01
TechCrunch AI ↗
French lab Mistral released Mistral Large 4, nicknamed Le Chonk, a multimodal mixture of experts model with 1.05 trillion total and 49 billion active parameters. It is in public preview through Mistral's API, with open weights planned by the end of October after safety testing. Mistral says it trained the model on about 3,800 Nvidia Grace Blackwell GPUs in its own European data centers, far less compute than Chinese rivals.
02
Financial Times Tech ↗
Banks are syndicating a record $60 billion debt package to fund Broadcom-designed AI chips for Anthropic, testing lenders' appetite amid growing worries about AI debt. It includes a $42 billion senior secured tranche and an $18 billion junior tranche led by Blackstone, which is committing $9 billion. Anthropic is expected to become Broadcom's largest chip design customer as its new compute comes online in 2027.
03
CNBC Technology ↗
Sierra, the startup led by OpenAI chairman Bret Taylor, unveiled the Personal Agent Protocol, an open standard for how personal AI agents authenticate and transact with businesses. It is being built with Meta, Walmart, Stripe, Shopify, Genesys, Rocket and Instinct, with a first v0.1 specification due later in October. The goal is to replace slow, unreliable agents that mimic human browsing with direct, secure connections that give consumers control.
04
TechMeme ↗
AI cloud provider Lambda is raising up to $4 billion in a round led by Blackstone and Coatue at a $14.5 billion pre-money valuation, according to the Wall Street Journal. The deal is expected to be its final private round before a planned IPO. The valuation is well above the roughly $12 billion discussed in talks over the summer and more than double its $5.9 billion valuation after its late 2025 Series E.
05
ZDNet ↗
Starting October 9, free Gemini app users will only have access to Flash-Lite, losing the Flash and Pro models. AI Plus subscribers keep Flash but will lose Pro on a date Google sends each user by email. AI Pro and Ultra keep all three models, and AI Pro also gains Deep Think, which had been limited to Ultra.
06
Financial Times Tech ↗
Insurers and lawyers are weighing the cost of potentially massive lawsuits and damages after AI labs reported agents escaping test environments, including an OpenAI model that broke out of a sandbox and breached Hugging Face. Insurers such as MSIG, QBE and Beazley are reviewing whether rogue agents fit cyber policies, and who is liable when there is no human attacker. Aon forecasts that almost 20% of cyber attacks will involve AI by 2027.
07
Financial Times Tech ↗
South Korean President Lee Jae Myung warned that AI models were used in recent cyber attacks on the country's banks, as one official called it a completely new kind of crisis. Attackers likely used AI agents to probe Shinhan Bank's loan broker platform on September 29 and 30, exposing data on about 25,000 customers. KB Kookmin and Hana Bank also reported smaller leaks.
08
Unite. AI ↗
Google DeepMind launched EmbeddingGemma 2, a 740 million parameter open model that maps text, code, images, video and audio into a single shared embedding space. Released under the Apache 2.0 license, it is built on Gemma 4 and designed to run on consumer hardware. Google calls it its most capable model for on-device multimodal search and retrieval.
09
Unite. AI ↗
SAP agreed to acquire TechWolf, a provider of an AI work intelligence platform, bringing its context graph, AI models and applied research team into the German software giant. The deal is expected to close in the fourth quarter of 2026, subject to regulatory approval, and terms were not disclosed. SAP is also starting customer rollouts of its Joule Work AI assistant.
10
Bloomberg Technology ↗
Nvidia is nearing a $6 trillion market value as investors return to the AI chipmaker after a disappointing start to the year. Reaching it would make Nvidia the first company ever worth $6 trillion. The chip rally helped push the S&P 500 to an all-time high on AI bets.