Nvidia agrees to buy Hugging Face for $12.9 billion, reports ↗
Nvidia is reportedly moving to acquire Hugging Face for $12.9 billion, potentially giving the chip giant control of one of the most important distribution hubs for open AI models, datasets and developer tools.
There’s a fairly big caveat - neither Nvidia nor Hugging Face had publicly confirmed the deal, while separate reporting suggested the talks may not have produced a signed agreement. So, a huge move… or a huge almost-move.
Strategically, it makes sense. Hugging Face sits right where developers decide which models to run, while Nvidia increasingly wants influence far beyond simply selling the GPUs underneath them.
Exclusive-Anthropic planned, then abandoned $7 billion purchase of MatX, sources say ↗
Anthropic discussed buying AI-chip startup MatX for roughly $7 billion as it searched for a faster route into custom silicon.
Those acquisition talks are no longer active, with discussions reportedly shifting toward a possible partnership. MatX was founded by former Google TPU engineers - exactly the sort of talent Anthropic needs if it wants to reduce its dependence on outside chip suppliers.
Basically, Claude's hardware ambitions are getting considerably less theoretical.
Major tech companies call for defensive surge to defeat AI-driven hacks ↗
OpenAI, Anthropic, Microsoft, Alphabet and Amazon joined more than 100 organizations in calling for a major coordinated push against AI-powered cyberattacks.
The group warned that increasingly capable models could make sophisticated attacks far more widespread, and urged governments and businesses to treat cyber defence as an immediate leadership priority.
A slightly peculiar moment, frankly - some of the companies building increasingly capable AI are simultaneously warning everyone to prepare for what increasingly capable AI might do.
Anthropic unveils new framework allowing AI agents to operate physical devices ↗
Anthropic unveiled a research preview of Model Hardware Standard, or MHS, designed to let AI agents safely communicate with and operate physical equipment.
Think microscopes, liquid handlers, robotic arms and even laser systems used with quantum computers. Agents could coordinate several machines, modify parameters and potentially recover from hardware problems without constant human intervention.
Anthropic plans to test the framework with partners and build safety evaluations before eventually open-sourcing it. Claude is, in a rather literal sense, starting to grow hands.
Gemini Omni 1.1 Flash lets you build with more control ↗
Google released Gemini Omni 1.1 Flash with a sizeable batch of upgrades aimed at making AI-generated video more controllable and production-ready.
Developers can extend scenes, specify starting and ending frames, create cheaper 360p drafts and upscale finished footage to 4K. Scene extensions can stretch video sequences while trying to preserve visual consistency.
It’s less "type prompt, pray for cinema" and more like a proper editing workflow now… or so it seems.
Huawei plans for more AI pharma tie-ups, says healthcare president ↗
Huawei wants to deepen its AI partnerships with pharmaceutical companies, expanding beyond drug manufacturing into development, clinical work and eventual implementation.
Its healthcare business already offers AI-assisted compound-screening technology powered by Huawei's Ascend and Kunpeng chips.
The bigger play is obvious - AI drug discovery is becoming another battlefield where compute companies, cloud providers and pharmaceutical firms are all suddenly wearing each other's coats.
SK Hynix holds groundbreaking ceremony for $4 billion Indiana AI chip packaging facility ↗
SK Hynix broke ground on a more than $4 billion advanced packaging and research facility focused heavily on memory technology for AI systems.
The site will include next-generation high-bandwidth-memory packaging alongside AI semiconductor research and development. HBM has become one of the quiet choke points of the AI boom - GPUs aren't much use sitting around waiting for memory.
The investment is another sign that the AI infrastructure race is spreading from model labs and GPU design into practically every layer of the semiconductor supply chain.
FAQ
Why would Nvidia want to acquire Hugging Face?
Hugging Face is an important hub for AI models, datasets and developer tools, giving it considerable influence over how developers discover and deploy open AI technology. An acquisition could therefore extend Nvidia’s reach beyond supplying GPUs and into the software and model-distribution layer. However, the reported $12.9 billion deal had not been publicly confirmed by either company, while separate reporting suggested no agreement may have been signed.
Why is Anthropic investing more attention in AI chips and hardware?
Anthropic’s reported discussions with MatX suggest it is exploring ways to gain greater control over the computing infrastructure behind its AI models. MatX was founded by former Google TPU engineers, making its expertise particularly relevant to custom silicon development. Although acquisition talks reportedly ended, a partnership remained possible, indicating Anthropic’s hardware strategy could continue without requiring the company to purchase MatX outright.
How could Anthropic’s Model Hardware Standard change AI agents?
Model Hardware Standard, or MHS, is designed to help AI agents communicate with and operate physical equipment. Potential applications include microscopes, robotic arms, liquid handlers and specialised laboratory hardware. The framework could allow agents to coordinate multiple machines, adjust operating parameters and respond to some hardware problems, although Anthropic plans further testing and safety evaluations before eventually making the technology open source.
What makes Gemini Omni 1.1 Flash more valuable for AI video production?
Gemini Omni 1.1 Flash introduces greater control over how AI-generated video is created and edited. Developers can specify starting and ending frames, extend existing scenes, generate lower-cost 360p drafts and upscale completed footage to 4K. These features move the workflow closer to conventional production tools by giving creators more opportunities to iterate, refine sequences and maintain visual consistency across longer generated scenes.
Why are semiconductors and cybersecurity becoming central to AI industry news?
AI growth increasingly depends on more than better models. Companies need GPUs, high-bandwidth memory, advanced packaging, secure infrastructure and protection against increasingly capable automated cyberattacks. SK Hynix’s Indiana investment highlights the pressure to expand the semiconductor supply chain, while the cybersecurity initiative involving major technology companies shows that defensive capabilities are becoming an equally important part of scaling AI safely.