Listen to the audio version of this article (generated by AI).<\/p>\n
Key Points<\/p>\n
OpenAI\u2019s second-quarter results showed continued heavy cash burn, with a $12.3 billion loss even as revenue increased just 18% quarter over quarter.
\n Token prices have fallen sharply over the past three months, roughly halving and putting further pressure on OpenAI\u2019s path to profitability.
\n OpenAI remains central to the broader AI trade, but it must continue raising capital or companies that depend on its spending could face significant risk.<\/p>\n
OpenAI released some high-level financials from the company\u2019s second quarter to investors, and the results were jaw-droppingly terrible. The results are so poor that they should call into question the viability of the artificial intelligence (\u201cAI\u201d) company and perhaps the entire standalone AI model business.<\/p>\n
For the quarter, OpenAI reported a loss of $12.3 billion, including stock-based compensation, widening from the first quarter\u2019s $9.3 billion loss. That\u2019s an enormous loss, of course, but what may be perhaps more concerning for investors is the quarter\u2019s relatively slow sales growth.<\/p>\n
Sales climbed 18% to $6.7 billion in the second quarter, up from $5.7 billion. The nearly 20% quarter-over-quarter gain would look fantastic for many non-AI companies. Still, it looks mediocre as key rival Anthropic doubled its revenue in the same period to $11.6 billion.<\/p>\n
And just so the message is not lost: OpenAI\u2019s margins actually worsened as it grew. The extra $1 billion in revenue in the second quarter led to an incremental loss of about $3 billion. While it\u2019s fair for OpenAI to run losses as it reinvests in its business, the actual fundamentals are worsening by the week with token prices sinking.<\/p>\n
That\u2019s a problem that might not go away anytime soon, as the company has slashed the price of recently released high-end models by up to 80% to compete. The move threatens to accelerate a price war that low-cost Chinese AI labs are already well-positioned to win.<\/p>\n
If all those issues aren\u2019t enough turmoil, OpenAI is also suffering from a mass exodus of key executives this year \u2013 plenty more reason to be skeptical of the company\u2019s staying power.<\/p>\n
At least 12 executives have left the company in 2026, including operating chief Brad Lightcap, applications CEO Fidji Simo, and chief revenue officer Denise Dresser. Dresser had been with OpenAI since December 2025, making such a quick departure worrisome.<\/p>\n
All this turmoil in a company that a year ago was hailed as the leading AI player. Now, with massive and growing operating losses, OpenAI must keep raising funds to survive. And that\u2019s not a one-off \u201craise cash and we\u2019re golden\u201d kind of moment. The structural lack of profits in the AI model business means that OpenAI has to go to the funding well to plug the massive hole it has dug for itself.<\/p>\n
Eventually, investors won\u2019t keep writing $100 billion-plus checks to keep OpenAI afloat, as they did with its $122 billion capital raise in March<\/a>. Worse, the cost of keeping OpenAI afloat may keep rising as token pricing plummets, an accelerating trend since May.<\/p>\n If OpenAI doesn\u2019t survive, it will renege on hundreds of billions of dollars in spending across the AI industry, putting other highly leveraged players such as Oracle (ORCL) and CoreWeave (CRWV)<\/a> in jeopardy.<\/p>\n AI Token Pricing Plummets, Hurting OpenAI and Industry Revenue<\/p>\n One of the key issues plaguing OpenAI and other AI labs is the plummeting price of tokens that are used to pay for AI. They\u2019re how model companies get paid for their work. So if token prices fall, labs such as OpenAI generate less revenue even with the same operating costs.<\/p>\n What we\u2019re seeing in token prices over the past three months is devastating, at least according to Silicon Data\u2019s LLM Token Expenditure Index. This index tracks what AI users are paying per million tokens, including open models as well as closed proprietary models such as OpenAI\u2019s.<\/p>\n The index peaked in late May, reaching more than $2 per million tokens. Since then, it has been almost all downhill. By late August, the index was about $1.02 per million tokens, effectively cut in half in just three months.<\/p>\n The index is showing quantitatively what we\u2019re seeing qualitatively. A falling index shows that cheaper, lower-end models are capturing market share. That fits well with reports that low-cost, open Chinese AI models are grabbing share, notably even in the critical enterprise sector.<\/p>\n One conclusion is that it will be tough for makers of high-end, frontier models such as OpenAI and Anthropic to compete profitably, especially as AI outputs become even more commoditized.<\/p>\n The eagle-eyed may also notice that this index\u2019s token pricing peaked just days before OpenAI and Anthropic announced that they had filed confidentially to go public in early June<\/a>. While it may be a coincidence, it looks almost as if the companies timed the move. As their businesses looked increasingly unable to self-fund, they accelerated their path to go public.<\/p>\n OpenAI Is Key to the Whole AI Trade<\/p>\n OpenAI must have funding to survive. If it doesn\u2019t get this funding, it\u2019s not only OpenAI that will go bust. It will be all the companies whose businesses are built on getting paid by OpenAI.<\/p>\n Now, at this point, some AI bulls will see plummeting token prices and invoke what\u2019s known as Jevons Paradox. It\u2019s the idea that innovations that raise the efficiency of a process actually increase their use over time. So, the greater the efficiency of oil extraction, the more oil is used.<\/p>\n That\u2019s an okay point if you\u2019re talking about the use of AI as a whole, just as it was for Internet usage in from 1999 to 2000. Service quality moved from the poky dial-up speeds of America Online, which initially charged per-hour rates to unlimited high-speed Internet \u2013 and web use multiplied.<\/p>\n But it didn\u2019t multiply without plenty of uneconomical business models going bust in the process.<\/p>\n AI will get cheaper and more efficient over time \u2013 we\u2019re seeing it right now with the shift to open models and Chinese models. But that doesn\u2019t mean that any individual company will make it through, especially ones with a high-cost delivery model offering a commoditized service.<\/p>\n If OpenAI gets in serious trouble, funding for the whole AI sector would likely seize up in terror.<\/p>\n OpenAI has huge spending commitments to dozens of other companies \u2013 notably Oracle, but also Cerebras Systems (CBRS), where OpenAI makes up the majority of near-term revenue<\/a>. Any significant disruption to OpenAI will spill over to these firms, as well as the various AI data centers backed by hyperscalers. Even the Big Tech companies here are not immune.<\/p>\n AI may (or may not) become as pervasive and world-changing as the Internet, but it may take down several big companies along the way. If investors aren\u2019t willing to perpetually fund the unending losses \u2013 spoiler alert: they never are \u2013 the AI trade will quickly become a dicey proposition.<\/p>\n Believing in the future of AI as a service doesn\u2019t mean any single AI business is a good investment.<\/p>\n Regards,<\/p>\n James Royal, PhD<\/p>\n Editor\u2019s Note: Sam Altman is officially launching a new venture\u2026 and it could be 500 times bigger than ChatGPT, unlock an industry worth more than $1 million per American, and send one small group of stocks soaring as this news spreads. Click here to learn about the stocks tied to Sam Altman\u2019s Next Venture<\/a>.<\/p>\n","protected":false},"excerpt":{"rendered":"Listen to the audio version of this article (generated by AI). Key Points OpenAI\u2019s second-quarter results showed continued…\n","protected":false},"author":2,"featured_media":146678,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[7],"tags":[24,1986,157,305],"class_list":["post-146677","post","type-post","status-publish","format-standard","has-post-thumbnail","category-openai","tag-ai","tag-ai-stocks","tag-openai","tag-tech-stocks"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/146677","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=146677"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/146677\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/146678"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=146677"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=146677"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=146677"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}