Newly reported internal projections suggest OpenAI is heading into several years of deep financial losses as it continues to pour unprecedented amounts of money into building and running advanced artificial intelligence systems. According to details surfaced by The Information, the company expects the red ink to continue well into the decade, with profitability only coming into view around 2029.
At the center of these projections is a striking figure: OpenAI is said to be preparing for a loss of roughly $14 billion in 2026 alone. That number is significantly worse than earlier estimates for 2025 and highlights just how expensive it has become to operate at the cutting edge of AI development. When viewed over a longer horizon, the picture looks even starker. From 2023 through the end of 2028, OpenAI reportedly expects cumulative losses of about $44 billion, before finally swinging into profit toward the end of the decade.
Mounting Costs and a Long Road Ahead
The projected losses reflect how quickly OpenAI’s expenses are growing as it scales its technology and infrastructure. Training larger and more capable AI models requires enormous computing power, specialized hardware, and vast amounts of electricity, all of which translate into soaring operational costs. As models become more complex, each new generation becomes significantly more expensive than the last.
At the same time, the report points out an apparent mismatch between OpenAI’s long-term loss projections and its recent spending patterns. While future losses are expected to reach tens of billions of dollars, OpenAI’s cash burn in the most recent reporting period was reportedly lower than many observers anticipated, totaling around $340 million in the first half of the current financial year. The report does not clearly explain how this relatively modest short-term spending reconciles with the massive losses forecast in coming years, leaving open questions about accounting timelines, future commitments, and how quickly costs are expected to accelerate.
What seems clear is that OpenAI’s leadership is planning for a long and expensive runway. The company appears willing to absorb years of financial pain in exchange for what it believes will eventually be a dominant position in the global AI market.
A $200 Billion Bet on Artificial Intelligence
Perhaps the most eye-catching element of the projections is the sheer scale of OpenAI’s planned spending. The company is said to expect total expenditures of around $200 billion by the end of the decade. A majority of that amount, estimated at 60% to 80%, would be devoted to training AI models and running them at scale for users.
These costs span a wide range of necessities, including data centers, high-performance chips, cloud computing resources, and the day-to-day expense of serving AI responses to customers, often referred to as inference. Training new frontier models remains one of the most capital-intensive undertakings in the technology world, and OpenAI’s plans suggest those costs are only going to rise in the near term.
Such figures underline a growing reality in the AI industry: only a handful of companies with access to vast financial resources can afford to compete at the very top. Smaller firms may innovate in niche areas, but building and operating large, general-purpose AI models increasingly looks like a game reserved for the best-funded players.
A Bold Vision for Revenue Growth
Despite years of projected losses, OpenAI’s internal forecasts reportedly outline an exceptionally optimistic future. By 2029, the company’s for-profit operations are expected to generate $100 billion in annual revenue, up dramatically from an estimated $4 billion in 2025. If achieved, that kind of growth would rank among the fastest and largest revenue expansions ever seen in the tech sector.
The projections also suggest OpenAI could record a $14 billion profit in 2029, marking a dramatic turnaround after years of heavy spending. Such a shift would depend on both rapid customer adoption and OpenAI’s ability to maintain strong pricing as competition intensifies.
Comparisons With Nvidia Highlight the Ambition
The scale of OpenAI’s revenue expectations has naturally drawn comparisons with Nvidia, one of the biggest beneficiaries of the current AI boom. Nvidia generated around $130 billion in revenue in 2025, largely thanks to its dominant position in supplying chips used in AI data centers worldwide.
For OpenAI to approach similar revenue levels within just a few years would be extraordinary, particularly given that it operates primarily as a software and services company rather than a hardware supplier. Achieving that goal would require not only widespread adoption of its products but also sustained demand at premium prices in an increasingly crowded market.
Breaking Down OpenAI’s Expected Revenue Streams
The internal forecasts reportedly outline a diversified revenue mix. Just over half of the projected $100 billion in annual revenue is expected to come from ChatGPT, including consumer subscriptions and enterprise offerings. Around 20% is forecast to come from API access, where developers pay to integrate OpenAI’s models into their own products and services. The remaining share would come from other products, such as video generation tools, AI-powered search, and potential new services like research assistants.
This mix suggests OpenAI is betting on AI becoming deeply embedded across many industries, rather than relying on a single blockbuster product.
Signs of Cost Relief on the Horizon
There are also indications that some costs could ease over time. OpenAI reportedly expects the cost of inference to decline as hardware becomes more efficient and software improves. The company also anticipates spending less on training data, with annual data acquisition costs projected to fall from about $500 million this year to roughly $200 million later in the decade.
How OpenAI plans to achieve these reductions is not entirely clear, but one possibility is increased use of AI-generated or synthetic data in training future models.
