California's AI Safety Act heads to the governor, Europe's AI enforcers target financial institutions, and an AI medical platform just crossed a $12 billion valuation.
California's AI Safety Act passes final vote, now waits for the governor
The Frontier AI Safety Act cleared its final California Assembly vote on August 10 and is now sitting on the governor's desk. The bill would require AI companies to prove their most powerful models cannot be used to cause mass harm, and would hold them legally liable if they cannot. Every major AI lab in the world has been lobbying against it.
Why it matters: Silicon Valley is in California. A signature here becomes an effective global standard: AI labs can't build one product for California and another for everywhere else. This is the biggest regulatory decision in AI since the EU AI Act.
EU AI Act enforcement week two: France's financial regulator goes after 14 banks
Nine days into live enforcement of the EU AI Act, France's data protection authority issued formal information requests to 14 financial institutions using AI-powered credit scoring, demanding technical documentation proving compliance with the law's high-risk system requirements. This is the first targeted enforcement action since the rules kicked in on August 2.
Why it matters: Regulators aren't waiting. If you work in finance, insurance, hiring, or healthcare and your organisation uses AI to make decisions about people, these rules are now live, and enforcers are already moving.
Medical AI platform OpenEvidence raises $250M, now valued at $12 billion
OpenEvidence, which helps doctors quickly surface relevant clinical evidence and medical literature at the point of care, raised $250 million in a Series D round at a $12 billion valuation. It is one of the largest healthcare AI fundraises of 2026 and signals that investors believe AI tools designed specifically for clinicians are past the proof-of-concept phase.
Why it matters: Healthcare AI is no longer running small pilots, it is being funded at sovereign wealth scale. If you work in medicine, public health, or health tech in India, the international benchmark for what these tools can do is being set right now.
Google's AI will now call businesses, check stock, and buy things on your behalf
Google launched consumer AI agents that can independently make real phone calls to businesses (checking inventory, making reservations, and completing purchases), without you ever picking up the phone. Gemini 3.5 also received updates pushing it deeper into search, coding, and multi-step task automation for everyday users.
Why it matters: AI that acts in the world on your behalf is meaningfully different from AI that answers questions. This is what the shift from chatbot to agent looks like for regular users, not just developers.
National Grid invests $1.75B to power AI data centers, the energy problem is real
National Grid acquired a 35% stake in Joulent, a US energy company focused on supplying power to AI data centres, for $1.75 billion. Separately, reports from Michigan documented the measurable impact that AI data centres are already having on local water supplies and electricity costs, with communities near large server farms seeing utility bills and grid strain rise.
Why it matters: AI's environmental and infrastructure cost is no longer theoretical. The electricity and water needed to run the world's AI systems is becoming a tangible policy and community issue, not just a sustainability talking point.
India's AI ambition meets its infrastructure reality
India's IndiaAI Mission is backed by a ₹10,371 crore government budget and is administered through MeitY's principles-based advisory framework, a deliberate rejection of the EU's prescriptive rulebook. Outlook Magazine's latest edition covers what it calls the "AI Divide": the growing gap between India's high-level AI investment narrative and the ground-level infrastructure, teacher training, and access gaps that shape whether AI education actually reaches students.
Why it matters: The budget and the intent are in place. The harder question (whether delivery infrastructure keeps pace) will determine whether India's AI push is transformational or another top-down initiative that stalls in implementation.
The governor's pen: California's AI Safety Act signing decision
A signature makes California the world's de facto AI safety regulator, and sets a legal liability standard every lab anywhere in the world will have to meet. A veto signals that even in the most progressive US state, AI industry lobbying can block major safety legislation. Both outcomes are consequential and neither is predictable. The decision is expected within the coming days and is the single most watched policy moment in AI right now.
Abu Dhabi closes $49B AI fund • LeapXpert raises $180M • Swiss lab launches open-source 170B model
Abu Dhabi's MGX sovereign fund closed its first fund at $49 billion (exceeding its target) with investments spanning semiconductors, AI infrastructure, and platforms. LeapXpert raised $180 million for AI-governed business communications across WhatsApp, Teams, and SMS, aimed at regulated industries. Swiss AI lab Libertas launched Solidus, a 170-billion-parameter open-source language model built around ethical AI and transparency principles.