AI chip news
AI chip news from every vendor: Nvidia, AMD, Intel, Google TPU, AWS Trainium, custom accelerators and startups.

Google TPU 8i Max and TPU 8tx, announced November 9, 2026: what Google published
TPU 8i Max is two TPU 8i chips on one package; TPU 8tx pairs TPU 8t silicon with AMD EPYC hosts.

SambaNova SN50 heads to Supermicro inference racks on November 3, 2026
SambaNova SN50 specs, and Supermicro's November 3, 2026 plan for 10,000 SN50 chips paired with Nvidia GPUs.

AMD Helios racks running at Meta, AMD says on October 27, 2026
AMD said on October 27, 2026 that its first Helios racks are installed and in testing at Meta, ahead of a wider rollout.

Google Cloud A5X with Vera Rubin NVL72 announced October 21, 2026
Google Cloud A5X instances on Vera Rubin NVL72, announced October 21, 2026, reach general availability next month.

vLLM and SGLang on TPU: Google's PyTorch engines, October 20, 2026
Google's new vLLM and SGLang engines run natively on TPU through TorchTPU, launched October 20, 2026.

Ultra Ethernet 1.1 adds scale-up links and collectives
Ultra Ethernet 1.1 adds a local accelerator profile, in-network collectives and new congestion controls.

Microsoft unveils Azure Maia 300 on October 12, 2026: 96-chip racks
Microsoft unveiled Azure Maia 300 on October 12, 2026: 288 GB of HBM4 per chip, 96 chips per 167 kW rack.

NVIDIA Opens RTX Spark Preorders for Windows AI PCs
Laptops are due October 16 and compact desktops in November; DGX Station for Windows remains a preview.

US eases AI chip export rule for the UAE on July 14, 2026: the changes
A US rule published on July 14, 2026 lets approved UAE entities receive advanced computing items without a license.

CoreWeave puts Vera Rubin NVL72 into production with Cognition
CoreWeave opened Vera Rubin NVL72 on September 30, 2026; Cognition is the first customer running production on it.

Huawei Ascend roadmap: 960, 970 and 980 dates set September 17, 2026
At Huawei Connect 2026 on September 17, Huawei set dates for the Ascend 960DT, 960PR, 970 and 980 AI chips.
Broadcom AI semiconductor revenue: $16.7 billion, reported Sep 2, 2026
Broadcom said on September 2, 2026 that AI semiconductor revenue reached $16.7 billion in its third fiscal quarter.

Nvidia Q2 fiscal 2027 data center revenue: $89.0 billion, Aug 26, 2026
Nvidia reported on August 26, 2026 data center revenue of $89.0 billion for the second quarter of fiscal 2027.

Nvidia Groq 3 LPX enters full production on August 24, 2026
Nvidia said on August 24, 2026 that its Groq 3 LPX inference accelerator is in full production, with Nebius first.

Cerebras unveils CS-4 on Aug 18, 2026: WSE-3T, three wafers per rack
Cerebras unveiled the CS-4 system on August 18, 2026, built on its WSE-3 Turbo wafer-scale processor.

SK hynix HBM4 mass production 2026: shipments began in Q2, Q4 ramp expected
SK hynix began HBM4 mass shipments in Q2 2026; its Q3 results expect HBM4 to be mainstream in Q4.

AMD launches Instinct MI400 Series on July 23, 2026: MI455X and MI430X
AMD launched the Instinct MI400 Series on July 23, 2026: the MI455X for frontier AI and the MI430X for HPC.
Coverage includes Microsoft Maia, Meta MTIA, custom chips from Broadcom and Marvell, Cerebras, Groq and Chinese AI chips, plus the supply chain (TSMC, HBM from SK hynix, Samsung and Micron, ASML) and export controls, from the companies’ own datasheets, releases and filings. Specs are compared in our AI chips table, rental prices in GPU prices, and background in our chip explainers.
