OpenAI disclosed a striking set of metrics as it frames 2026 as a year to move from fascination to utility. In a long post, Chief Financial Officer Sarah Friar said the company’s 2025 annual recurring revenue topped $20 billion and that weekly and daily active user numbers reached new highs. Friar presented a simple causal loop: large investments in compute fuel breakthroughs in research, which produce more capable models, better products and wider adoption, which in turn funds the next wave of compute.
The financial headline is impressive, but the rest of the picture is more ambiguous. OpenAI reports that compute capacity grew threefold in 2025 and nearly 9.5 times since 2023, while revenue rose roughly tenfold over the same period. The company, however, is not yet consistently profitable: Microsoft has hinted that OpenAI recorded substantial losses in 2025, and the firm is negotiating major fundraising and long-term compute deals that industry sources say exceed a trillion dollars in aggregate.
Friar framed the company’s near-term objective as closing the gap between what advanced AI can do and how people, firms and governments actually use it. She singled out health care, scientific research and enterprise applications as especially fertile ground, arguing that smarter systems can translate directly into better outcomes and measurable value. Friar also defended recent product decisions, including advertising in ChatGPT, as attempts to turn exploration into action in ways that add user value rather than detract from it.
The scale of OpenAI’s ambition has attracted equally large financial commitments and risks. The firm has been reported to be in talks to raise as much as $100 billion, while SoftBank has committed a headline $40 billion investment and Microsoft remains a pivotal partner and customer. CEO Sam Altman has repeatedly described OpenAI’s infrastructure spending as a strategic ‘‘gamble’’—one that treats compute as the lever that ultimately determines the company’s revenue ceiling because inference costs will eventually overtake training costs.
That calculus has consequences far beyond OpenAI’s balance sheet. A company that ties its growth to exponential increases in computation places enormous demand on chips, data‑centre capacity and energy. It also concentrates bargaining power with cloud and hardware suppliers, while intensifying competition among a small set of big tech players and investors willing to underwrite an arms race in infrastructure.
The path to monetisation is diversifying. Friar argued that as AI becomes habitual, platforms gain predictable economic value and can support new pricing models. Beyond subscriptions and API fees, OpenAI is experimenting with advertising and exploring value‑share models tied to domain‑specific improvements in science, drug discovery, energy systems and finance. These are high-margin possibilities but require deeper integration into specialised workflows and long sales cycles.
Public markets loom as the ultimate test of that strategy. Altman has suggested an IPO could come in 2027, and media reports have speculated about a private valuation approaching $1 trillion. If OpenAI were to list at that scale, it would crystallise a new class of market expectations for AI companies and for the compute investments that underpin them. If it fails to translate investment into durable profits, it could become a cautionary example of mispriced infrastructure gambles.
For global audiences, the story is about more than one company’s ledger. OpenAI’s model—massive upfront capital for compute, rapid productisation and bets on downstream monetisation—has become a template for an industry. Governments, investors and corporate customers must weigh the upside of faster scientific and commercial innovation against the systemic risks of concentrated compute demand, geopolitical supply‑chain strain for chips, and the fiscal fragility of firms that burn cash while chasing scale.
In short, OpenAI has turned its capabilities into substantial revenue, but its strategy remains a high‑stakes wager on continued demand, favourable financing and the ability to convert broad technical advances into reliable, high‑value applications.