Welcome to Eye on AI. Emily Forlini here, filling in for Jeremy one last time as his vacation comes to a close. In today’s issue:
- Juicy details OpenAI doesn’t want you to see in its new report
- Anthropic reportedly plans a $2 trillion IPO in October—the largest ever
- OpenAI replaces its chief revenue officer after less than a year
- Google pronounces Sam Altman dead—for 41 minutes
It really sunk in for me this week just how much money is flowing in the AI industry.
I spoke with two former OpenAI employees who made about $10 million in a day by selling shares in an internal tender offer, which Bloomberg reports totaled $7 billion across the staff. Then, this morning, on Fortune‘s weekly AI podcast, my coworker Beatrice Nolan and I interviewed the CEO of Lovable. This week, the old Stockholm-based firm, which is only three years old, doubled its valuation to $13.3 billion.
A couple million, a hundred billion, a trillion (or two, in the case of Anthropic’s upcoming IPO)—what’s the difference at this point? There’s just one big problem looming in the background: The ROI of AI adoption is still not clear for companies.
OpenAI grapples with this existential question in a 69-page report published on August 11 on the enterprise adoption of ChatGPT. On its face, the report tells the a story of exponential AI usage growth across all seniority levels and job functions, highlighting what it calls a “frontier gap,” in which companies who are using AI are pulling ahead of those who aren’t. In other words, if you’re not using AI—especially agents you can delegate tasks to—you’re losing.
But the fine print tells a different story.
More AI doesn’t mean more money
In one small table on page 35, the researchers report no statistically significant correlation between the revenue per employee, and how much those employees use AI, measured in messages sent and tokens used.
“Revenue per employee is not meaningfully associated with output tokens per employee or messages per active user once other controls are included,” the report explains.
It says that companies that already have higher revenue per employee tend to be early ChatGPT adopters. Also, companies that use the tech more tend to have higher revenue per employee in general. In other words, big, lucrative companies are more likely to have hopped on the AI train. However, the study doesn’t clearly establish that the more AI they use, the more money they make.
Executives are using AI the least
Executives may not be best equipped to gauge ROI because they are using it the least—another nugget buried in the report. It’s not just that companies have fewer executives than they do general employees. But what’s interesting about the graph on page 29 is that most senior employees are using it less intensely, with the least weekly messages per user.
Early career employees have by far the most usage, a point OpenAI CFO Sarah Friar highlighted in her LinkedIn post about the report: “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.”
OpenAI’s enterprise sales had a rough Q4’2025
Surprisingly, OpenAI’s overall usage within enterprises completely flatlined from about October 2025 to December 2025. In a graph (page 26) depicting output token growth, the black line representing “total” growth is almost perfectly flat for that time period. During this time, Anthropic’s Claude Code was taking the corporate world by a storm, becoming the go-to platform at many places.
To the company’s credit, in January 2026, the line thrusts upward into an exponential curve. As one VC told me yesterday, “OpenAI’s run rate in 2026 has been pretty incredible.” OpenAI attributes the growth not only to adding new clients, but also also to its current clients deepening their use. We also know CEO Sam Altman has been reorganizing the company around enterprise sales, and killing what the company called “side quests,” such as the video app Sora.
In a sprint to accelerate this line—Or, maybe to get it going again? Who knows, the graph ends at March 2026—OpenAI today announced it hired a new Chief Revenue Officer, Dali Rajic, who will replace Denise Dresser. It’s an aggressive move; Dresser held the role for less than one year. Rajic’s focus will be accelerating customer adoption and helping businesses measure impact as the company sprints towards its IPO.
OpenAI paid the academics who contributed to the report
Lastly, two of the five authors are academics that OpenAI paid to help with the report. The other three are OpenAI employees. Including academics in a paper like this typically implies greater credibility and the impartiality of an outside research institution, but the waters are a little muddier here.
On the first page, David Holtz and Prasanna Tambe are affiliated with Columbia Business School and Wharton at the University of Pennsylvania, respectively. But a footnote specifies that both “contributed to this work in their capacity as paid contractors for OpenAI.”
Did the researchers find more that they didn’t publish, as they typically would for an academic paper? We’ll never know, but it’s another reminder of what has always been true: You’ll have to measure AI’s impact based on your own first-hand experience—not the hype.
With that, here’s more AI news.
Emily Forlini
[email protected]
@emilyforlini
FORTUNE ON AI
Microsoft begins to merge consumer and enterprise Copilot apps in push for super app — by Sebastian Herrera
CoreWeave CEO Michael Intrator cites ‘sold out’ capacity as revenue more than doubles and backlog swells to $104 billion — by Amanda Gerut
When OpenAI employees have a problem, they email this special address to see if Sam Altman will solve it immediately — by Emily Forlini
CIOs and CTOs spent years lauding AI. Now, with costs rising, they’re putting limits on how it’s used — by John Kell
Forget the gala, these Silicon Valley schools run their own venture capital funds — by Amanda Gerut
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. 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. 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. 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 said Altman was still the current CEO of OpenAI. Hard to do from the grave, but we’ve seen crazier things in AI. Maybe it’s a premonition for future CEOs who carry on through AI avatar reproductions of themselves.
Someone had apparently vandalized Altman’s Wikipedia page, claiming he was assassinated in Seattle. The malefactor re-wrote the entire introduction to Altman’s Wiki entry in the past tense, according to Gadget Review. Wikipedia’s automated tools flagged the change, and a human editor fixed it 41 minutes later—one minute for each of Altman’s supposed years of life.
Google responded on X after the incident, saying, “When people vandalize public info sources, this can affect the information that appears in Search.” It’s a silly, but stark, reminder of how fragile the internet’s information ecosystem is. Some blame Wikipedia for allowing rogue editors, while others say Google’s on the hook for choosing to display inaccurate information. It’s unclear if Google would have caught the issue on its own, in 41 minutes or less.
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.
