AI news, 28 August: Nvidia nears $13 billion Hugging Face takeover
Nvidia reportedly agrees to buy Hugging Face for $12.9 billion. According to The Information, the chip giant has struck a deal to acquire the leading open-source AI model hub, a move that would rank among Nvidia’s largest ever and give it control of a key platform used by millions of developers to share and download models and datasets. Business Insider noted talks valued the company above $13 billion but said no signed agreement was yet confirmed; neither company has publicly commented. The reported price marks a sharp rise from Hugging Face’s $4.5 billion valuation in 2023 and comes as open-source systems challenge closed models from firms such as OpenAI and Anthropic. Analysts say the deal would extend Nvidia’s reach beyond chips into model distribution and cloud services.
Tech giants warn of imminent AI-powered cyberattacks. OpenAI, Anthropic, Google, Microsoft and more than 100 other companies published an open letter on Thursday calling for urgent collective action, saying AI-enabled hacks on hospitals, water systems and other critical infrastructure will grow far more sophisticated within months. The signatories, including cybersecurity firms and banks, urged governments and organizations to make cyber defense a top priority, fix known weaknesses, and give defenders better access to advanced AI tools. The appeal follows recent incidents in which AI agents, including some linked to OpenAI, breached systems during testing, raising fresh questions about control of increasingly autonomous systems.
US judge rules Pentagon blacklisting of Anthropic unlawful. A federal judge in California found that the Trump administration illegally retaliated against Anthropic by designating it a national-security supply-chain risk and barring it from government work, according to the New York Times and Reuters. Judge Rita Lin wrote that the move punished the company for constitutionally protected speech about how its AI should not be used for mass surveillance or autonomous weapons, calling the national-security justification an “empty invocation.” Anthropic welcomed the ruling; the administration can still choose not to use its models and may appeal.
Anthropic launches standard for AI agents to control lab and factory hardware. The company released a research preview of its Model Hardware Standard, a framework that lets AI agents safely operate physical devices such as microscopes, robotic arms and manufacturing tools with far less custom coding. Anthropic said the standard can cut setup times from weeks to hours and enable round-the-clock automated experiments. It is sharing the early version with partners for safety testing before a planned open-source release, aiming to speed scientific and industrial work while adding guardrails against misuse.
Bill Gates renews call for stronger AI preparation. In a lengthy essay and interviews, the Microsoft co-founder warned that society remains unprepared for job disruption, cyber risks and other upheavals from advanced AI, criticizing tech firms for downplaying dangers while racing ahead. He advocated international coordination, possible taxes on AI that replaces workers, and reserved roles for humans.