- S&P Global Ratings stress-tested four APAC tech-hardware sectors — foundries, memory, cooling components, and ODMs — against declining AI capex and concluded that foundries (TSMC, SMIC) are the most insulated because leading-edge capacity remains supply-constrained across multiple end markets.
- Memory and cooling suppliers are the most exposed.
TSMC
47 stories mentioning TSMC
- TSMC reported August revenue of NT$514.8 billion (about $16.35 billion), up 53.3% year over year and 10.1% from July — a fourth consecutive monthly increase.
- Year-to-date revenue through August reached NT$3.39 trillion, up 39.3%.
- TrendForce put TSMC's Q2 foundry share at 72.5%, with 5nm, 4nm, and 3nm capacity fully booked;
- At Hot Chips 2026, d-Matrix detailed Raptor, an inference accelerator that bonds a TSMC 4nm compute die face-to-face onto custom DRAM at a 36-micron pitch, delivering roughly 100 TB/s of memory bandwidth per card.
- The company claims about 0.37 pJ/bit — on the order of one-sixth to one-tenth the energy per bit of HBM3.
- Nvidia announced full production of the Groq 3 LPX rack, commercializing technology from its $20B December acquisition of Groq assets — its largest deal on record.
- Each rack packages 256 Groq 3 chips and is claimed to deliver up to 3,400 tokens per second on an Artificial Analysis benchmark, deployed alongside Vera CPUs and Rubin GPUs at neocloud provider Nebius starting later this year.
- Xiaomi unveiled three TSMC-fabricated Xring chips: the 3nm Xring O3 flagship phone SoC already in mass production, the 6nm Xring O100 NPU designed to run its MiMo LLM on-device, and the 3nm Xring D100 for autonomous driving.
- The company says the O3 broke five million on AnTuTu.
- Xiaomi framed the launch explicitly as reducing dependence on Qualcomm and MediaTek — another data point in the vertical-integration trend running through consumer AI silicon.
- Samsung lifted prices on new 4nm, 5nm and 8nm foundry orders placed in July, with increases reaching 15% and Chinese customers absorbing the steepest hikes.
- Its 4nm Pyeongtaek lines are reported at full capacity as AI demand spills past TSMC's constrained allocation.
- Expect cost pass-through into accelerators, networking silicon and devices over the next two to three quarters.
- Microsoft is reported to be in talks with TSMC to produce more than 300,000 Maia 300 AI accelerators for 2027 delivery, with the chip expected to be unveiled this fall.
- Maia 200 is already deployed in Azure data centers while Maia 300 remains in design.
- Microsoft hopes large Azure customers — Anthropic among them — will adopt the in-house silicon.
- Microsoft is planning to significantly increase production of its next-generation Maia 300 chip, with a public unveiling potentially as soon as next month.
- The company is in talks with TSMC to secure capacity for over 300,000 chips for 2027 delivery — an order of magnitude above the tens of thousands of Maia 200 chips produced so far.
- TSMC reported July revenue of NT$467.58 billion (about $14.5 billion), up 44.7% year on year and running ahead of its own raised full-year guidance of slightly above 40% growth in dollar terms.
- The company has lifted 2026 capital expenditure guidance to $60–64 billion, and high-performance computing — where AI chip revenue is booked — accounted for 66% of second-quarter revenue.
- Executive Summary Monday’s cycle was defined by two forces pulling in opposite directions.
- Capital is flooding into AI silicon — Intel raised $15B in equity, TSMC posted 45% YoY revenue growth, and Microsoft is quietly planning a 10× production ramp of its next-gen Maia chip.
- Simultaneously, Washington shifted from rhetoric to demands for accountability: House Democrats want the CEOs of OpenAI and Anthropic under oath, a new FLI safety index gave no lab better than a C+, and TechCrunch published a deep analysis arguing safety testing itself has become a systemic risk.
- The Information reported that TSMC is developing advanced chip-packaging technology similar to Intel's offering.
- The effort matters because packaging has become a strategic bottleneck for AI accelerators, where memory, interconnect, and manufacturing capacity increasingly determine who can deliver frontier-scale compute.
- TSMC said it is gradually resuming operations at its Kumamoto, Japan, plant after a magnitude 7 earthquake.
- The company said building structures were safe and employees were evacuated, while inspections continued to determine possible wafer or equipment impact.
- The incident underscores that AI chip supply chains remain exposed to physical disruption as well as geopolitics and export controls.
- A supply-chain roundup detailed OpenAI’s device plans: a screenless smart speaker co-designed with Jony Ive (early 2027, around $200–$300, running ChatGPT/GPT-Live) and an “AI agent phone” targeting mass production in the first half of 2027 (MediaTek Dimensity 9600, TSMC N2P, Luxshare).
- The effort stems from OpenAI’s $6.5B acquisition of io and 400-plus ex-Apple hires — the same hiring that underpins Apple’s July 10 trade-secrets lawsuit.
- TSMC commits an additional $100B, lifting its total Arizona pipeline to $265B.
- Raised full-year capex guidance to $60–64B and framed 2-nanometer as the next revenue driver.
- Despite U.S. fab costs running 4–5× higher than in Taiwan, TSMC intends not to "leave any food on the table" as AI demand compounds — a hard signal the buildout is structural, not cyclical.
- TSMC CFO Wendell Huang told CNBC the company is speeding up capacity expansion at its Arizona fabs because of robust AI chip demand.
- The acceleration deepens U.S. advanced-manufacturing capacity and shortens supply for leading-edge accelerators.
- It also reinforces TSMC's pricing power as the near-sole-source foundry for most frontier AI silicon.
- WSJ reports that TSMC plans a further $100 billion U.S. investment after AI demand helped drive a surge in earnings.
- The move would deepen the geographic redistribution of advanced semiconductor capacity and reflects sustained customer demand for AI compute.
- For executives, the important signal is that the AI infrastructure cycle is still extending into multi-year manufacturing commitments despite questions about near-term AI returns.
- TSMC reported record second-quarter revenue of about $40.2 billion and net profit up 77.4% year over year, with a 67.7% gross margin and a raised full-year outlook.
- As the leading-edge foundry for Nvidia, AMD, and Apple silicon, TSMC's results remain one of the cleanest signals on whether AI capex is real.
- ASML raised its full-year guidance for the second time this year and beat quarterly estimates, now guiding to 43-45 billion euros in 2026 sales at 54-56% gross margins.
- CEO Christophe Fouquet called first-half orders extremely strong as customers like TSMC accelerate capacity expansion for AI chips.
- As the sole EUV lithography supplier, ASML's upgrade is one of the cleanest reads on sustained AI-infrastructure demand.
- Meta will scale Hyperion to 5 GW at >$50B — up from initial ~$10B/2 GW.
- Partner Entergy will build ~7 GW of new generation.
- Aggregate Louisiana AI commitments now exceed $250B.
- The physical layer, not model architecture, remains the binding constraint.
Internal documents show Meta plans to begin manufacturing its custom data-center accelerator, codenamed Iris, in September as part of a four-generation MTIA roadmap scaling toward 14 GW of compute by 2027. Built with Broadcom and TSMC, it reportedly passed testing in six weeks — Meta’s most aggressive push yet to reduce reliance on Nvidia and AMD GPUs.
- Today's cycle is about the economics of AI rather than new model launches.
- TSMC posted record quarterly revenue and Intel committed €5B to expand European fab capacity, even as the Associated Press flagged that roughly $700B in 2026 data‑center spend has become a measurable inflation risk feeding into the Fed's rate path.
- TSMC reported second‑quarter revenue of NT$1.27 trillion (~$39.6B), up 36% year‑over‑year and above its own guidance, with June sales alone up 67.9%.
- Analysts at SemiAnalysis note the foundry is "sold out" on its N3 node — targeted by essentially every leading AI GPU and CPU this year — and estimate TSMC is tracking to over $40B in AI‑chip revenue in 2026.
- The last 24 hours were defined by capital and governance rather than model launches .
- Four separate multi-billion-dollar infrastructure commitments — from Meta, Intel, Samsung, and TSMC — landed inside a single day, reinforcing that the durable economics of the AI build-out still sit in silicon, memory, and advanced packaging rather than the model layer.
- Three new CoWoS fabs at Chiayi Science Park — the bottleneck step gating AI-accelerator supply.
- Once both phases run: >$9.35B annual output.
- TSMC also reported record Q2 revenue of ~$39.62B (up 36% YoY).
- Samsung Group plans to announce a decade-long investment of about 1,000 trillion won (~$647.5 billion) in South Korea, including a potential 300 trillion won for chip factories in the country’s southwest, plus AI data centers, batteries, and displays.
- The plan, to be unveiled at a meeting with President Lee Jae Myung, aims to harness surging AI-driven memory demand to drive national economic growth.
- IBM said its Albany research lab produced the first chip technology to operate below 1 nanometer — a 0.7nm (7-angstrom) transistor built in three dimensions via a "nanostack" architecture, packing ~100 billion transistors onto a fingernail-sized area.
- Though IBM exited chip manufacturing years ago, the result positions its research ahead of TSMC, Samsung and Intel on transistor scaling and points to a longer runway for compute density as AI demand surges.
Google is in talks with Samsung as a third foundry partner alongside TSMC and Intel. The multi-foundry approach reflects supply chain risk management and could become the template for hyperscaler chip procurement.
Alphabet tapped Intel to manufacture three million in-house AI chips — a significant win for Intel's foundry ambitions and Google's effort to diversify its chip supply chain beyond TSMC. Validates Intel's 18A process as production-ready for a major hyperscaler customer.
TSMC warned AI chip demand is straining the entire supply chain — "we can only support so much" — and confirmed it would like to raise prices. For infrastructure planners, the warning signals chip allocation challenges and cost increases through 2027.
Cerebras Systems' post-IPO rally — shares surged 68% on debut earlier this month — continues to draw institutional flows, with ARK adding to its position. Separately, CEO Andrew Feldman warned that US chip manufacturing catch-up versus TSMC could take up to 15 years, framing his pitch for domestic AI silicon.
The newsletter corpus treats NVIDIA GTC Taipei 2026 as a high-signal infrastructure event: NVIDIA's first GTC Taipei conference, focused on accelerated computing, sovereign AI infrastructure, robotics simulation, Blackwell Ultra production systems, Rubin roadmap previews, and Taiwan-centered AI factory partnerships. The event reinforced a core corpus theme: frontier AI competition is constrained not only by models, but by GPUs, networking, manufacturing ecosystems, and regional cloud capacity.
- The Information’s AM coverage highlighted Huawei’s efforts to narrow the chip gap with TSMC despite U.S. sanctions.
- The Cowork newsletter framed the development alongside Jensen Huang’s comments about China and DeepSeek’s price cuts, underscoring how compute access, export controls, and model pricing are converging into one strategic issue.
- Computex 2026 appears as an additional high-signal hardware/platform event in the corpus, especially because it anchors NVIDIA's post-Blackwell roadmap in Taiwan's manufacturing ecosystem.
- The May 23 digest says Jensen Huang used Computex in Taipei to unveil the Vera Rubin AI superchip platform, SpectraLink photonic networking for rack-scale AI clusters, and a Jetson Thor robotics developer kit.
Bloomberg's Odd Lots podcast featured Cerebras CEO Andrew Feldman discussing the company's wafer-scale chip design (~58× the size of a standard GPU), competitive positioning against Nvidia, the TSMC manufacturing relationship, and the open- vs. closed-source model debate — all in the week of Cerebras' record tech IPO. A useful deep-dive on the hardware architecture bets underpinning the AI infrastructure race.
- Apple signed a preliminary manufacturing agreement with Intel for US-based chip production, responding to White House pressure to reduce dependency on TSMC amid geopolitical risk.
- Intel's stock has risen 240% year-to-date on surging Xeon CPU demand for AI inference workloads, as the hardware focus shifts from GPU training to CPU-driven inference at scale.
- OpenAI's $18 billion custom AI chip initiative with Broadcom — code-named Project Nexus — has hit a major financing wall.
- Broadcom will not commit production capacity at TSMC unless Microsoft agrees to purchase approximately 40% of the chips upfront, essentially using Microsoft's creditworthiness as a guarantee.
- SpaceX has filed plans for a $55B semiconductor fabrication facility in Texas dubbed "Terafab," positioning the company as a domestic chip manufacturing play alongside its Colossus AI supercomputer.
- The filing comes days after Anthropic secured the entire Colossus 1 cluster (220,000+ NVIDIA GPUs, 300MW) under a long-term compute contract.
- Claude Mythos Security Breach Highlights Dual-Use AI Risks at Frontier Labs The Claude Mythos access incident (detailed in Model Releases above) carries significant policy implications: it is one of the first known cases of unauthorized external access to a classified-as-high-risk pre-release AI system.
- Per Epoch AI data cited in the 2026 AI Index, global AI compute capacity has tripled annually since 2022 and is now 30x its 2021 baseline, with NVIDIA accounting for ~60% of installed compute.
- Amazon and Google rank second and third on the back of their custom silicon stacks.
- The directional read is that the compute build-out has not yet plateaued — and the supply chain still hinges on TSMC.
- The 2026 Stanford AI Index documents that global AI compute capacity has grown 30-fold since 2021, at a compounding rate of 3.3× annually.
- The U.S. hosts 5,427 data centers — more than 10× any other country — with a single foundry (TSMC) fabricating almost all leading chips.
- Training carbon costs have reached alarming levels: training xAI's Grok 4 generates an estimated 72,000–140,000 tons of CO₂-equivalent.
Cerebras Targeting April IPO at $22–25B Valuation AI chip startup Cerebras Systems is targeting an April 2026 IPO at a valuation of $22–25 billion, aiming to raise approximately $2 billion in what would be one of the largest AI hardware public offerings since Nvidia's rise. Cerebras's wafer-scale engine architecture offers an alternative inference paradigm to GPU clusters, and the company has been gaining enterprise traction among organizations seeking lower-latency inference at scale.
The corpus previews GTC Taipei as a delivery-story event: N1X ARM-based laptop SoC, Vera Rubin NVL72 production progress, partner assets, and Taiwan's AI supply-chain role. - NVIDIA's official COMPUTEX/GTC Taipei page highlights Jensen Huang's keynote, expert sessions, training, demo showcase, AI Factory MGX ecosystem, and OpenClaw/NemoClaw Build-a-Claw demos.
Nemotron 3 Nano Omni: Covered as a unified multimodal reasoning model released at GTC. - OpenClaw and NemoClaw: The corpus links NVIDIA's GTC narrative to cross-vendor agent runtime work and safer agents that run locally, in cloud VMs, and at the edge. - SAP partnership: Several entries describe enterprise agent runtime collaboration with SAP.
- NVIDIA's GTC cycle appears repeatedly in the corpus as the infrastructure counterweight to software-centric AI events.
- The March GTC narrative centered on agentic AI, physical AI, robotics, Nemotron models, Vera Rubin systems, NVLink Fusion, and AI factory economics.
- GTC Taipei, scheduled for June 1–4 at the Taipei International Convention Center, extends that story into Taiwan's semiconductor and manufacturing ecosystem, with the corpus highlighting a Jensen Huang keynote, N1X ARM laptop SoC expectations, Vera Rubin delivery updates, and OpenClaw/NemoClaw agent demos.
GTC 2026 is consistently framed as NVIDIA's pivot from model acceleration to embodied AI: robotics, simulation, factory autonomy, autonomous workloads, and GR00T/humanoid foundation-model updates. - Later corpus entries connect GTC's physical-AI narrative to NVIDIA Research's ICRA robotics papers and to Jetson Thor edge robotics.
AI factory lock-in: NVIDIA is positioning the rack, network, software runtime, and agent safety layer as one integrated system. - Physical AI as growth vector: Robotics and embodied autonomy become the next demand driver after LLM training and inference. - Taiwan as strategic center: GTC Taipei ties NVIDIA's platform roadmap to the manufacturing base that makes accelerated computing possible. - AI PCs and edge expansion: N1X, Jetson Thor, and Alpamayo-style AI PC references show NVIDIA expanding beyond data centers.
The corpus describes Vera Rubin as NVIDIA's next-generation AI factory platform, with Rubin GPUs, Vera CPUs, NVLink 6, HBM4-class memory, and NVL72 rack-scale deployment. - Reported metrics include sharply higher FP4 inference throughput, improved performance per watt, and a claimed 10x reduction in inference cost per token versus Blackwell-era systems. - Hyperscaler demand is a recurring theme, with AWS, Azure, Google Cloud, and Oracle described as preparing or evaluating large-scale deployments.