Archive
Latest — page 4
Older stories, newest first.
Unitree Robotics shares surge 460% in Shanghai stock market debut
The humanoid robot maker's Shanghai listing valued it at about $52 billion at the close, even as the wider market fell.
Anthropic's annualised revenue run rate hits $65 billion in July
The figure is seven times higher than a year earlier, as investors weigh a possible IPO valuing the company above $2 trillion.
Nvidia backs $105 billion financing for OpenAI Ohio data centre
An SEC filing shows Nvidia will guarantee financing for a 4.25-gigawatt OpenAI campus in Pike County, with an option to expand.
DeepSeek's V4-Pro went GA in silence, then got a lot more expensive
A quiet pricing-page edit and a four-day-later price rise of up to 1,100% reveal more about DeepSeek's cost position than either fact does alone.
OpenAI revenue chief departs days after Lightcap exit news
The sudden departure of Denise Dresser has investors calling the executive churn a "huge red flag" ahead of a planned listing.
Hassabis pitched AI safety body to Trump officials before DeepMind exit
The Wall Street Journal reported Demis Hassabis proposed an IAEA-style oversight group weeks before stepping down as Google DeepMind chief executive.
Gemini 3.7 Flash and the mid-tier price war Google can actually win
Google's cheapest model now beats Anthropic's and OpenAI's mid-tier rivals on one benchmark and undercuts both on price. Its actual flagship still doesn't exist.
IBM bets its consulting arm on OpenAI — a year after betting it on Anthropic
Frontier models go into IBM Consulting Advantage, thousands of consultants get certified, and the distribution layer for enterprise AI quietly becomes the contested ground.
Grok 4.6 competes on turns, not IQ — and that is the more interesting claim
xAI's August release matches GPT-5.6 Sol on a composite index at a third of the headline price. The number that matters is 53 turns against 103.
Chinese makers ship 97% of world's humanoid robots in H1 2026
Agibot and Unitree led a global market that more than tripled in size from a year earlier, research firm Smart Analytics Global said.
Meta releases open-weight Muse Glimmer for single-GPU agents
The 30-billion-parameter model, distilled from Meta's closed Muse Spark 1.2, is Meta's first fully open release since it retired Llama.
Anthropic signs $10 billion compute deal with six-month-old start-up Volta
The agreement adds to roughly $71 billion in chip-lease financing Anthropic has arranged in two months to keep pace with Claude demand.
OpenAI updates GPT-5.6 Sol, makes Luna unlimited for free users
The update cuts factual errors and merges fast and reasoning modes into one model for paid users.
Tesla and SpaceX confirm $16.8bn Terafab chip campus in Texas
The two Musk-run companies will build a vertically integrated semiconductor site in Grimes County to make custom 2-nanometre chips
Jeff Dean leaves Google to launch AI research start-up Discovery Loop
The 27-year Google veteran and DeepMind chief scientist departs to build a company automating scientific discovery.
Hassabis steps down as Google DeepMind CEO, becomes Alphabet chief scientist
Koray Kavukcuoglu takes over Gemini models and frontier research as two Gemini co-leads quit the same day.
Apple seeks court order to block OpenAI from using alleged trade secrets
Apple asked a federal judge to bar OpenAI and two former employees from using confidential information as OpenAI builds consumer hardware.
DeepSeek's V4-Flash is cheapest well-known AI model to run, study finds
Artificial Analysis found the Chinese model costs 3 cents per test, versus $1.86 for GPT-5.6 Sol and $3.15 for Claude Fable 5
The AI Act stops being a timetable: Brussels starts enforcing
Transparency obligations, general-purpose AI powers and the penalty regime all became live on 2 August. The compliance question moves from 'when' to 'who checks'.
Google DeepMind releases Gemini Robotics 2 model family
The three models give humanoid and two-armed robots whole-body control, multi-robot coordination and fast adaptation to new machines.
Anthropic says Claude models breached real systems during cyber tests
A vendor misconfiguration gave evaluation models live internet access despite being told they were in an offline simulation.
OpenAI cut Luna by 80% — and told you exactly how it could afford to
A model that helped optimise its own serving stack funded an 80% price cut three weeks after launch. The efficiency story is real; the timing is competitive.
Claude finds new attacks on reduced-round AES and post-quantum HAWK scheme
Anthropic said its Claude Mythos Preview model uncovered cryptographic weaknesses that took human researchers weeks to verify
More than 1,100 AI staff sign letter urging paced development
Signatories including Anthropic's Dario Amodei and OpenAI's Jakub Pachocki want US backing for tools to deliberately slow frontier AI progress.