Pay Us Now, Get Rich Later

OpenAI is going after private equity. Similar to the “better” deal the ChatGPT-maker is offering the Pentagon, CEO Sam Altman is reportedly dangling a “guaranteed minimum return of 17.5%” to big private equity firms like TPG, Bain Capital, Advent International, and Brookfield Asset Management.

OpenAI, like Anthropic, is pitching a joint venture aimed at “accelerating adoption,” of enterprise AI products, per Reuters. They are hoping to raise about $4 billion at a pre-money valuation of approximately $10 billion.

The deal is structured via a joint venture. That means OpenAI is only offering the above terms if the firms roll out their AI tools within their hundreds portfolio companies, which could be a massive market opportunity. In other words, they’re offering the picks and shovels now in exchange for money, with the idea of giving them gold later.

TPG and Advent will also reportedly getting early access to OpenAI’s newest models as a deal sweetener. But it’s important to note that OpenAI is burning cash, up to $665 billion per estimates, so these promises may be fantastical.

If you zoom out, this is just the latest chapter in the circular financing playbook that Nvidia, Oracle, and SoftBank have been running for years. OpenAI is trying to hook investors with a return on a future investment that’s only worth anything if those same investors start pushing the product they just bought.

The problem with this strategy is they are creating a market and demand that doesn’t yet exist, and forcing this software down the throats of PE firms and others as the IPO buzzer ticks downward. If you’re Microsoft, which has OpenAI representing some 45% of its entire commercial backlog, you’re worried about this almost too-good-to-be-true kind of deal.

OpenAI COO Brad Lightcap made no secret of the company’s intention to dominate the more lucrative enterprise business, in a bid to capture market share from rival Anthropic.

Our analysts just identified a stock with the potential to be the next Nvidia. Tell us how you invest and we’ll show you why it’s our #1 pick. Tap here.

“One of the interesting things and some of the inspiration for the work we’ve been doing lately around OpenAI Frontier is we have not yet really seen enterprise AI penetrate enterprise business process,” he said in February at the India AI Impact Summit in New Delhi. He also boasted about how the company has too much demand.

“We almost always find ourselves having to manage too much demand,” he added. “We are still an organization that is growing, and so there is this global demand factor that we would love to be able to meet, and we are working as best as we can to be able to meet.”

Yet, almost a month later, OpenAI appears ready and willing to give away new models, as well as granting big private equity firms firms seniority, “guaranteed returns,” and downside protection, meaning the PE firms get paid first.

Anthropic, for its part, has over 300,000 enterprise customers, accounting for about 80% of its revenue, Reuters reported last year. By contrast, OpenAI surpassed 1 million business customers last year, describing this as the “fastest growth in business platform history.” Despite this, high inference costs are nuking profitability and put into question the company’s “bigger is better” thinking. OpenAI’s inference costs — essentially how much it is to run an AI model on new data — quadrupled last year, causing adjusted gross margin to drop from 40% to 33%. SaaStr, a networking and business intelligence firm for software firms, estimates that Anthropic may generate around $211 in revenue per monthly user, while OpenAI’s comparable figure is about $25 per weekly user.

This is a substantial revenue gap. Considering OpenAI’s debt structure and projected cash burn, its plan to onboard more users with $4 billion in new funding is nothing short of worrisome.

OpenAI’s strategy goes back to Peter Thiel’s core thesis in “Zero to One.” The idea being that a startup’s goal is to escape competition by becoming a monopoly. But this does not appear to be working for a company that now looks increasingly over-leveraged and deeply indebted. In the first nine months of 2025 alone, OpenAI spent $8.67 billion on inference.

This is bad for OpenAI, but possibly even worse for companies like Oracle, which is OpenAI’s primary cloud provider and holds some $130 billion in total debt with $248 billion in new lease commitments. And then there is Coreweave, with $21.4 billion in debt and $29 billion in total liabilities, versus $3.9 billion in equity. And this is all while OpenAI only has $4 billion in revolving credit, largely untapped.

In other words, OpenAI has other companies in debt based on the hope that their AI enterprise efforts will one day pan out, which, by the numbers today or in the future, does not look to be the case.

One stock. Nvidia-level potential. 30M+ investors trust Moby to find it first. Get the pick. Tap here.