Dario told Dwarkesh three weeks ago that Anthropic has been growing revenue 10x per year. $0 to $100M in 2023. $100M to $1B in 2024. $1B to $9-10B in 2025. Then added “another few billion” in January 2026 alone. This chart is what that 10x looks like on the consumer side. The revenue curve has been exponential for three years, but almost nobody could see it because 70-75% of Anthropic’s revenue comes from API and enterprise. The growth was hidden inside developer terminals and corporate workflows. Claude Code alone went from $500M run-rate in September to $2.5B by February. What changed in October is the consumer curve started matching the revenue curve. Claude Code launched on the web. Opus 4.5 shipped. Vibe coding went viral over the holidays. And suddenly the DAU chart caught up to the financial reality that was already there. Here’s the number that puts this in perspective. ChatGPT has 900 million weekly active users. Anthropic’s run-rate revenue is $9-10B. OpenAI’s ARR is about $10B. Nearly identical revenue, 1% of the consumer user base. Anthropic’s revenue per user is roughly 80-100x higher because the money comes from developers and enterprises building on the API, not $20/month chat subscriptions. Dario also said something most people skipped past: “Even though a part of my brain wonders if it’s going to keep growing 10x, I can’t buy $1 trillion of compute in 2027. If I’m just off by a year, or the growth rate is 5x instead of 10x, then you go bankrupt.” That’s the CEO of a company growing 10x per year telling you he’s terrified of betting on 10x continuing. Because at this scale, the difference between 10x and 5x is the difference between the most valuable company ever built and bankruptcy. There is no middle ground on an exponential. This DAU chart is one data point. The revenue curve is three years of data points. And Dario is simultaneously the most bullish and most cautious CEO in tech because he understands exactly what exponential means: the upside is a trillion-dollar company and the downside is zero. Same curve, one year of timing difference.