Foxconn Q3 2026: Record Profit, AI Rack Growth Seen in 2027
Foxconn Q3 2026 net profit rose 34% to a record NT$77.5B on AI servers. It sees high double-digit AI rack shipment growth in 2027 but tight chip supply.

News on the chips, memory and servers that power AI. Recent stories cover Intel's Data Center GPU D-Series for AI inference, Samsung making Nvidia NVHBM base dies on TSMC 3nm, and OpenAI's custom Jalapeño inference chip shown at Hot Chips.
Foxconn Q3 2026 net profit rose 34% to a record NT$77.5B on AI servers. It sees high double-digit AI rack shipment growth in 2027 but tight chip supply.

Google's Gemini glasses from Gentle Monster and Warby Parker start at $395, ship Dec 10, with an LED light and face redaction for bystanders.

SMIC's Q3 2026 revenue rose 33.9% to $3.19B and gross margin hit 29.9%, both above its own guidance. Utilization was 97.3%. Q4 guide: revenue up 4% to 6%.

TSMC reported October 2026 revenue of NT$586.89 billion, up 59.7% from a year earlier. Here are the numbers and how they compare with Q4 guidance.

Financial News, citing October 31 meeting Q&A materials, reported that Sam Altman told Samsung executives, "There has been double-counting of demand."

NVIDIA's 2B-class Alpamayo 2 Edge drives at 160 to 190 ms on DRIVE AGX Thor, and GR00T 2 Turbo runs at 124 ms on a humanoid's Jetson Thor.

Anthropic's IPO roadshow says Claude Mythos ported Claude to AMD MI355X GPUs in about three days. The footnote covers prompt processing for one model.

AMD's 6th Gen EPYC Venice CPUs are shipping: up to 256 cores, 1.6 TB/s memory, PCIe 6.0. Meta, Azure and OCI are first, and AMD's Vera claims are estimates.

Anthropic says Claude Fable 5.2 designed a speakerphone, a voice hub and a virtual pet with Diode, and the boards worked with no circuit fixes.

Google added TPU 8tx for confidential inference and TPU 8i Max, two TPU 8i chips on one package, to its eighth-generation TPU lineup.

Reuters says about 600,000 Nvidia H200 chips sit in Hong Kong as Beijing withholds approval for large-scale imports into mainland China.

Nvidia GPUs will run prefill and Rebellions' R200 NPUs will run decode in SK Telecom racks linked by NVLink Fusion. R200 systems are expected in 2028.

CoreWeave's credit default swaps sit above 1,000 basis points, a level analysts call distress. The debt figures, its reply and what to watch on November 10.

A fire at Panasonic's Hukou plant in Taiwan on November 4 hit a supplier to circuit-board makers used in AI servers. Estimates of the supply impact conflict.

YC startup Subvocal opened $199 refundable preorders for radio-wave glasses that turn silent mouthing into text. The page now says sold out. Shipping is estimated for Q1 2027.

Google says its Suncatcher prototype returned the first AI response from a data center class TPU in orbit. Here is what it carries and what comes next.

Supermicro Q1 FY2027 sales were $19.9B with a 12.7% gross margin, down from 17.5%. Q2 guide is $19.5B to $21.5B; FY2027 rises to $72B to $82B.

Nothing CEO Carl Pei says the first AI-native product ships in December with no new screen. What is confirmed, what is rumor, and what to ask before buying.

Caiwen and TechNode report DeepSeek plans to train V5 on Huawei Ascend 950 chips. Neither report has a DeepSeek or Huawei statement.

AMD's Q3 2026 revenue rose 45% to $13.4 billion as data center sales jumped 97% to $8.5 billion. MI455X began volume shipments, Q4 guidance is about $17 billion, and AMD expects 2027 revenue growth above 80%.
