debut on record. A $26 billion private placement at the same price takes the total to $107.3 billion. Trading starts on Nasdaq on Thursday under ANT.
That money already has a destination. Anthropic has committed to nearly $900 billion of compute spending with partners including Amazon and Google, Reuters reports. Compute costs rose another $5 billion in the third quarter, tipping it to a $567 million operating loss.
The filing also names the soft spot in the buildout. Its compute providers lean on third-party financing to buy chips, and if that dries up, contracted capacity “may be delayed, repriced, reduced, or not delivered.”
Source: Anthropic · Reuters via Investing.com · NPR
OpenAI’s models are redesigning AMD’s GPUs
OpenAI said on Wednesday its models rewrote the SystemVerilog code behind AMD’s flagship GPU IP and found up to 14% gains in power, performance and area over heavily optimized baselines. AMD has already put some of the changes into a design it sent to manufacturing, and added others to production files for future chips.
OpenAI calls it roughly a year of progress in under five months. AMD also verified single-digit PPA gains on key blocks of its next flagship CPU, and the same approach cut standard-cell area by about 13% in the vector unit of OpenAI’s own Jalapeño chip.
Every rewrite was checked with Synopsys tools, and OpenAI is now training a model with Synopsys to drive EDA software like an expert designer. “Up to” is doing some work in that headline number, since OpenAI gave no block-by-block breakdown.
Supported by Meter
You have the GPUs and the power. Now who builds the network, and who do you call when it breaks?
Meter designs, deploys, and runs frontend, storage, and management networks on its own switches and its own network OS.
One call reaches the team that designed the hardware and wrote the software.
Industry Chatter
These are just unconfirmed industry whispers.
Nvidia’s next CEO — Heidrick & Struggles, the search firm hired by Nvidia’s board, has approached Arm CEO Rene Haas about the job (Bloomberg Law · Business Standard).
Intel 18A yields — Intel’s 18A yields have climbed to about 70%, a sharp turn from the “horrible” yields Lip-Bu Tan inherited (TrendForce).
SK hynix hybrid bonding — SK hynix has installed hybrid bonders from Hanwha Semitech and Besi–Applied Materials at Cheongju to test 20-high HBM5 stacks (Munhwa Ilbo).
Terafab tools — Musk’s Terafab told Korean toolmaker GST it needs thousands of scrubbers and chillers and asked how many it can deliver (The Elec).
Anthropic and SpaceX — Elon Musk says Anthropic’s most likely plan is to rent another 10 gigawatts of compute from SpaceX (Benzinga · investingLive).
Sector Watch
Earnings
Kulicke & Soffa — Fiscal fourth-quarter revenue more than doubled to $410.3 million, and the bonder maker guided flat for the December quarter. (Source: Kulicke & Soffa)
Soitec — The wafer maker signed about $1.2 billion of deposit-backed photonics-SOI capacity deals and expects that business to more than triple this year. (Source: Soitec · Tech Times)
Lenovo — Revenue jumped 48% to a record $30.2 billion on AI demand, but non-cash charges left a $640 million net loss. (Source: Lenovo · Bloomberg)
Memory, packaging and fabs
JEDEC — The standards body previewed DDR6, with server modules starting at 12.8 GT/s and over 50% more bandwidth than DDR5. (Source: JEDEC · TrendForce)
ASE — ASE’s Taiwan Foredge unit broke ground on a NT$248.9 billion AI chip test plant in Kaohsiung, due in 2028 and 2029. (Source: Digitimes · Liberty Times)
Rapidus — Japan’s government is weighing up to ¥54 billion of tax relief when it hands the Hokkaido fab over to Rapidus. (Source: Jiji Press · The Japan Times)
Optics and AI cloud
Eoptolink — The Chinese optical module maker denied order-cut rumors, saying its 1.6T modules are in volume production and its backlog is ample. (Source: Reuters via TradingView · Zhitong via Investing.com)
Nscale — The neocloud brought its first Vera Rubin NVL72 racks online on Dell systems, with a wider rollout planned over coming months. (Source: Nscale · Dell)
Also worth a read
Austin’s latest on Chipstrat, Intel Custom ASICs?
Vik’s latest on Vik’s Newsletter, HBM: Lower Die-Count Stacks and Custom Base-Die Performance
Tell us how to make Semi Doped better! Give us your feedback
Want to become a sponsor? Send an email to sponsor@semidoped.com

