OpenAI is a Multi-Trillion Dollar Opportunity
Don’t be “Precautionary”
OpenAI’s $6.5 billion venture round produced lots of reaction this week. Many attempted to understand its $150 billion valuation. Some tried to use numbers to do so, most interestingly Cory Weinberg from The Information, Eric Newcomer and MG Seigler on Spyglass.
The results were varied. Weinberg starts his essay with:
OpenAI’s newest investors have signed up for a bumpy and expensive ride. The company’s projections suggest it won’t turn a profit until 2029, when its revenue would hit $100 billion.
He includes the following chart.
MG speculates on how OpenAI will ever produce an outcome for its investors, musing about how public markets will be “willing to tolerate a massively money-losing company.”
What is missing from this is any analysis of the enterprise value OpenAI can expect from its revenues. Or any objective analysis of whether the projections will likely be missed, met, or exceeded.
Google has an enterprise value of $2 trillion and annual revenues of $330 billion. So the right question might be: Can OpenAI be as big or bigger than Google? And how fast?
Its revenue journey to $4 billion is far faster than Google’s. As Reid Hoffman speculates in “Superagency,” OpenAI may have a far more significant impact on humanity than Google search, enormous as that is. At $100bn in revenue, OpenAI is worth 30% of Google. But if Reid is right about the impact and the timing is faster than anybody envisages, OpenAI could eclipse Google’s revenue numbers.
An investor would construct a J Curve of costs and Outcomes over time. By the time OpenAI gets to $44 billion in revenue (2027 in these projections), I estimate it will have spent an accumulative $30 billion on costs. That would take five years (2023–2027). The following two years would add $141 billion to the cash flow.
This is a spectacularly good J Curve. And the rate of growth would far exceed Google’s.
What would have to occur for OpenAI to be this big? It’s not that much beyond its projections. And those projections could be either too low or too high. My bet would be too low.
Bill Durodie has a wonderful research paper on “How the EU strangles innovation” focused on the “Precautionary Principle.” He notes:
In 1992, the notion of a precautionary principle was incorporated into the founding Treaty on European Union (Maastricht Treaty), as Article 130r, Point 2 of Title XVI on the Environment.1 It came to take on a central role in the fledgling Union. Its reach and stature subsequently grew, primarily through various legal rulings, until it now shapes most policy areas across the EU.
OpenAI clearly has not been influenced by the “Precautionary Principle” and neither has Reid Hoffman. In his essay about “Superagency” he writes:
AI will enable our next great leap forward. In contrast to innovations like books or how-to videos on YouTube, AI isn’t just a way to manufacture and distribute knowledge, as valuable as that is. Because an AI has the capacity to be agentic itself, setting goals and taking actions on its own to achieve them, you can leverage AI in two distinct ways. In some instances, you might want to work closely with an AI — such as when you’re learning a new language or practicing mindfulness skills. In others, such as optimizing your home’s energy consumption based on real-time energy prices and weather forecasts, you might prefer to let an AI handle that by itself.
He differentiates AI from prior innovations due to its ability to act as an agent itself. And he postulates:
Superagency is what happens when a critical mass of individuals, personally empowered by AI, begin to operate at levels that compound throughout society. In other words, it’s not just that some people are becoming more informed and better-equipped thanks to AI. Everyone is, even those who rarely or never use AI directly.
None of this is baked into OpenAIs numbers. But it will appear in its actual numbers.
$6.5 billion at $150 billion valuation is cheap if you can get in (which you probably cannot).
Other exciting themes this week. What is real-time news becoming? Max Read describes how badly social media deals with news:
None of these people are particularly informative or helpful; they’re barely even entertaining. All these videos — which are generally what surfaces to the top when you search “Milton” on TikTok or Twitter — are new recombinations of the same kinds of awful reality-TV motifs that now dominate every social network: endurance challenges to unclear end; wealth porn and self-aggrandizement; and, most importantly, mental illness as a form of entertainment. It’s just that, instead of watching them because you’re bored, you watch them because they’re what comes up when you’re looking for information about an ongoing news event on social media.
That contrasts with Zhen Yang’s piece about how data science is impacting the New York Times.
But the hat tip of the week goes to the New Yorker for its story about how Silicon Valley is becoming good at politics. The article charts the career of Chris Lehane, who has helped many Valley companies achieve their policy goals. Lehane understands that big tech companies have going audiences in the tens or hundreds of millions:
“If Airbnb can engage fifteen thousand hosts in a city, that can have an impact on who wins a city-council race or the mayoralty,” Lehane told me. “In a congressional or Senate race, fifty thousand votes can make all the difference.” Of course, simply having a huge user base doesn’t guarantee that Airbnb can get everything it wants. Voters respond only to enticements that they find persuasive. But companies like Airbnb, Lehane understood, could make arguments faster, and more efficiently, than nearly any political party or other special-interest group, and this was a source of considerable power. “The platforms are really the only ones who can speak to everyone now,” Lehane said.”
And if you can deploy the power of those people in politics, then politicians tend to listen:
The only fixed truth about technology is that change is inevitable. Most of the tech industry “has run independent of politics for our entire careers,” Andreessen wrote when he announced that his political neutrality was over. Going forward, he would be working against candidates who defied tech. As Andreessen saw it, he didn’t have a choice: “As the old Soviet joke goes, ‘You may not be interested in politics, but politics is interested in you.’ ”
It is true that we live in a time when “politics is interested in you.” We humans like it when our lives improve. OpenAi, AirBnB, Google, Amazon, Apple, and others deliver on that. Social Media has more questionable upside. The “Precuationary Principle” in the EU would never give those benefits. Indeed it does the opposite, slowing or disabling them.
I know which side I am on.
Hat Tip to this week’s creators: @BillDee3723, @reidhoffman, @coryweinberg, @EricNewcomer, @johnsfoley, @cduhigg, @PeterJ_Walker, Zhen Yang, @readmaxread, @miro, @ColinRugg, @cookie, @ttunguz, @jasonlk
Contents
- Editorial:
- Essays of the Week
- How the EU Strangles Innovation — Bill Durodie
- Superagency-Reid Hoffman
- OpenAI Projections Imply Losses Tripling to $14 Billion in 2026
- The Bear Case for OpenAI at $157 Billion
- Google break-up reads like antitrust fan-fiction
- Silicon Valley, the New Lobbying Monster
- The startup world is weird right now.
- 10 things to know about the Series A funding environment in 2024
- How The New York Times incorporates editorial judgment in algorithms to curate its home page
- The future of live news online sucks
- Video of the Week
- Reversing Aging with Juan Carlos Izpisua Belmonte of Altos Labs
- Interview of the Week
- Neal Baer on CRISPR
- Startup Of the Week
- Miro’s Innovation Workspace
- Post of the Week
- Democrats Fail on Elon Musk
- News of the Week
- Geoff Hinton and John Hopfield win Nobel Prize in Physics for their work in foundational AI
- DeepMind’s Demis Hassabis and John Jumper scoop Nobel Prize in Chemistry for AlphaFold
- Amazon’s new warehouses will employ 10x as many robots
- How stuck is the startup exit market? Pretty stuck, says Pitchbook
- The Quiet Liquidity Crisis in SaaS
- OpenAI pursues public benefit structure to fend off hostile takeovers
- Alto Reveals New Fund In Partnership With SignalRank
- Fourteen AGs sue TikTok, claiming that it harms children’s mental health
- Russia bans Discord chat program, to the chagrin of its military users
- Turkey blocks instant messaging platform Discord