Quarterly cash consumption exceeds half of revenue! OpenAI's performance reveals the explosive trend of enterprise AI applications, also exposing the heavy pressure of computing costs.

Media reports cite a document disclosed to shareholders, stating that the developer of ChatGPT, OpenAI, burnt through $3.7 billion in the first quarter of 2026, exceeding about half of the quarterly revenue of $5.7 billion for the period. Just the day before, it was reported that as the ChatGPT developer prepares to go public on the US stock market, its spending in 2025 was around $34 billion, raising questions about the financial sustainability and fundamental growth prospects of this globally highest-valued AI application company among investors. However, from a positive perspective, OpenAI's strong revenue growth data highlights a surge in global demand for AI applications. Nevertheless, the business model of leading AI large model companies has entered a phase of "intense revenue growth, high capital consumption, high depreciation pressure," known as the heavy capital AI arms race, where the so-called "revenue surge and loss black hole" coexist phase. OpenAI continues to burn through cash before the IPO The shareholder document quoted by the media shows that the continuous increase in expenses reflects the company's significant investment in AI computing infrastructure to meet the increasing AI training/inference computing power demands. This includes model training, AI data center construction processes, talent recruitment, and market share expansion. Performance operating data also shows that massive expenditures have increased its net loss from $5 billion in 2024 to around $39 billion in 2025. Despite the high cost base of AI computing power, OpenAI achieved a strong growth pace with approximately $13 billion in revenue during the same period, showing exponential growth compared to the previous years. This growth was mainly due to an increase in subscription paying users for ChatGPT's high-end AI large model services. By the end of 2025, its monthly revenue run rate reached $2 billion, significantly higher than the level of $1 billion per quarter by the end of 2024. Additionally, OpenAI's initial financial data in 2025 exhibits typical features of "super growth and super cash consumption coexisting": with an annual revenue of around $13 billion, a growth of over 2.5 times from 2024's $3.7 billion, and a monthly revenue run rate reaching $2 billion by the end of the year. This underscores the emergence of real business demand for high-end ChatGPT service subscriptions, enterprise APIs, intelligent agent tools, and developer ecosystems. The net loss of $39.5 billion cannot simply be understood as an operating loss out of control, as the majority of it comes from non-cash accounting charges related to the company's structural transformation. During the transition from a non-profit structure to a Public Benefit Corporation (PBC), the convertible equity previously held by investors is typically considered a liability on the accounting books, and is revalued as the valuation increases, resulting in an accounting expense of around $30 billion. Losses related to the change in fair value of convertible equity and warrants total approximately $41.5 billion. Excluding such non-cash items, equity incentives, Microsoft compute resource offset, and other factors, the actual operational loss is around $8 billion. Therefore, OpenAI did not truly burn through $39.5 billion in cash within one year, but even after excluding accounting adjustments, its operating model is still in a phase of intense loss expansion. Multiple media reports quoting insiders reveal that OpenAI aims to IPO on the US stock market by the end of 2026 as soon as possible. Real and strong AI demand, but OpenAI's profit path remains to be seen Combining the previously disclosed expenditures of around $34 billion in 2025, revenue of around $13 billion, and net loss of around $39 billion, of which $30 billion is related to non-cash accounting charges from the old equity structure, adjusted losses of around $8 billion, it is evident that the issue for OpenAI is not whether they have revenue, but whether their revenue growth curve can catch up with the curve of model training, inference, data center, sales expansion, and talent costs. Therefore, for the capital market, OpenAI's IPO will become one of the most critical valuation benchmarks in the super AI bull market. On the positive side, OpenAI has proven that generative AI possesses a rare revenue growth rate, with $13 billion in annual revenue and a $2 billion monthly revenue run rate to support its position as the next generation enterprise AI software, AI developer ecosystem platform, and intelligent agent gateway; the negative aspect lies in the fact that its profit path still heavily relies on the reduction of computing costs, improvement in inference efficiency, penetration of enterprise payments, closure of the intelligent agent business loop, and return on data center capital expenditure. For AI leaders like OpenAI, Anthropic, and Google dominating the AI arms race, these data points do not signal the peak of AI computing demand but rather indicate that leading AI application developers are raising the competition threshold from "AI large model training costs" to a comprehensive AI reasoning warfare involving capital, computing power, electricity, chip supply chain, and data center execution capabilities. The Stargate project led by OpenAI, Oracle, and SoftBank has pushed the US AI data center planning to a capacity of nearly 7 gigawatts over the next three years with investments exceeding $400 billion, continuing to advance towards a goal of $500 billion and 10 gigawatts. For investors currently concerned about the prospects of enterprise AI applications, OpenAI's latest performance data delivers a dual signal: enterprise AI applications are transitioning from pilots to agent-driven AI intelligent workflows, but commercialization must prove that high-frequency usage can high-quality cash flow. OpenAI officials claim that Codex has become one of their fastest-growing enterprise products, with over 4 million developers using it weekly for code reviews, resource allocation, test coverage, incident response, and large codebase reasoning in various software development lifecycle phases. OpenAI is also leveraging consulting and systems integration partners like Accenture, Deloitte, PricewaterhouseCoopers, Infosys, etc., to accelerate Codex's entry into large enterprise workflows.