Today's digest
AI news today: Friday, November 13, 2026
NVIDIA names Colette Kress interim CEO; SMIC warns AI chip supply will stay tight; open inference gains a new consortium
Get tomorrow's issue in your inbox.
- 01
SMIC says AI chip supply chain may stay tight through 2027
Source: Reuters
On November 13, SMIC co-CEO Zhao Haijun said demand from data centres and private clouds is keeping the AI-related semiconductor supply chain tight, with constraints likely to persist through 2027. He also said some equipment orders take more than 24 months to arrive and that memory supply may not rebalance until mid-to-late 2028. These are management outlook comments, not an independently established forecast.
Why it mattersAI infrastructure depends on more than accelerator design: foundry capacity, memory and manufacturing equipment remain bottlenecks. Buyers should treat long lead times as a planning risk rather than assume announced capacity targets will become operational on schedule. More on AI chips & infrastructure →
- 02
NVIDIA announces Jensen Huang's resignation and appoints Colette Kress interim CEO
Source: NVIDIA
NVIDIA's board announced on November 12 that Jensen Huang had resigned as president, CEO and a board member, effective immediately. CFO Colette Kress was appointed interim CEO while the board searches for a permanent successor, and Harvey C. Jones was appointed the company's first board chair. NVIDIA did not state a reason for the resignation in the release and said it expects operations, customer commitments and the product roadmap to continue without interruption.
Why it mattersNVIDIA is a central supplier in the AI-computing ecosystem, so its leadership transition is significant for governance, execution and customer confidence. The company says its operating plans and product roadmap remain unchanged; any longer-term impact on strategy or market position is not yet established. More on NVIDIA →
- 03
Major AI hardware and cloud firms form the SGLang Consortium for open inference
Source: LMSYS Org
LMSYS Org announced a consortium with NVIDIA, AMD, Intel, Google, Meta, Cloudflare, Modal and RadixArk to work with SGLang maintainers on open-source inference software. Its initial focus includes frontier-model support, reasoning and multimodal workloads, distributed inference and hardware optimization. SGLang maintainers retain responsibility for technical decisions and code acceptance, and the project remains open to the wider community.
Why it mattersAs inference becomes a larger share of AI computing, coordinated engineering across hardware and cloud providers could improve efficiency and portability. This is a collaboration and roadmap forum, not evidence that every vendor will support every feature or that inference costs will immediately fall. More on NVIDIA →
- 04
Cerebras lifts its 2026 outlook after strong cloud and inference growth
Source: Cerebras Systems
Cerebras reported third-quarter 2026 GAAP revenue of US$210.8 million, up 55% year over year, and cloud and other services revenue of US$141.9 million, up 265%. Its non-GAAP Core revenue was US$230.9 million, up 70%, while GAAP operating margin was negative 106% and Core operating margin was negative 14%. The company raised its full-year Core revenue outlook to US$940 million–US$960 million and said it had signed a cloud deal with an unnamed top-three hyperscaler.
Why it mattersThe results show strong demand for high-speed inference while also illustrating the cost of infrastructure expansion. Readers should distinguish GAAP from company-defined non-GAAP Core metrics and treat capacity, customer and future-margin statements as forward-looking; Cerebras itself identifies customer concentration and data-centre availability as risks. More on AI funding & valuations →
- 05
AMD and PsiQuantum expand their partnership to co-design quantum-computing infrastructure
Source: PsiQuantum and AMD
The companies announced an expanded partnership combining AMD EPYC CPUs, Instinct GPUs, Versal adaptive computing and ROCm software with PsiQuantum's silicon-photonics platform. AMD plans an equity investment but did not disclose its size. The companies say initial applications will include drug discovery and cite early GPU-based simulations of covalent inhibitors; this is a company-reported research claim, not evidence of a deployed fault-tolerant quantum computer or demonstrated quantum advantage.
Why it mattersThe announcement highlights the hybrid classical and quantum infrastructure being assembled for future quantum systems. Near-term work still relies on conventional CPUs and GPUs alongside photonic hardware; commercially useful, fault-tolerant quantum computing remains a future objective. More on AI funding & valuations →
- 06
AMD adds MI450X to its Instinct MI400 lineup and publishes MI455X inference results
Source: AMD
AMD introduced the MI450X option for its Helios rack-scale platform, with 288 GB of HBM4 memory per GPU, and described MI430X scientific-computing capabilities. AMD said its internal testing measured 21,333 tokens per second per MI455X GPU on DeepSeek-R1 in an MLPerf Inference Offline configuration, up to 30% above NVIDIA's latest published result for its Vera Rubin platform. The comparison is vendor-reported and depends on the stated configuration.
Why it mattersThe release signals continued competition over inference throughput, memory and rack-scale systems. Buyers should compare independent benchmark submissions, total system cost, power, software maturity and availability rather than treat one vendor test as a universal performance ranking. More on NVIDIA →
- 07
Nonprofits sue major AI developers seeking a court order to halt frontier-model development
Source: Reuters
Reuters reported on November 12 that several nonprofits sued OpenAI, Anthropic, Meta, Google's parent Alphabet and SpaceX's AI division in California, asking a court to block further development of frontier AI products. The complaint alleges unfair business practices and public nuisance and cites safety concerns; these are plaintiffs' allegations, not judicial findings. Reuters said Meta declined to comment and other companies had not immediately responded.
Why it mattersThe case adds a litigation route to debates over frontier-model risk, governance and development controls. Its significance will depend on court decisions and evidence; filing a complaint does not itself impose a development ban or establish the allegations. More on OpenAI →
New here? How to read this digest
More recent digests
- Sunday, November 15, 2026Big Nvidia investors and a 115,000-signature petition push for Huang's return; Nvidia drops nearly 13% on the week; Asia moves to steady chip supply
- Saturday, November 14, 2026OpenAI lets ChatGPT agents pay for you; VCs fear an Nvidia funding gap after Huang's exit; LangChain hits a $3B valuation
- Thursday, November 12, 2026Meta's Muse raises questions about AI-generated podcasts; Gemini Intelligence reaches Pixel 10 in India; Adobe and Google deepen AI integration
- Wednesday, November 11, 2026Anthropic begins its IPO roadshow; TSMC backs Anthropic and OpenAI; Google releases Gemini 4 Carbon
- Tuesday, November 10, 2026Anthropic upgrades its frontier models; Google expands its AI chip lineup; Australia moves toward mandatory AI incident reporting
- Monday, November 9, 2026Anthropic commits $3.4B to secure data-centre power; OpenAI raises $11B at a reported $1.25T valuation; GPT-6 Cyber puts AI-assisted cyber defence in focus
- Thursday, November 5, 2026Moonshot seeks Nvidia Blackwell access via Europe data centers; YC startup Discovered Materials offers $1M for five AI materials problems; and Midterm wins raise risk of slower data-center buildout for AI
Free daily digest
The AI news that matters, in five minutes.
5–10 hand-picked stories a day, each with a plain-language line on why it matters. Every item links to its source.
Free · Unsubscribe any time · Privacy