OpenAI has closed the largest private funding round in history, securing $122 billion at a post-money valuation of $852 billion. The round was anchored by SoftBank, Andreessen Horowitz, D.E. Shaw Ventures, MGX, and TPG, with continued support from Amazon, Nvidia, and Microsoft. Roughly $3 billion came from individual retail investors through bank channels, marking the first time everyday investors have been able to participate in an OpenAI raise at this scale.
The numbers are staggering: OpenAI now generates $2 billion in monthly revenue, with enterprise clients accounting for more than 40% of that figure. The company reports over 900 million weekly active users and more than 50 million paying subscribers. Its ads pilot, launched just weeks ago, is already generating over $100 million in annualised recurring revenue.
Most significant is the strategic direction the funding will support. CEO Sam Altman outlined plans for an AI “superapp” that would merge ChatGPT, Codex, web browsing, and agentic capabilities into a single unified experience. The company also intends to accelerate work on its next-generation model, codenamed “Spud,” which reportedly finished pretraining on 25 March.
The round comes with a notable counterpoint: on secondary markets, OpenAI shares have become increasingly difficult to sell, with some holders describing them as “almost impossible to unload.” Investors are pivoting toward Anthropic, whose $380 billion valuation offers what many see as more upside.
This is not just a funding round; it is a statement of intent. At $852 billion, OpenAI is now valued higher than all but a handful of public companies on the planet. The superapp strategy signals that OpenAI no longer sees itself as a model provider but as a platform company, one that wants to own the entire user experience from search to code to commerce. Merging ChatGPT, Codex, browsing, and agentic tools into a single experience is a direct challenge to Google, Apple, and Microsoft’s hold on the digital gateway.
The secondary-market dynamics tell an equally important story. While OpenAI raises record capital from institutions, its existing shares are going cold. Investors are quietly pivoting to Anthropic, where they see more upside at a $380 billion valuation. Banks are offering OpenAI shares without carry fees while Anthropic shares still command a premium. Both companies are weighing IPO plans, and the market is essentially placing a bet on which horse will win the race to public listing.
The bottom line: OpenAI has the capital to attempt something no AI company has ever tried, building a unified AI superapp at global scale. But the secondary-market exodus suggests even some insiders question whether $852 billion is justified. The next six months, as both OpenAI and Anthropic race toward IPOs, will reveal whether this is the crowning moment of AI’s golden age or its most expensive bet.
The Great AI Investor Pivot. This week’s numbers tell a striking story of divergence. OpenAI is raising record capital at unprecedented scale, yet its secondary-market shares are going cold. Anthropic, valued at less than half of OpenAI, is where the smart money is flowing. Meanwhile, AI-driven job cuts have become the top reason for layoffs in the US, and Waymo’s autonomous rides are doubling at a pace that puts a million weekly trips within reach by year-end.
The signal for enterprise leaders: the AI economy is maturing beyond hype into hard economics. Companies that can demonstrate clear unit economics, whether in enterprise software, autonomous mobility, or advertising, are pulling ahead. The era of valuation-by-vibes is ending.
Google DeepMind has released Gemma 4, its most powerful open-source model family yet, in four sizes from 2.3B to 31B parameters, all under Apache 2.0 licensing. The flagship 31B dense model ranks third on Arena AI’s text leaderboard, outperforming models twenty times its size. With 256K token context windows, native vision and audio processing, and support for over 140 languages, Gemma 4 represents a direct challenge to proprietary models for on-device and edge AI.
Since the original Gemma launch, developers have downloaded the models over 400 million times, creating more than 100,000 variants. The combination of commercially permissive licensing and competitive performance makes Gemma 4 the most compelling argument yet that open-source AI can rival proprietary frontier models for many enterprise use cases.
Read the Google Blog →