Anthropic’s Surprise Leap Into Operating Profit
Anthropic is on track to achieve what many assumed was years away for frontier AI labs: operating profitability at massive scale. Investor materials reviewed by The Wall Street Journal indicate the company expects an operating profit of USD 559 million (approx. RM2.58 billion) in the second quarter on USD 10.9 billion (approx. RM50.12 billion) in revenue, more than double the USD 4.8 billion (approx. RM22.07 billion) recorded the prior quarter. This surge has pushed Anthropic’s growth rate beyond that of Zoom during its breakout period and even ahead of Google and Facebook before their stock market debuts. In contrast, OpenAI is not projected to reach profitability until 2029 or 2030 and is reportedly bracing for a single-year loss of USD 74 billion (approx. RM340.4 billion) in 2028. The divergence underscores how Anthropic’s Claude revenue model is converting demand into cash flow far earlier than its closest rival.
The USD 41 Million-a-Day Compute Deal With SpaceX
Behind Anthropic’s profitability story sits one of the most ambitious compute deals ever disclosed. SpaceX’s S-1 filing reveals that Anthropic is paying USD 1.25 billion (approx. RM5.75 billion) every month for access to the Colossus 1 and Colossus 2 supercomputer campuses, equating to USD 41 million (approx. RM188.6 million) per day and USD 15 billion (approx. RM69.06 billion) per year, with up to USD 45 billion (approx. RM207.18 billion) over the contract term through May 2029. The agreement gives Anthropic full access to a 220,000‑GPU infrastructure backed by more than 300 megawatts of power, while allowing either party to exit with 90 days’ notice. For SpaceX, the arrangement validates a new line of business—AI compute as a service—just as it prepares for a landmark stock offering. For Anthropic, it locks in capacity critical to keeping up with demand for Claude while still driving toward operating profit.
Inside Anthropic’s Claude Revenue Model and Cost Discipline
Anthropic’s ability to post an operating profit despite eye‑watering compute spending highlights a distinct approach to AI company economics. The firm expects its compute costs to fall from 71 cents to 56 cents per dollar of revenue within a single quarter, even as it ramps infrastructure investment under the SpaceX deal. CEO Dario Amodei told developers that Anthropic had prepared for 10‑fold demand growth but instead saw 80‑fold annualized growth in the first quarter, straining compute supply yet turbocharging revenue. The Claude revenue model—focused on rapid monetization of high‑value enterprise and developer workloads—appears to convert infrastructure into income more efficiently than many rivals. While the company cautions that profitability may be uneven as it continues to invest, the combination of falling unit compute costs and explosive top‑line expansion suggests a path to sustainable margins rather than perpetual cash burn.
OpenAI Competition and Diverging Paths to Scale
Anthropic’s projected profitability years before OpenAI underscores two sharply different strategies for scaling frontier AI. OpenAI, according to investor expectations cited in the same materials, may not turn profitable until 2029 or 2030 and could incur a USD 74 billion (approx. RM340.4 billion) loss in 2028 alone. Anthropic, by contrast, is pressing toward operating profit while simultaneously negotiating a funding round that could value the company at USD 900 billion (approx. RM4.14 trillion), edging ahead of OpenAI’s USD 850 billion (approx. RM3.91 trillion) valuation. Both companies are racing to build the most capable models and capture platform‑level positions, but Anthropic’s emphasis on monetization discipline and compute efficiency is yielding earlier financial payoffs. The rivalry is no longer just about model benchmarks; it is about proving which economic architecture can support the next generation of AI safely and profitably.
What Anthropic’s Economics Signal for AI Investors
For investors and AI startups, Anthropic’s numbers reset expectations about what sustainable AI economics can look like. Delivering USD 10.9 billion (approx. RM50.12 billion) in quarterly revenue while shouldering a USD 1.25 billion (approx. RM5.75 billion) monthly compute bill demonstrates that frontier models can be commercially viable long before they reach technological maturity. At the same time, the company acknowledges profitability may fluctuate as it continues heavy infrastructure spending, especially under multi‑year commitments to providers like SpaceX. The market response has been swift: Anthropic is reportedly in talks to raise USD 50 billion (approx. RM230.21 billion) at a USD 900 billion (approx. RM4.14 trillion) valuation, a signal of growing investor confidence in AI companies that pair hypergrowth with improving unit economics. The message to the broader sector is clear: burn‑rate alone is no longer a badge of ambition—disciplined, revenue‑backed scaling is becoming the new benchmark.