Breakingviews: OpenAI’s $157 bln valuation requires hand waving

It’s possible OpenAI is worth $157 billion; it just requires a bit of hand waving. Technology giants, venture capital firms, and the usual suspects in the hedge fund and investment world piled into the company started by Sam Altman to give the artificial intelligence company $6.6 billion in new funding. Assuming these investors want a return commensurate with risk, say 20% annually for a decade, OpenAI’s value needs to grow to $1 trillion. That doesn’t immediately compute.

OpenAI estimates revenue will triple next year to $11.6 billion, and Altman thinks revenue could reach $100 billion by 2029 according to the New York Times. Few companies grow that fast for that long, but in the tech world, it’s possible. Meta Platforms was growing at about the same pace as OpenAI in 2009, and over the following four years revenue grew nine times bigger. If OpenAI achieves similar growth, it would slightly exceed Altman’s target.

Now imagine EBITDA margins were 50%, which is what Microsoft earns. That company is valued at 20 times EBITDA. Slap the same multiple on OpenAI, and the company’s investors would be able to hit their target return.

Still OpenAI is a different beast than other tech giants. Meta was solidly profitable at the same point in its history and could afford to finance growth from cash flow. OpenAI expects to lose around $5 billion this year, or over $1 billion more than the revenue it will pull in. A new $4 billion credit facility announced Thursday adds to its cash stockpile, but if OpenAI can’t fund its own growth, further hefty capital raises will be necessary – and probably be highly dilutive.

Its business model also needs to adapt. About 70% of revenue now comes from consumers, based on figures from the Times. A similar percentage in 2030 would imply $70 billion of consumer revenue, or $40 a month with 150 million subscribers. Seems doable, but when companies scale, pricing power often compresses. For example, Netflix has almost twice as many subscribers, but only pulled in $33 billion of revenue last year.

Finally, there’s the issue of margins – a 50% assumption seems almost undoable. Competition is fierce, with Alphabet, Microsoft, Meta and numerous startups pouring money into developing the most advanced AI models. There are also massive costs associated with AI – for example, a ChatGPT text search consumes 10 times the power of a Google search, per Goldman Sachs analysts, and that power is expensive. OpenAI’s insistence that investors not provide funding to two other start-ups shows just how difficult success will be.