By Lina Irawan 
My week began with a vivid illustration of AI’s burgeoning wealth. I spoke with two former OpenAI employees who collectively pocketed approximately $10 million in a single day. This windfall came from selling shares in an internal tender offer, a mechanism increasingly common in high-growth, pre-IPO tech companies to provide liquidity to early employees and investors. As reported by Bloomberg on August 10, 2026, this particular tender offer totaled an astonishing $7 billion across OpenAI’s staff, signaling immense confidence in the company’s future trajectory and rewarding its foundational talent handsomely. Such figures are a testament to the perceived value locked within these cutting-edge AI ventures.
The financial fireworks continued. This morning, on Fortune’s weekly AI podcast, my colleague Beatrice Nolan and I had the privilege of interviewing the CEO of Lovable. This Stockholm-based "vibe-coding" startup, a mere three years old, announced a monumental achievement: it had doubled its valuation to an eye-watering $13.3 billion this week. Lovable, which specializes in leveraging AI to understand and generate content based on emotional nuances and user sentiment – essentially coding for human "vibes" – exemplifies the rapid and often exponential growth seen in niche, yet highly impactful, AI applications. Their recent $400 million funding round, as reported by The Wall Street Journal, is earmarked for aggressive expansion, including scaling their workforce and deepening their presence in the Latin American market, alongside a commitment to robust safety and security protocols.
These individual success stories, however impressive, are dwarfed by even grander projections. Anthropic, another major player in the AI lab space, is reportedly planning an initial public offering (IPO) that could value the company at an unprecedented $2 trillion, potentially launching as early as October 2026. If realized, this would shatter all previous IPO records, eclipsing even SpaceX’s historic debut. This staggering valuation is predicated on investor belief that Anthropic’s annual revenue run rate could reach between $100 billion and $120 billion by the end of the year. Such figures – a couple million, a hundred billion, a trillion (or two) – have become the new normal in the AI lexicon, creating a sense of dizzying detachment from traditional financial metrics.
Yet, amidst this torrent of capital and sky-high valuations, a significant, often overlooked, challenge persists: The tangible return on investment (ROI) for companies adopting AI solutions remains largely unclear. This existential question isn’t just a concern for cautious CFOs; it’s one that even OpenAI, at the forefront of the AI revolution, is grappling with.
On August 11, OpenAI published a comprehensive 69-page report titled "How Organizations Use ChatGPT," delving into the enterprise adoption of its flagship AI model. On the surface, the report presents an optimistic narrative of explosive AI usage across all seniority levels and job functions. It introduces the concept of a "frontier gap," positing that companies actively integrating AI are rapidly outperforming those that aren’t. The implication is clear: businesses that fail to leverage AI, particularly advanced "agents" capable of automating complex tasks, risk falling irrevocably behind. This narrative fuels the urgency and competitive pressure driving AI adoption today.
However, a closer examination of the report’s "fine print" reveals a more nuanced and, frankly, sobering reality.
More AI Doesn’t Necessarily Mean More Money
Buried within a small table on page 35, the OpenAI researchers present a critical finding: there is no statistically significant correlation between a company’s revenue per employee and the extent to which those employees utilize AI, whether measured by messages sent or tokens consumed. The report explicitly states, "Revenue per employee is not meaningfully associated with output tokens per employee or messages per active user once other controls are included."
This finding is a significant counterpoint to the prevailing hype. While the study acknowledges that companies with higher existing revenue per employee tend to be earlier adopters of ChatGPT, and generally, companies that use the technology more also tend to have higher revenue per employee, it stops short of establishing a causal link. In essence, it suggests that successful companies are more likely to jump on the AI bandwagon, but the report doesn’t conclusively prove that increased AI usage directly leads to increased revenue. This highlights a crucial challenge for businesses: how to translate AI adoption into measurable financial gains beyond mere efficiency or cost-cutting. Many experts suggest that the true ROI of AI might be long-term, requiring significant strategic overhaul and integration, making immediate correlations difficult to establish. It could also be that current AI applications are primarily enhancing productivity in ways that are hard to quantify directly in revenue per employee, or that the market is still too nascent to see these effects clearly.
Executives Are Using AI the Least
Another intriguing insight from the report, tucked away on page 29, reveals a significant discrepancy in AI usage across different organizational levels. Executives, the very individuals responsible for strategic decisions and investment in AI, are using it the least. It’s not merely a matter of fewer executives compared to general employees; the data shows that senior leaders engage with AI less intensely, measured by weekly messages per user.
This finding carries profound implications. If top-level decision-makers are not directly experiencing the capabilities and limitations of AI, their ability to accurately gauge its ROI, formulate effective AI strategies, and champion its widespread adoption could be compromised. They might be relying on second-hand information or generalized enthusiasm rather than hands-on understanding. In contrast, early-career employees exhibit the highest usage, a point emphasized by OpenAI CFO Sarah Friar in her LinkedIn post about the report. She wisely noted, "For leaders, that’s a reminder that competitive advantage comes from the people closest to the work. Listen to them, learn from them, and help the rest of the organization catch up." This underscores the importance of bottom-up innovation and empowering the workforce to drive AI integration, rather than solely relying on top-down mandates. Bridging this "usage gap" between leadership and the frontline workforce will be critical for effective AI transformation.
OpenAI’s Enterprise Sales Hit a Rough Patch in Q4 2025
The report also sheds light on a surprising dip in OpenAI’s enterprise usage. A graph on page 26, depicting output token growth, shows that overall usage within enterprises completely flatlined from approximately October to December 2025. This period coincides with a surge in competition, particularly from Anthropic’s Claude Code, which was rapidly gaining traction and becoming a preferred platform in many corporate environments due to its specific strengths, perhaps in areas like code generation or longer context windows. This flatlining suggests a period of significant challenge for OpenAI in retaining and growing its enterprise footprint against aggressive rivals.
To the company’s credit, the narrative quickly shifts. In January 2026, the line depicting usage thrusts upward, transforming into an exponential curve. As one venture capitalist shared with me yesterday, "OpenAI’s run rate in 2026 has been pretty incredible." OpenAI attributes this remarkable resurgence not only to securing new clients but also to existing clients deepening their engagement and expanding their use cases. This period of renewed growth aligns with CEO Sam Altman’s strategic reorganization of the company, which has seen a significant pivot towards enterprise sales. Part of this shift involved "killing what the company called ‘side quests’," such as the experimental video generation app Sora, to re-focus resources on core enterprise products and revenue generation.
In a clear sprint to accelerate this upward trajectory – or perhaps to regain lost momentum, as the graph conveniently concludes at March 2026 – OpenAI today announced a significant leadership change. Dali Rajic has been hired as the new Chief Revenue Officer, replacing Denise Dresser, who held the role for less than a year. This aggressive move, with Dresser reportedly departing on her own accord to pursue other ventures, is widely interpreted as a direct response to the intense pressure to demonstrate sustained growth ahead of a potential IPO. Rajic’s mandate is explicitly defined: accelerate customer adoption and, crucially, help businesses measure the impact of AI, directly addressing the core ROI challenge that permeates the industry. This move also highlights the growing influence of co-founder and president Greg Brockman in shaping the company’s strategic direction as it gears up for a public offering.
OpenAI Paid the Academics Who Contributed to the Report
Finally, a subtle but important detail concerning the report’s credibility. Two of the five authors listed are academics, affiliated with Columbia Business School (David Holtz) and Wharton at the University of Pennsylvania (Prasanna Tambe). Typically, the inclusion of external academic researchers in such a paper lends an air of greater credibility and impartiality, suggesting an independent, rigorous analysis. However, a footnote on the first page clarifies that both academics "contributed to this work in their capacity as paid contractors for OpenAI."
This contractual relationship inevitably muddies the waters regarding the report’s objectivity. While it doesn’t automatically invalidate the findings, it raises questions about potential biases and whether the researchers might have been influenced by the interests of their paying client. In traditional academic research, findings are often presented in their entirety, regardless of whether they fully support a particular narrative. The question lingers: Did the researchers uncover additional findings or nuances that were not published, as they typically would for an independent academic paper? Without full transparency, we’ll never know. This serves as a potent reminder that in the fast-paced, hype-driven world of AI, businesses and investors must exercise critical judgment. The true impact of AI, especially its ROI, will ultimately need to be measured based on individual, first-hand experience and rigorous internal evaluation, rather than solely relying on external reports, particularly those with potential conflicts of interest. The allure of the trillion-dollar valuations must not overshadow the need for tangible, verifiable benefits.
FORTUNE ON AI
- Microsoft begins to merge consumer and enterprise Copilot apps in push for super app — by Sebastian Herrera: Microsoft’s strategic move to unify its Copilot offerings underscores a broader industry trend towards creating comprehensive, integrated AI ecosystems that cater to both individual and corporate users, aiming for seamless functionality across all platforms.
- CoreWeave CEO Michael Intrator cites ‘sold out’ capacity as revenue more than doubles and backlog swells to $104 billion — by Amanda Gerut: The insatiable demand for specialized GPU infrastructure to power AI development is driving astronomical growth for companies like CoreWeave, highlighting the critical bottleneck in AI expansion and the immense capital flowing into hardware.
- When OpenAI employees have a problem, they email this special address to see if Sam Altman will solve it immediately — by Emily Forlini: This anecdote reveals a unique, perhaps unconventional, internal communication channel at OpenAI, hinting at Sam Altman’s hands-on leadership style and an organizational culture that prioritizes rapid problem-solving, even at the highest levels.
- CIOs and CTOs spent years lauding AI. Now, with costs rising, they’re putting limits on how it’s used — by John Kell: As the initial enthusiasm for AI matures, corporate leaders are facing the reality of escalating operational costs, prompting a more disciplined and strategic approach to AI deployment, focusing on specific use cases with clearer ROI.
- Forget the gala, these Silicon Valley schools run their own venture capital funds — by Amanda Gerut: This innovative trend in Silicon Valley education showcases how institutions are directly engaging with the startup ecosystem, not just as academic incubators but as active financial participants, blurring the lines between academia and venture capitalism.
AI IN THE NEWS
- Anthropic plans $2 trillion IPO, possibly in October. It would be the largest ever on record, eclipsing SpaceX, but the valuation and IPO timing are not yet finalized. Investors believe the AI lab’s annual revenue run rate will be between $100 billion and $120 billion by the end of the year. The ambitious target reflects an aggressive bet on the future dominance of foundational AI models, yet market analysts caution that such projections, while exciting, often face intense scrutiny as companies transition from private to public markets. More in The Financial Times.
- Lovable raises $400 million, doubles valuation to $13.3 billion. The vibe-coding startup is only three years old. It plans to use the funds for expansion, including growing its workforce and expanding its Latin American customer base. It also plans to implement strong safety and security. This emphasis on safety is particularly relevant for a company dealing with nuanced human emotion and communication, underscoring the increasing regulatory and ethical considerations in AI development. More in The Wall Street Journal.
- OpenAI’s chief revenue officer is out after less than a year. The company has hired Dali Rajic to replace Denise Dresser, who says she left on her own accord to pursue other ventures. Many interpret the move as coming from the growing presence of co-founder and president Greg Brockman as the company gears up for a potential IPO. The swift executive change highlights the intense pressure for performance and revenue growth as OpenAI positions itself for a monumental public offering. More from OpenAI.
EYE ON AI NUMBERS
41
Lifespan of OpenAI CEO Sam Altman, according to Google’s botched algorithm. On Wednesday, the search giant proclaimed Altman dead, adding a death date of August 12, 2026, to the fact panel it displayed at the top of its search results. That same panel paradoxically stated Altman was still the current CEO of OpenAI. While seemingly absurd, this incident serves as a stark reminder of the fragile and often vulnerable nature of the internet’s information ecosystem, even for sophisticated AI-powered search engines.
The erroneous report originated from a vandalized Wikipedia page for Altman, where an unknown malefactor had rewritten the entire introduction in the past tense, falsely claiming he was assassinated in Seattle. According to Gadget Review, Wikipedia’s automated tools eventually flagged the malicious change, and a human editor corrected it 41 minutes later – a minute for each of Altman’s supposed years of life.
Google promptly responded on X (formerly Twitter) after the incident, stating, "When people vandalize public info sources, this can affect the information that appears in Search." This explanation, while technically true, ignited a debate about accountability. Some critics pointed fingers at Wikipedia for its open-editing model, while others argued that Google, as the primary aggregator and presenter of information, bears responsibility for its algorithms’ inability to swiftly verify or cross-reference critical biographical details, especially for high-profile individuals. It remains unclear whether Google’s internal safeguards would have detected and corrected the misinformation on its own within that crucial 41-minute window, highlighting the ongoing challenge of ensuring accuracy and preventing the spread of misinformation in an increasingly AI-driven information landscape. The incident underscores the critical need for robust fact-checking mechanisms and human oversight, even as AI aims to automate and streamline information delivery.
AI CALENDAR
Nov. 16-17: Fortune 500 Innovation Forum, Detroit. Apply here to attend.
Dec. 6-12: Neural Information Processing Systems (Neurips) conference. Sydney, Australia.
Dec. 7-8: Fortune Brainstorm AI, San Francisco. Apply here to attend.
