Leaked financials expose OpenAIs explosive growth and staggering cash burn

Tripled revenues and a $21 billion loss reveal the staggering financial reality of OpenAI's high-stakes race for market dominance.

June 20, 2026

Leaked financials expose OpenAIs explosive growth and staggering cash burn
The financial mechanics powering the artificial intelligence revolution have long been shrouded in mystery, but newly disclosed shareholder documents have exposed the staggering scale of OpenAI's cash consumption and explosive growth[1][2]. In the first quarter of the year, the creator of ChatGPT generated an impressive $5.7 billion in revenue, representing a tripling of its sales compared to the same period in the prior year[1][2]. However, the cost of capturing that market share was equally historic. OpenAI burned through approximately $3.7 billion in cash during those same three months, a figure that also tripled year-over-year[1][2]. This symmetry highlights the core paradox of the modern artificial intelligence industry: revenue growth and operational expenses are currently scaling in near-perfect lockstep, challenging traditional software-as-a-service business models and testing the patience of late-stage investors[1].
The parallel tripling of both income and cash burn highlights a structural characteristic of generative artificial intelligence that distinguishes it from previous technology booms[1][2]. In traditional software markets, companies benefit from massive operating leverage where serving an additional customer costs virtually nothing once the core platform is built[1]. For OpenAI, however, every query processed by its models requires significant computing power, meaning that a larger user base directly translates to a larger computing bill[1][3]. Despite these ongoing constraints, the company's financial records did reveal some efficiency gains. OpenAI reported gross margins of 39% for the first quarter, a notable improvement from 33% during the same period in the previous year[4][5]. This margin expansion was achieved even as the direct cost of revenue, primarily the cost of serving its frontier models, reached $3.5 billion for the quarter, suggesting that hardware optimization and engineering breakthroughs are starting to shave down individual query costs[5].
While gross margins showed a positive trajectory, the broader picture of profitability remains deep in the red, heavily weighed down by compensation and corporate restructuring[6][7]. OpenAI posted an operating loss of $9.3 billion in the first quarter, a figure heavily impacted by $2.3 billion in stock-based compensation for its employees, which more than doubled year-over-year as the company competed to recruit and retain elite engineering talent[5][7]. Furthermore, the net loss for the quarter reached a staggering $21.3 billion[8][7]. However, a significant portion of this net loss was not a direct cash drain; $12.4 billion was attributed to a non-cash accounting charge resulting from the revaluation of investor rights as OpenAI transitions away from its original non-profit governance structure[8][7]. This trend of monumental losses mirrors the company's performance in the previous fiscal year, when it recorded $13.1 billion in revenue against $34 billion in total expenses, culminating in an annual net loss of roughly $39 billion[9].
Despite the eye-watering losses, OpenAI's balance sheet remains highly resilient, insulated by an unprecedented war chest of private capital[6]. The company concluded the first quarter with more than $73 billion in cash and marketable securities, a dramatic rise from the $40 billion held at the end of the previous fiscal year[2]. This cash cushion was primarily fortified by a blockbuster private funding round that valued the company at $852 billion[1][2]. With such massive reserves, OpenAI is under very little near-term pressure to seek fresh capital, which could allow leadership to delay its highly anticipated public market debut[2]. OpenAI has already filed confidentially for an initial public offering in the United States, with some analysts predicting a listing that could value the company at up to $1 trillion, making it one of the largest public market debuts in financial history[1][10].
Looking forward, OpenAI's financial runway will face severe tests from both long-term infrastructure commitments and intensifying competition[6][4]. The shareholder documents reveal that the startup faces a jaw-dropping $665 billion in computing purchase commitments through the end of the decade, largely locked into partnerships with infrastructure giants like Microsoft, Oracle, and Nvidia[11][3][5]. At the same time, maintaining its market leadership requires aggressive pricing strategies[6]. Reports suggest that OpenAI is preparing to slash its API prices to squeeze its primary rival, Anthropic, which has recently faced developmental hurdles of its own[6][4]. While a price war could successfully defend OpenAI's market dominance, it threatens to erode the hard-won gross margin improvements achieved in the first quarter and push the company’s projected timeline for profitability even further into the future[1][6].
Ultimately, OpenAI’s leaked financial performance represents the ultimate high-stakes gamble in modern technology history. The company is actively betting that artificial intelligence will achieve a level of capability and ubiquity that will easily justify hundreds of billions of dollars in infrastructure investments and years of deep operating losses[1][3]. If OpenAI succeeds in pioneering artificial general intelligence and locking in enterprise dominance, its current multi-billion-dollar losses will eventually look like a bargain. However, if the technology's utility plateaus or if intense competition permanently depresses pricing power, OpenAI's aggressive burn rate could become a cautionary tale of speculative excess. For now, public markets and private backers alike are watching closely to see if the world’s most prominent artificial intelligence startup can translate historic growth into a sustainable corporate future.

Sources
Share this article