Anthropic overtakes OpenAI in revenue, India's AI funding soars 4x, the EU AI Act countdown is real, and a Chinese startup is matching frontier models at a fraction of the cost.
Anthropic quietly overtakes OpenAI in the revenue race
Anthropic has overtaken OpenAI in business subscriptions and self-reported revenue, on track to hit $47 billion and reach profitability a year ahead of OpenAI's schedule. Similarweb data shows ChatGPT's share of monthly visits fell below 50% of the generative AI market for the first time ever in May, while Claude's user base keeps climbing. This is the first real power shift at the top of the AI industry since ChatGPT launched.
Why it matters: For three years, OpenAI was the only name most people knew. That era is ending. The products you use for work and study will increasingly be shaped by whoever wins this race, and it is no longer settled.
White House drafts voluntary AI release standards as OpenAI offers US govt a 5% stake
The White House is in advanced talks with OpenAI, Google, and Anthropic on a framework of voluntary AI release standards, an announcement could come within days. Separately, OpenAI has proposed handing the Trump administration a 5% equity stake, an unprecedented move that would make the government a direct financial stakeholder in the country's leading AI lab. Also: Anthropic's Fable 5 model is back in global circulation after the US Commerce Department lifted export controls, following new safeguards implemented by the company.
Why it matters: Governments are moving from observers to participants. When a state holds equity in an AI company, it changes every safety, regulation, and competition conversation that follows. Worth watching very closely.
EU AI Act goes fully live August 2, deepfake labelling becomes mandatory in 29 days
The EU AI Act reaches full enforcement on August 2, 2026. From that date, providers must ensure AI-generated content is marked in a machine-readable format detectable as artificially made, and deployers must disclose synthetic content clearly at first interaction. This applies to images, video, audio, and text. Separately, the US DEFIANCE Act is now in force, giving victims of non-consensual sexual deepfakes a federal right to sue creators and distributors for up to $250,000 in damages.
Why it matters: The era of posting AI-generated content without disclosure is ending on both sides of the Atlantic. If you create content professionally, these rules affect what you can publish and how you must label it.
India's AI funding soars 4× in H1 2026: Sarvam AI leads with $234M raise
Indian AI startups raised $676 million in the first half of 2026 across 57 deals, a 4x jump year-over-year, with deal count up 90%. The biggest raise this week: Sarvam AI, the lab building India's full-stack sovereign AI, closed $234 million in a Series B led by HCLTech. The government is backing the surge with a $1.15 billion Fund of Funds and a Rs 1 trillion R&D scheme spanning AI, quantum computing, and deep tech.
Why it matters: India's AI ecosystem is no longer just potential, it is capitalising at pace. Sarvam's raise signals that sovereign, India-first AI models are attracting serious institutional money, not just government grants.
Suno raises $400M for AI music. A medical AI for doctors hits $12B.
Suno, the tool that lets anyone create a full song from a text prompt, raised $400 million in a Series D round led by Bond, valuing the company at $5.4 billion. Separately, OpenEvidence (an AI medical platform used by over 700,000 physicians) raised $250 million at a $12 billion valuation. The two rounds together signal that AI is no longer just a productivity tool: it is becoming a creative and professional infrastructure layer.
Why it matters: When a music tool and a doctor's tool are both worth billions in the same week, it tells you how wide the AI market has become. The tools being funded today will be defaults in your field within two years.
Chinese AI startup GLM-5.2 matches Claude Opus and GPT-5.5 at a fraction of the price
Z.ai, a Chinese AI startup, has released GLM-5.2, a model whose coding and agentic AI capabilities now approach those of Anthropic's Claude Opus 4.8 and OpenAI's GPT-5.5: at a substantially lower cost. Companies are already routing simpler tasks to cheaper models and reserving premium models for complex work. Open-source and Chinese models are gaining adoption everywhere except industries where security concerns limit use.
Why it matters: When a cheaper model does 80% of the job, businesses stop paying for the expensive one for routine tasks. This reshapes pricing across the entire industry and makes frontier models compete on quality, not just brand.
OpenAI's 5% stake offer: what it means when AI labs hand equity to governments
OpenAI's proposal to give the Trump administration a direct equity stake is unlike anything the tech industry has done before. A government that owns a share of an AI company has a financial interest in that company's success, which creates very different incentives than a government that merely regulates it. If this deal goes through, it could become a template other countries use to demand equity positions from AI labs operating in their jurisdictions. That changes the global governance map.
Google brings AI upgrades to every school + AI-designed vaccine clears first human trial
At ISTE 2026, Google announced six AI updates for education, including Read Along, a real-time reading fluency tool now available across all Google Workspace for Education tiers, including the free Fundamentals tier used by most schools. Separately, a University of Cambridge team reported that an AI-designed vaccine component completed its first human trial successfully, a milestone in using AI to accelerate how new vaccines are built and tested.