Why OpenAI may well be completely Zuck’d

Regular readers of this Substack will recall that I have never been bullish on OpenAI. Last Fall, I warned that they might someday be known as the WeWork of AI, and in late January I wrote an essay about some of the strong headwinds they faced, despite their seeming invincibility, called “OpenAI’s Got 9.9 Problems and Twitch Ain’t One”.

Among other things, I warned that “although the underlying technology initially improved rapidly, leading to a lot of excitement, it may soon, perhaps this year or next, reach a plateau”, that “is still not clear that there will be enough adoption to justify the company’s stratospheric valuations”, “internal tensions are likely to remain”, and that “OpenAI lacks both profits and a moat”.

Six months later, almost to the day, all that has borne out. Long-term adoption (vs mere experimentation) has been modest, there is no significantly better tech now than there was in January; GPT-5 is still a no show.

The occasion of writing today’s piece however is the latter bit: profits and moat.

Let’s start with profits. I have long suspected that OpenAI was losing money, and lots of it, but never seen an analysis, until this morning. The Information just dropped a (tentative) analysis, and it’s not pretty:

That’s not great news for OpenAI, and you can see why they haven’t been, um, Open, about their financials (reporting only revenue but not costs and profits), if The Information is even close to right. (Ditto for Anthropic, which faces similar challenges).

That’s profit. How about moat, Silicon Valley-ese for a way to hold off your competition? Facebook’s initial moat, for example, was its social network, which simply couldn’t be touched; they could afford to lose money for a long time at the beginning, because of that moat.

OpenAI, as far as I can tell, doesn’t really have any moat whatsoever, beyond brand recognition. Customers can easily switch to other providers, almost anytime they want, and for month there had been more and more alternatives, including Anthropic and Google, at more and more attractive prices, forcing OpenAI itself to cut prices just a few days ago. But then was then.

Yesterday was something even more dramatic: MetaAI all but pulled the rug out from OpenAI’s business, offering a viable competitor to GPT-4 (similar on many benchmarks) for free.

LLMs have just became exactly the commodity I predicted they would become, at the lowest possible price. How can OpenAI survive in that context?

§

Obviously, their only hope is to raise more cash, and they will certainly try. Altman is a genius at raising money. But investors really ought to ask some tough questions, such as these:

• What is their moat?

• What is their unique technology?

• Do they have a killer app?

• Will the tech ever be reliable?

• What is real and what is just demo?

• What is their route in profitability when Meta is giving away similar tech for free?

All of these questions are hard, with no obvious answer; the last may be fatal. OpenAI may well be Zuck’d.

§

Ditto for Anthropic, and a number of others.

Gary Marcus’s predictions over the last couple years have been astonishingly on target.