Summary
The AI Daily Brief covers several major developments reshaping the competitive landscape among frontier AI labs. The lead story is Anthropic’s announcement — disclosed inside a blog post about a new compute partnership — that it has reached $30 billion in annualized recurring revenue, a 3x increase since the end of 2025 and a 58% jump since February 2026. That figure, if calculated on a comparable basis, appears to put Anthropic ahead of OpenAI in ARR for the first time.
The episode also covers a Wall Street Journal deep-dive into the financials of both companies, sourced from fundraising disclosures. OpenAI projects spending $30 billion on model training in 2026 — triple last year — while Anthropic forecasts $28 billion in training costs by 2028. Both companies are forecasting small profits this year when training costs are excluded, with Anthropic projecting GAAP-style profitability by 2028 and OpenAI targeting cash-flow positivity by 2030. The WSJ frames both companies’ IPO narratives as bets that public markets will absorb massive ongoing capital burns.
Rounding out the briefing: Anthropic’s expanded Google and Broadcom compute deal adds 3.5 gigawatts of TPU capacity coming online from 2027; Google launches AI Edge Eloquent, a fully on-device AI dictation app powered by Gemma 4; and Gemma 4 hit 2 million downloads in its first week following its open-source release.
📺 Source: The AI Daily Brief: Artificial Intelligence News · Published April 12, 2026
🏷️ Format: News Analysis
