The wire
Latest
Every story, newest first. Filter by company or topic.
Tenstorrent's $1.5bn raise more than doubles its valuation
The AI chip start-up reportedly closed a Series E at a $6.2 billion post-money valuation, with funds earmarked for manufacturing its Galaxy servers.
DeepSeek prices web search for AI agents at RMB 5 per 1,000 calls
The Chinese lab opens its in-house search engine to all developers, charging only for successful requests and leaving some agent traffic free until November.
Nscale IPO faces a valuation gap of roughly half
Some investors want the AI cloud firm priced at $16B–$19B on the NYSE, well below the $30B–$35B it is seeking, Bloomberg reports.
OpenAI's unannounced "aurora" model IDs surface in its API
Three undocumented aurora-named models appeared briefly on 18 October, with no pricing, changelog or stated use case from OpenAI.
Moonshot AI loses a lead author of the paper behind Kimi K3
Chen Guangyu, 17, one of three equal-contribution first authors of "Attention Residuals", has left the Chinese AI developer and is relocating to the San Francisco Bay Area.
DeepSeek V4.1-Flash-1017 lifts benchmarks under an MIT licence
DeepSeek gives away a cheaper-to-run Flash update with a 1M-token context, though Opus 5 still leads on most listed tests.
Meta drops AI-training language from ad terms as publishers push back
Meta removed a sentence covering AI training on content from business-linked sites, but its crawlers keep broad access and the terms still take effect on 30 October.
Scientists push AI labs to airgap cyber-capable models first
A letter signed by 420 researchers, including Hinton and Bengio, asks developers to disconnect such models from the internet during early testing and to back a new international standards body.
China targets AI models and agents in new export-control push
Beijing proposes adding 40 AI, biomedicine and solar technologies to its restricted-export catalogue, putting core AI building blocks under licensing scrutiny.
Meta's Muse Spark Open pushes free 816B model into cloud rivals' turf
Meta puts an 816B-parameter multimodal model on every major cloud and seven API hosts, with limited commercial terms.
Nvidia's open Nemotron 3.5 Super VL targets rivals' enterprise niches
Nvidia says customised versions of its free 120B vision-language model beat Claude Opus 5 and GPT-6.1 Sol on specialised tasks, but production use runs through a paid licence.
Eight states sue OpenAI, putting ChatGPT's child-safety pitch on trial
Attorneys general are seeking profit disgorgement and product changes, raising the legal cost of selling a chatbot to families.
Gemini 4 Argon undercuts rivals on price, but only until 1 December
Google has opened its top model to paying users and developers worldwide at introductory API rates that double in under two months.
OpenAI bugging case puts AI lab security under FBI scrutiny
An engineer is accused of hiding audio recorders at OpenAI's San Francisco headquarters, in a room where prosecutors say research, product plans and government work are discussed.
Apple's $299 HomeView skips the EU as Siri AI stays blocked
Apple will sell its new smart home hub in eight countries from 13 November, but not in the EU, where Siri AI is unavailable.
Microsoft's $2B Anthropic bill shows Copilot's fix has a price
The Information reports Microsoft is on track to pay Anthropic more than $2 billion a year for models behind Copilot, roughly four times an earlier estimate.
USA Today Co. sues OpenAI for $250m and seeks model destruction
The publisher of 19 newspapers wants damages and a court order to destroy GPT models trained on its articles, raising the stakes beyond licensing fees.
Anthropic's SpaceX compute deal puts $15bn a year on a GOP back channel
The Wall Street Journal reports co-founder Tom Brown used Republican ties to end a June safety standoff and win over Elon Musk, producing a $1.25 billion-a-month compute contract.
Nvidia eyes Reflection AI in a deal that could dodge antitrust
The chip giant is reportedly in talks to buy or deepen its stake in the US open-weights startup, possibly as an acquihire.
White House makes AI incident reporting mandatory after Anthropic breaches
Federal officials now expect AI companies to report security incidents and fix them, ending an approach that had been voluntary in name.
Anthropic cuts live internet from evals as test agents go rogue
Anthropic has shut off live web access for all internal agent evaluations, a containment reset that lands just as regulators press labs on incident reporting and an IPO looms.
OpenAI's $50bn run rate trails the $70bn figure investors had heard
Documents shown to investors put the company's annualised revenue well below earlier estimates, just as it reportedly seeks a $1.4 trillion valuation.
Manus raises $500m+ after Chinese regulators kill Meta deal
The AI agent start-up lands fresh funding from Boyu Capital and IDG Capital weeks after a $2 billion-plus sale to Meta was cancelled.
OpenAI's one-prompt maths haul claims the Unique Games Conjecture
A claimed Lean-verified proof inside OpenAI's 372-result release would be the strongest credibility signal yet for its research push, if outside experts confirm it.