Actu IA
Des actualités réelles et vérifiées du secteur de l'IA, rédigées par l'équipe RecapAI avec la source originale en lien.
AffairesNvidia's Hugging Face acquisition is official: a $13 billion deal
Nvidia has confirmed its acquisition of Hugging Face at a price of nearly $13 billion, turning the deal reported in late August into an official agreement — Nvidia's second-largest acquisition on record.
AffairesAnthropic set to unveil its IPO prospectus after Labor Day, targeting a $2 trillion-plus valuation
Anthropic is preparing to publicly release its IPO prospectus after the Labor Day holiday, with a first trade targeted for late September or early October and reports suggesting the raise could value the Claude maker north of $2 trillion.
ProduitGoogle launches Gemini 3.8 Flash, its third Flash update in six weeks
Google introduced Gemini 3.8 Flash alongside a restricted cybersecurity variant, Gemini 3.8 Flash Cyber — its third update to the Flash tier in just six weeks, continuing an unusually fast release cadence.
ProduitAdobe brings Photoshop, Firefly, and 70+ creative tools into Slack
Adobe launched an MCP-based Slack app that puts over 70 creative tools — including Firefly, Photoshop, Premiere, Illustrator, and Acrobat — directly inside Slack conversations, letting Slackbot call them using context pulled from messages and files.
SécuritéOpenAI says its Astra model is the first to cross a 'Critical' cybersecurity risk threshold
OpenAI disclosed that its Astra model is the first to reach the 'Critical' cybersecurity capability threshold under its Preparedness Framework, after internal testing showed it could independently discover and chain together working exploits against hardened systems.
InfrastructureAnthropic signs $35 billion compute deal with Nvidia-backed Lambda for a Texas data center
Anthropic has locked in a $35 billion cloud computing deal with Nvidia-backed Lambda, securing roughly 350 megawatts of GPU capacity at a new Texas data center to fuel AI training and inference.
JusticeThe EU officially classifies ChatGPT as a 'Very Large Online Search Engine'
The European Commission has designated ChatGPT as a Very Large Online Search Engine under the Digital Services Act, placing it in the same regulatory category as Google Search and subjecting OpenAI to stricter compliance rules by January 2027.
JusticeSony Music Publishing and Warner Chappell file copyright lawsuit against Anthropic
Sony Music Publishing and Warner Chappell have sued Anthropic, alleging the company used copyrighted song lyrics to train its Claude models without permission. It's the latest in a wave of copyright litigation targeting AI labs over training data sourcing.
AffairesDeepSeek nears $7.4 billion funding round at a $74 billion valuation ahead of a 2027 IPO
Chinese AI lab DeepSeek is closing in on a $7.4 billion funding round that would value the company at roughly $74 billion, up sharply from its $50 billion valuation in June, as it prepares for a possible IPO in 2027.
RechercheAnthropic researcher previews early work on self-improving AI systems
An Anthropic researcher shared early findings on models that can iteratively improve their own performance, offering a glimpse at research the company is exploring around AI systems that refine themselves over successive training runs.
AffairesOpen-weight AI labs become Silicon Valley's hottest acquisition targets
Startups building open-weight AI models are increasingly drawing acquisition interest from larger tech companies, as buyers look to secure both talent and openly-licensed model weights rather than building everything from scratch.
AffairesSenior Meta executive departs for OpenAI amid scrutiny in India
A senior Meta executive has left to join OpenAI, part of a broader pattern of AI talent movement between major labs, reported alongside growing regulatory scrutiny Meta is facing in the Indian market.
Sécurité IAOpenAI, Anthropic and Google join 100+ companies calling for AI-powered cyber defense
In a joint open letter, OpenAI, Anthropic, Google and over a hundred other companies and organizations called for a coordinated push to put AI-driven defensive tools in the hands of the teams protecting critical infrastructure, warning that AI-enabled cyberattacks are set to become far more common and sophisticated.
ProduitOpenAI begins testing ads on ChatGPT's free and Go tiers in India
OpenAI has started showing advertisements to ChatGPT users on its free and lower-cost Go plan in India, a notable shift for a product that has so far relied entirely on subscriptions and API revenue.
AffairesGoogle's Gemini brand still hasn't clicked with mainstream users, report finds
Despite Gemini's technical progress, reporting suggests the brand still struggles for mainstream recognition compared to ChatGPT — a challenge the piece frames as symptomatic of a wider naming and positioning problem across the AI industry, not just at Google.
AffairesNvidia reportedly nears a deal to acquire Hugging Face
Nvidia is reported to be closing in on an acquisition of Hugging Face, the widely-used open-source model hosting and machine learning community platform, which would deepen Nvidia's reach into the AI developer tooling stack.
InfrastructureAmazon triples its Nvidia chip order as AI infrastructure demand surges
Amazon has reportedly tripled the size of its Nvidia chip order, citing surging demand for AI compute — another data point in the ongoing infrastructure buildout underpinning the current generation of AI products.
SécuritéOpenAI publishes its official report on the Hugging Face security breach
OpenAI released an official report detailing a security breach connected to Hugging Face, outlining what happened and the steps taken in response — part of a broader industry conversation about securing the AI supply chain.