Docs: Anthropic expects to break even in 2028, while OpenAI projects ~$74B in operating losses, or ~75% of revenue, that year before turning a profit in 2030
Financial documents from both companies show the different approaches they are taking to the AI boom
Context & Ripple Effects
Earlier documents had OpenAI targeting profitability in 2029, alongside sharply rising projected losses; this report extends that long-horizon profitability path by another year. Anthropic, meanwhile, had projected lower cash burn as it pursued revenue growth, providing a useful contrast in financing posture. Anthropic's earlier burn-reduction outlook makes the new break-even target more consequential.
The disclosures put the commercial debate less on near-term revenue alone and more on the duration and scale of losses each company expects to carry while building and serving frontier models.
First-order effects
- Anthropic’s 2028 break-even expectation gives it a materially earlier projected path to self-sustaining operations than OpenAI’s 2030 target.
- OpenAI’s projected 2028 operating loss of roughly $74B makes continued access to large-scale financing central to its stated growth plan.
Second-order effects
- Investors and strategic partners are likely to compare AI companies more closely on operating-loss duration and cost discipline, not just revenue growth.
- The gap raises pressure on both companies to show that training and inference spending can be converted into durable margins; prior disclosures already indicated inference costs exceeding half of revenue in their projections.
Third-order effects
- If these divergent timelines persist, frontier AI competition may increasingly favor firms with either faster cost absorption or deeper, more patient capital pools.
- Profitability forecasts will become a key test of the AI infrastructure-finance model, though projections remain sensitive to future compute costs, demand, and model-development spending.
The trend: Frontier AI is shifting from a race to scale toward a contest over who can finance compute-intensive growth long enough to reach sustainable margins.