Anthropic’s Surprise Turn to Operating Profit
Investor materials reviewed by The Wall Street Journal indicate that Anthropic is poised to post its first operating profit of USD 559 million (approx. RM2.57 billion) in the second quarter, on USD 10.9 billion (approx. RM50.1 billion) in revenue. That more than doubles the USD 4.8 billion (approx. RM22.1 billion) it generated in the previous quarter, highlighting how quickly the Claude AI business model is scaling. The company’s quarterly growth is now described as faster than Zoom at its pandemic peak, and even ahead of Google and Facebook before their public listings. Crucially, Anthropic’s compute costs are expected to fall from 71 cents to 56 cents per dollar of revenue, giving it enough margin to tip into the black. By contrast, OpenAI does not expect to reach profitability until 2029 or 2030, underscoring a widening gap in AI company finances and capital efficiency.
SpaceX Supercomputers and a USD 1.25 Billion Monthly Bet
Anthropic’s path to profitability is unfolding alongside an enormous infrastructure commitment. SpaceX’s S-1 prospectus reveals that Anthropic is paying USD 1.25 billion (approx. RM5.75 billion) every month for AI compute access at the Colossus 1 and Colossus 2 supercomputer campuses through May 2029. That equates to roughly USD 41 million (approx. RM188.6 million) per day and about USD 15 billion (approx. RM69.1 billion) per year, with the total contract value reaching up to USD 45 billion (approx. RM207.3 billion). In return, Anthropic gets access to 220,000 Nvidia GPUs and more than 300 megawatts of power, effectively renting a purpose-built AI factory at massive scale. Either party can exit with 90 days’ notice, but for now the deal crystallises how central compute access has become to AI company finances—and how aggressively Anthropic is willing to spend to secure capacity for Claude’s growth.
Why Anthropic Is Hitting Profitability Before OpenAI
The contrast between Anthropic profitability projections and OpenAI’s extended losses stems from diverging growth and spending profiles. OpenAI expects to remain unprofitable until 2029 or 2030 and is forecasting a USD 74 billion (approx. RM340.7 billion) loss in 2028 alone, reflecting massive investment in models, infrastructure, and ecosystem bets. Anthropic, by comparison, is achieving break-even sooner because revenue is scaling faster than its already enormous compute bill. CEO Dario Amodei has said the company prepared for 10-fold growth but experienced 80-fold annualised growth in the first quarter, straining infrastructure but dramatically improving unit economics. As compute costs fall to 56 cents per revenue dollar, Anthropic’s Claude AI business model is demonstrating higher near-term operating leverage. The company still warns profitability may be bumpy as it ramps infrastructure, yet the timing gap versus OpenAI underscores sharply different risk and capital allocation strategies.
Investor Confidence, Valuations, and Competitive Positioning
These financial trajectories are already influencing market perceptions. Anthropic is said to be raising USD 50 billion (approx. RM230.5 billion) at a USD 900 billion (approx. RM4.15 trillion) valuation, which would edge past OpenAI’s reported USD 850 billion (approx. RM3.92 trillion) value. For investors, a clear line of sight to operating profit—even amid a USD 1.25 billion (approx. RM5.75 billion) monthly compute commitment—signals a business capable of converting AI demand into real cash flow. That strengthens confidence in Anthropic’s ability to keep funding its rapid expansion and negotiate favourable infrastructure deals, such as the SpaceX partnership. At the same time, OpenAI’s longer road to profitability suggests a bet on bigger, riskier platforms that may pay off later. The result is a competitive landscape where Anthropic markets itself as the more capital-efficient, revenue-dense alternative—potentially reshaping funding flows and customer decisions across the AI industry.