When rattus rattus is seen deserting a ship, must you ask: “Is the vessel sinking?”
This from The Economist: When entrepreneurs walk into the offices of Andreessen Horowitz, a big American venture-capital firm, the odds these days are that their startups are using AI models made in China. “I’d say 80% chance [they are] using a Chinese open-source model,” says Martin Casado, a partner at AH.
Sequoia Capital and Andreessen Horowitz (a16z for short) are regarded as the two titans of tech venture capital in Silicon Valley. a16z’s corporate motto is “Software is eating the world” and this essay aims to show precisely that… and underline the fact that it is more and more Chinese software and less and less American hardware that is doing the eating. That said, a revolutionary Chinese hardware system is being put in place to complement China’s software.
And yet the prevailing narrative – unsurprisingly fostered by US Big Tech and echoed by the talking heads of Wall Street – could hardly be more different. If this narrative were your only source of information, you would surely believe that the US is winning the world’s tech race and that the US AI ecosystem reigns supreme.
Why then are stock market returns, year to date and in US Dollars, twice as good in Shanghai (+33%) as they are on the S&P (+15%)? Granted, Nvidia (+78%) is the world’s best performing big tech stock.
But then comes China’s Mag 8:
- Xiaomi (+71%);
- SenseTime (+67%);
- JD.com (+52%);
- Cambricon (+42%);
- Tencent Holdings (+41%);
- Baidu (+38%);
- Hunyuan (+34%); and
- SMIC (+34%).
Then comes Microsoft (+31%). Note that Huawei, DeepSeek, Biren and Z.ai are unlisted.
To use an appropriate phrase: “Something does not compute.” Granted, after recent stock price run-ups, there may even be a degree of financial bubble in Chinese tech stocks too as there most surely is in US ones! Valuation wise, US tech stocks are almost twice as expensive as their Chinese peers: the S&P 500 Information Technology sector’s P/E ratio is 40 compared with the Hang Seng TECH Index’s 22 and the MSCI China Tech 100 Index’s 23.
I used the following Darwin quote to conclude my first long AI piece (available upon request) but I cannot help but return to it every day.
As natural selection acts by competition, it adapts the inhabitants of each country only in relation to the degree of perfection of their associates; so that we need feel no surprise at the inhabitants of any one country, although on the ordinary view supposed to have been specially created and adapted for that country, being beaten and supplanted by the naturalised productions from another land. – Charles Darwin – On the Origin of Species: By Means of Natural Selection (1859)
There are two main differences between the Chinese and US approaches to AI.
First, Chinese companies constantly benchmark themselves against US peers whereas US players seem largely oblivious of their Chinese counterparts: January 2025’s DeepSeek “moment” was forgotten in a week.
US AI largely lives with its head in its own “AI cloud”, with all but a few US commentators aware of the “naturalised productions from another land”, with that “another land” of course being China.
Second, China is building a tech ecosystem – integrated, self-upgrading, symbiotic – that prizes coopetition (a business strategy where companies simultaneously cooperate with and compete against each other) among its national players even as it must compete with individual US players. Of course, some might see the recently convened “six-way” love-in between Nvidia, Microsoft, Oracle, CoreWeave and AMD, all with OpenAI as coopetition US-style!
As valuations above hint, storm clouds are on the horizon, suggesting a financial bubble might be forming in US tech. Alarm bells are ringing across the commentariat: academics, technology analysts, investors, even industry insiders.
Here are a few of the growing list of relevant quotes from outside observers on Wall Street valuations:
Jeffrey Pfeffer, Stanford GSB: “We’re seeing irrational valuations across AI. The belief that more compute equals more value is like thinking bigger sails make a boat unsinkable. But if the hull is rotten… no sail will save it.”
Julien Garran of MacroStrategy Partnership: “The AI bubble is 17 times larger than dotcom and four times bigger than subprime.”
Leading industry insiders are feeling queasy about the high capital expenditure Big Tech has been and still is undertaking. Such insiders fear that heavy data centre investments being made will not see appropriate returns:
Jeff Bezos, CEO Amazon: “I don’t see how these data centres pay for themselves.”
Mark Zuckerberg, CEO Meta: “But there is definitely a possibility, based on past large infrastructure buildouts… that something like that (a bubble) could happen here. If we end up misspending a couple of hundred billion dollars, that’s going to be very unfortunate, obviously.”
Elon Musk, X post: “The AI bubble is real. Not because the tech fails. Because the economics do. You can’t build sustainable intelligence on $30,000 chips.”
Greg Brockman, president, OpenAI: “We must prove ROI before capital runs dry.”
Perhaps the most disingenuous insider admission comes from Bezos, founder of Amazon. He suggests that today’s stock markets are in a “good” bubble, defining current euphoria as a “kind of an industrial bubble as opposed to financial bubbles.”
I predict the future will likely deliver – albeit perhaps not as fast as Amazon Prime! – a different package from the one Bezos is currently expecting, perhaps on a scale to the “parcel bomb” he received in 1999 when the dotcom bubble burst and Amazon’s share price fell from $113 to $6!
Investors are growing uneasy too:
Harris Kupperman, CEO of Praetorian Capital: “This is one of those rather surreal situations where everyone senior in this AI ecosystem knows that the math doesn’t work.” (Kuppy’s essay is a must-read for financial analysts.
David Einhorn, founder of Greenlight Capital: “The numbers that are being thrown around are so extreme that it’s really, really hard to understand them… There’s a reasonable chance that a tremendous amount of capital destruction is going to come through this cycle.”
Then there are comments that focus on the big picture of the US stock market and the US economy at large while not specifically mentioning AI:
David Solomon, CEO Goldman Sachs: “I wouldn’t be surprised if in the next 12 to 24 months we see a drawdown with respect to equity markets.”
Leon Cooperman, CEO of Omega Advisors, and quoting Warren Buffett: “We are in the late innings of a bull market where bubbles can form and risks rise.”
Jerome Powell, chairperson of the US Federal Reserve: “We are seeing unusually large amounts of economic activity.” (He was alluding to the data centre capex build of perhaps $400-billion in 2025.)
And here is one comment that bluntly mentions AI and the scale of its influence:
Ruchir Sharma, chairperson of Rockefeller International: “America is now one big bet on AI.”
There has been bubble talk for many months but, until MIT released its August 2025 report titled “The GenAI Divide: State of AI in Business 2025”, such warnings were brushed off by the US AI fan club. But even for these New Era optimists, the MIT study’s findings were too stark to be dismissed out of hand… though the apologists for the sky-high stock market valuations have tried, mostly by ignoring the report!
MIT made many disturbing findings, none more so than the one captured in the report’s opening line:
“Despite $30–40-billion in enterprise investment into GenAI, this report uncovers a surprising result in that 95% of organisations are getting zero return.”
The report goes on to say that just 5% of integrated AI pilots are extracting millions in value, while the vast majority remain stuck with no measurable P&L impact.
Meanwhile, enterprise-grade systems, custom or vendor-sold, are being quietly rejected; 60% of organisations evaluated such tools, but only 20% reached pilot stage and just 5% reached production.
The nub of the findings speak to my central contention in this essay. As big as the Wall Street financial bubble might indeed be, the biggest bubble in the US is not so much financial as technological. DM
*Read part two here and part three here.