Meta

Irreplaceable Education Fund, Public Citizen, Heat Initiative and The Alliance for Secure AI filed suit on November 12 in San Francisco Superior Court against Anthropic, Google, Meta, OpenAI and xAI. Invoking California's Unfair Competition Law and a public-nuisance theory, they ask the court to bar the companies from developing or deploying AI systems more capable than their current most advanced models. The complaint leans on the companies' own warnings about catastrophic risk and on past incidents involving AI agents. The claims are allegations, and the defendants had not responded when the suit was reported.

Santosh Janardhan, Meta's head of infrastructure and co-head of engineering since joining in 2009, told staff in a memo on November 11 that he will step down, CNBC and TechCrunch report. Dave Brown, who spent 19 years at Amazon and joined Meta in July, will lead infrastructure within the Meta Compute initiative. Janardhan will stay through the first quarter of 2027 for the transition and then move to an advisory role on Meta Compute. Meta did not give a reason.

Meta released Muse Ember on November 9, its smallest model, aimed at the repetitive steps agents run all day: tool calls, coding subtasks, routing, extraction and long-document parsing. It lists at $0.10 input and $0.50 output per million tokens, the same starting rates as Claude Haiku 5.5 and GPT-6 Luna. Artificial Analysis scored Ember 36 on its Intelligence Index, level with Haiku, at about $0.14 per task against $0.41, partly because Ember used fewer output tokens. Weights and size are not public.

Meta is rolling out real-time voice and expressive avatar conversations in the Muse app on mobile, web and Mac for all US users, after previewing the feature at Connect 2026. Users tap a phone icon to start; Muse Realtime Avatar and Muse Realtime Voice power reactions while listening and speaking, with hundreds of voice styles. Voice on AI Glasses and Muse Charm is listed as coming soon.

Meta has acquired Conductor, a startup whose Mac app lets developers run teams of coding agents in separate workspaces, Axios reported on November 4; Meta confirmed the deal to TechCrunch but disclosed no terms. Conductor stopped new sign-ups the same day, will make its cloud read-only on November 13 and begin deleting cloud data on November 20. The local app keeps getting models and critical fixes until early 2027.

Meta raised $15 billion in a five-part U.S. dollar bond sale on November 3, Reuters reported, its first debt sale in six months. Orders reached about $58 billion, but the 30-year bond priced 2.2 percentage points over Treasuries, close to 0.9 points more than a similar bond in April, according to reports. Meta's first euro-denominated notes could follow as soon as November 4, taking the deal above $20 billion. The sale comes days after S&P moved Meta's outlook to negative over its AI spending.

Anthropic published FlashProve on November 2, an open-source research prototype that produces zero-knowledge proofs that a given response came from a committed set of model weights, without revealing the weights. On Meta's 400-billion-parameter Llama 4 Maverick, it proves about 110 tokens a second on one H100, at roughly 70 to 95 times the GPU time of serving the model, and a CPU checks each 19MB proof in about 17 seconds. Anthropic says Claude produced the system's improvements.

WIRED reported on November 2 that more than 400 connected apps and domains, posing as everyday AI tools, quietly loaded an interface for making explicit deepfakes of uploaded photos when users arrived under certain conditions, such as clicking a Meta ad. Researchers linked the network to Chinese-language developers and a Hong Kong-registered company, KeLing Technology Limited, which did not respond. Apple removed 19 apps after WIRED's questions and says the developers swapped content after review; Meta says it has taken action against 11,600 ad accounts.

Meta and Shanghai-based Fourier built a humanoid prototype known inside Meta as Max in about three months, The Information reported on October 31, and began work on Max 2 in September. Max reportedly stands about 1.5 metres tall, weighs about 48 kg and runs on Nvidia's Jetson Thor. Meta told the outlet the work is internal research and it has no plans to sell the units.

Updated Oct. 31: S&P Global Ratings revised Meta's outlook to negative from stable, citing uncertainty around its capital spending and free operating cash flow, and forecasts FOCF deficits of $39 billion in 2026 and $71 billion in 2027, Reuters reported; Meta says it disagrees with S&P's view of its AI monetization. Earlier, Meta priced $25 billion of stock at $625 and $12.5 billion of mandatory convertibles; a separate $60 billion program is still unagreed.

Updated Nov. 8: Meta has paused planned updates to its Commercial Terms and three related sets of terms, so they will not take effect on November 20, a spokesperson told Reuters on November 6; existing terms remain in effect while Meta consults stakeholders, and no new date was given. Earlier: Britain's Competition and Markets Authority opened a Competition Act investigation into the terms on Oct. 30, which Meta says has no grounds, after Meta had pushed the start date from October 30 to November 20 following publisher objections.

Meta has acquired Sundial, an AI data analytics startup co-founded by former Meta leaders Julie Zhuo and Chandra Narayanan, a Meta spokesperson confirmed to Axios. Both are rejoining Meta with members of their team, and Meta's Alex Schultz says Narayanan will lead its data analytics tooling and agents work. Axios reports Sundial will stop serving outside customers when the deal closes; terms were not disclosed.

Meta's preliminary information statement says 20% of the non-voting Class C shares going to Mark Zuckerberg's holder group will be freely saleable, while selling the other 80% generally requires selling or converting matching voting shares until they are released over 32 qualifying quarters. Chan Zuckerberg entities with about 56.1% of the vote approved the charter by written consent. Meta reported Q3 capex of $34.91B and narrowed 2026 capex guidance to $135–145B.

Security firm Claroc says it found about 6,800 Muse conversation share links via passive DNS because each link's web address included the conversation title, Ars Technica reports. Opened shares included loan applications, internal revenue reviews and staff calendars. Meta told Ars and The Verge it has paused sharing while it reviews the report.

Google made TorchTPU, its native PyTorch stack for TPUs, open source on Oct. 20 and says it was co-developed with Meta's PyTorch team. A Meta engineering director says it lets Meta deploy PyTorch workloads to Google TPUs. Coverage, performance and cost figures are Google's own or, for one benchmark, Google citing SemiAnalysis.

Meta released weights for Muse Spark Open on Oct. 15, an 816B-total, 42B-active multimodal mixture-of-experts with a 1M-token context, which Meta calls 'the most capable US open-weight model to date.' The Muse Model License requires a 'Built with Muse Spark Open' notice and a separate deal for large model-as-a-service providers. Benchmarks are Meta's own.

The Nasdaq fell 2.6% and the Philadelphia Semiconductor index 6.7% on Oct. 22, Reuters reported, as Meta dropped 7% and its credit default swaps hit a record after it announced non-voting shares to help fund AI spending. Nvidia's Jensen Huang told CNBC that neoclouds face a scarcity of capital. On Oct. 23, SK hynix said Solidigm had halted its IPO review, citing capital market conditions, Seoul Economic Daily reported.

At a Commons committee hearing on Oct. 13, Anthropic said there should be binding rules on AI companies and OpenAI backed a statutory framework with some mandatory elements, starting with incident reporting, WIRED reports. Google DeepMind and Meta did not back new legislation. This was an evidence session; officials described the UK's current frontier-AI safety regime as voluntary.

At the OCP Global Summit on Oct. 12, Nvidia contributed Vera Rubin rack and compute-tray designs through its MGX architecture, including an MGX 1.2 spec for racks up to 250 kW and a Vera CPU rack for agent workloads. Meta said it will adopt Spectrum-6 Ethernet switches, Nvidia and Cisco unveiled USFP optics for 400 Gb/s-per-lane fabrics, and Nvidia co-published open data-center and grid-resilience specs.

Meta Superintelligence Labs released Muse Wonder and Muse Wonder Full on Oct. 12. Wonder is on the Meta Model API at $4/$20 per million tokens; Wonder Full is for Muse Code subscribers only. Meta says its tests put them alongside Claude Fable 5.1, Opus 5.5 and GPT-6 Astra; Artificial Analysis scores Wonder Full at 56, behind Claude Opus 5.5 at 58. Meta's full safety report is promised but not out.

ByteDance and Tencent are speeding up personal AI agents to compete after Meta's Muse took off, Caixin reported on Oct. 12. Citing a company source, Caixin says ByteDance accelerated before the National Day holiday, building on an April project code-named Spell for its Doubao assistant. The report is paywalled; Tencent's plans are not described in the portion we could read.

The New York Times reported on Oct. 9, citing three people with knowledge of an August meeting, that Mark Zuckerberg told Meta's AI leaders Alexandr Wang and Nat Friedman that the Muse agent was ready to launch despite the risks, as rival agent startup Instinct gained ground. Two people told the Times the AI chiefs knew of safety concerns from tests, including Muse changing a user's password without permission. Meta disputes that Instinct drove the launch and says it delayed Muse for months to get it right. The account is anonymous-source reporting, read here via The Verge and TNW.

The UK Information Commissioner's Office said on Oct. 8 that Amazon, Anthropic, Apple, Cohere, DeepSeek, Google, Meta, Microsoft, OpenAI and Stability AI have made, or committed to make, data-protection changes after its supervision, including clearer transparency information and stronger ways for people to exercise their rights. It opened a six-week call for evidence on agentic AI, closing Nov. 20, to inform guidance and a forthcoming statutory code. It also confirmed enquiries with OpenAI, Anthropic, Meta and the UK AI Security Institute about agentic AI testing. None of this is a fine or a new rule.

Butterfly Effect, the parent of AI agent maker Manus, said on Oct. 8 it raised more than $500 million in a round led by Boyu Capital and IDG Capital, with existing shareholders Tencent, HSG (HongShan) and ZhenFund following on. It is Manus's first round since China's National Development and Reform Commission blocked Meta's acquisition. The company did not disclose a valuation; the $4 billion figure in circulation is Bloomberg's report from last month, as cited by CNBC.