A hat tip here to Financial Times columnist Robert Armstrong, who said on the Prof G Markets Podcast on Jan. 30: “There is a new vision where [AI] is much more competitive, and profits are shared, and much of the value may be captured by consumers.” (Listen on Apple here or on Spotify here.)
The Answer Is “No”
A new technology emerges that ushers in remarkable productivity or, better yet, dopamine on demand. A group of talented people builds a thick layer of innovation on top of the tech (financed by government), and a small number of CEOs leverage storytelling to access cheap capital and overwhelm the competition with brute force. These companies engage in regulatory capture (i.e., they become johns for our elected whores), and create trillions in shareholder value. GPS, e-commerce, payments, search, social, streaming … all produced trillion-dollar-plus entities.
AI is likely as transformative a technology as the aforementioned, and it’s already setting records for consumer and enterprise adoption. So the question is: Which organization(s) will capture, and sustain, trillions in shareholder value? The answer may be … no.
Marc Andreessen says DeepSeek is AI’s “Sputnik moment.” That’s the wrong analogy. This isn’t a rival flexing technological superiority, but dispelling the myth that we (the U.S.) are the only game in town. The Soviet Union did that before Sputnik, in 1949, when it detonated its first nuclear weapon.
Bicycles & Vaccines
Bicycles, sanitation, airplanes, vaccines, and Crispr are just a few technologies that have changed the world, but their benefits were dispersed across society rather than being hoarded by a few shareholders. I increasingly believe AI will join this roster — it presents dynamics similar to most stakeholder vs. shareholder innovations:
- Government-backed or university-developed (the internet, GPS, vaccines);
- Open-source or public domain (Linux, Python, Wikipedia, USB); and
- Too foundational to be monopolized (bicycles, sanitation, airplanes).
In sum, the winners here will be stakeholders, not shareholders. Scan your emotions after reading the last sentence. Did you reflexively grab your shareholder pearls and feel this was, maybe, a bad thing? It isn’t.
America becomes more like itself every day, and our obsession with, and idolatry of, the dollar has put the public good in the back seat. Billions of stakeholders benefiting from AI, vs. one business becoming worth more than every stock market on Earth except Japan and the US, and one man being worth more than Boeing feels less sexy … less American. (Hint: NVDA & Jensen Huang.)
A Brief History of Last Week
In December, Chinese hedge fund High-Flyer released an open-source AI chatbot called DeepSeek that looks to be almost as good as OpenAI’s ChatGPT. It was reportedly designed in a matter of months by modestly paid millennial engineers and doesn’t run on the expensive Nvidia chips the U.S. prohibits from being sold to China. DeepSeek reportedly cost $5 million to train; it cost $100 million to train OpenAI’s LLM. Nvidia fell 17% on Jan. 27, losing over a half a trillion dollars in market cap. In one day, the company shed the value of the entire global auto industry (not including Tesla).
We don’t know if Jan. 27 was a speed bump in AI or the beginning of a tech market correction many have been expecting for 15 years. Some air definitely came out of the balloon, but just some — NVDA fell to its October ’24 price level (NBD). The DeepSeek revelation was shocking, but not surprising: A Chinese company knocks off an expensive U.S. product at lower cost (see above: China).
Gong
In retrospect, it’s easy to identify the action that rang the bell at the apex of a market. I believe the gong may have been the news that SoftBank is close to leading a $40b investment in OpenAI at a valuation of about $300b. This is double the valuation the firm raised at four months ago and a similar valuation to TikTok parent Bytedance. IP theft and addiction are both great businesses. However, crime pays more as OpenAI is being valued at 92x revenues, vs. Bytedance at 2.3x. Jesus, after reading the last sentence I wonder when Sam Altman is going to begin using terms such as “Community Based EBITDA.”
Masayoshi Son’s limited partners (i.e., his investors) are looking for venture-type returns (3x to 5x in 7 to 10 years), meaning Masa’s LPs (i.e., masochists) believe OpenAI will be one of the 10 most valuable firms in the world … soon.
Kill Orville Wright
Yesterday, I skirted along the edge of the atmosphere at four-fifths the speed of sound, traveling from London to New York in seven hours. The least expensive tickets were $400. Jet transport technology has changed the world. Sixty years ago, my mother crossed the Atlantic in a steamship: It took seven days and cost 4x what flying does today. Commercial aviation has created enormous value. However, the vast majority of that value has been captured by consumers and society, vs. airlines. Since 1945 the industry has experienced years of low-margin profitability only to have its gains wiped out by periods of huge losses (e.g., $128b in 2020).
Weak Flex
Starting in the 1980s personal computers put technology that had cost tens of millions 20 years earlier on nearly every person’s desk. The gains in productivity, globally, have been substantial. I was on the board of Gateway Computer in 2006 — weak flex. We were the second-largest manufacturer (by unit volume) of a technology that, at the time, had greater adoption and a bigger impact than AI has at this moment. I raised, and purchased, 18% of the firm for $90m. Eighteen months later, we sold it to Acer for $900m. Why? A: The CEO urged us to sell, as he felt there was a real risk we might run out of money. Think about this: NVDA shed 600 Gateways on the DeepSeek news.
AI could be enormously valuable, and at the same time a lousy business. As with email, the user may capture 99% of the value and the manufacturer 1%. What looked at first like a proprietary asset may turn out to be a public good. BTW, isn’t this how education and health care are supposed to work? But that’s another post.
Vaccines may be a useful analog here. I have a stock screen looking for potential fallen angels to buy. One of them is Moderna. At the height of the pandemic the company’s stock was nearly $500 a share. As I write this it’s $33. Vaccines may be the greatest innovation in modern history. However, their value to shareholders is fleeting.
Private assets can transform into public ones for a variety of reasons. Economists have different words for the process depending on the details: It might be called “decommodification,” “non-excludability through diffusion,” or “commonization.” They are all Latin for “there is no money here.”
Shoplifting
The ironies of DeepSeek are pretty rich. The biggest is that, as Jon Stewart pointed out, AI stole AI’s job.
Another one is the way Sam Altman has been bitching that DeepSeek stole some of his IP, “distilling” big OpenAI models to produce its own smaller, more efficient version. This conjures Steve Jobs whining that Bill Gates stole the idea of a graphic user interface from Apple. Gates responded that Apple had stolen the idea, too, when Xerox PARC left its garage door open. OpenAI is built on data it took from other people under the banner of “fair use.” Hannibal Lecter is irate that his neighbors aren’t vegan.
Another irony is that the U.S.’s attempt to keep American-made chips out of China may turn out to be a powerful argument in favor of free trade. Nvidia is now in a reasonable position to tell the U.S. government, “You’ve only created an incentive for adversaries to develop workarounds, destabilizing the AI industry and U.S.-China relations.” Anyways, it’s an argument.
Finally, though, what looks like it may be very bad for some businesses has the potential to be wonderful for the rest of us, not just for the public but also for the tech industry. Since the rise of Amazon and then Netflix, valuations have been driven not by innovation, but capital.
Kong
We thought King Kong was singular, and under U.S. control. And then a prehistoric reptile appeared on our shores, empowered by many years of exposure to nuclear radiation. FYI, Godzilla (#awesome) was meant to be a metaphor for nuclear weapons. Feels weird, but recently I find myself rooting for Canada, Denmark, and the lizard.
Life is so rich,
P.S. Introducing the Prof G Markets Newsletter, your weekly guide to the stories shaping the business world, designed for anyone who wants to understand the capital markets and establish economic security.
Subscribe here to receive the Prof G Markets Newsletter every Monday.
P.P.S. AI might not make Sam Altman (much) rich(er), but it can do a lot for you. Take 30% off Section’s AI Academy when you use code SCOTT30.
Source: www.medium.com
