**The AI Boom Faces a Structural Threat: Insights from Kamath and Armstrong**
A new warning has emerged from the financial world regarding the meteoric rise of premium AI companies. Nikhil Kamath, co-founder of Zerodha, and Brian Armstrong, CEO of Coinbase, have publicly sounded the alarm that the sky-high valuations of industry giants like OpenAI and Anthropic are on a collision course with a major structural correction. Their assessment draws unsettling parallels to past economic bubbles, suggesting the current AI frenzy is unsustainable and primed for a significant downturn.
### Why Kamath Would Short Every AI Company Today
During an appearance on the “People by WTF” podcast, Kamath and Armstrong dissected the current state of the AI market, finding troubling similarities to the dot-com crash of the early 2000s and the cyclical nature of crypto booms. Their shared concern centers on a fundamental market inefficiency: the disconnect between the immense cost of developing proprietary models and the rapid emergence of cheaper, capable alternatives.
Kamath articulated a stark investment thesis, stating that shorting every private AI company today could prove immensely profitable in five years. He directly compared the sentiment to the height of the “Internet bubble,” suggesting a period of reckoning is inevitable. “It feels a bit like… the ‘Internet bubble’,” he remarked, highlighting the risky nature of current valuations.
This prediction is underpinned by the expectation of market fragmentation. Kamath believes a future dominated by a few American AI titans will give way to a more regional landscape. In this new environment, nations will prioritize building their own domestic models through reverse-engineering and local development, reducing reliance on expensive, imported AI. “If the world goes in that direction, I don’t see the reason to pay the multiples that these private companies have today,” Kamath argued.
Adding weight to this scenario is the rapid advancement of competitors. A recent example highlighted is China’s Moonshot AI and its model, Kimi K3, which reportedly surpassed established models from Anthropic and OpenAI in coding benchmarks almost immediately upon release. This demonstrates how quickly the technological lead can shift, further undermining the perceived value of current market leaders.
### What is the 99% Cheaper Threat Armstrong Describes
Armstrong echoed Kamath’s bearish outlook, focusing specifically on the dramatic cost disparity between elite and open-source AI models. He pointed out that top-tier labs invest billions into creating cutting-edge models, while open-source alternatives, which are approximately six months behind in development, can achieve similar results for a fraction of the price—up to 99% less for inference.
“The future is open source. We need to embrace it and get on with it,” Armstrong stated, citing a push from figures like Chamath Palihapitiya. He enforces a clear divide in the market: elite models will retain value for highly specialized, high-stakes tasks (like discovering new physics), while the vast majority of consumer and business applications will become intensely price-sensitive.
Armstrong’s specific concern is that as open-source models become capable enough to run on everyday, commodity hardware, the defensive moat protecting high-value AI companies will crumble. This price-driven shift could render the business models of today’s most valuable startups obsolete. Reflecting on the parallels to the crypto market, Armstrong admitted, “It makes me a little nervous when I see these valuations growing this fast as well. Like I’ve seen things like this happen before… They correct, and then there’s real value under it, so then they grow later.”
### FAQ
**Q: Why are Nikhil Kamath and Brian Armstrong warning about an AI bubble?**
A: Both leaders compare the current AI market to historical bubbles like the dot-com boom and crypto mania. Their primary concern is that premium AI companies are valued far beyond their intrinsic worth, creating a massive structural threat. They point to the vast cost advantage of open-source models and the historical tendency for such frenzies to correct dramatically.
**Q: What does “shorting every AI company” mean in this context?**
A: Shorting a stock is a bet that its price will go down. By stating he would “short every private company in AI,” Kamath is expressing a strong belief that the current valuations of these companies are unsustainable and likely to plummet in the next five years, leading to significant financial losses for investors.
**Q: What is the “99% cheaper threat” referring to?**
A: This refers to the massive cost difference between developing proprietary AI models (which can cost billions) and using open-source models, which are roughly six months behind in development but cost up to 99% less to run. Armstrong argues this cost gap will cause a major shift in the market, as businesses and consumers inevitably choose the drastically cheaper, yet still capable, open-source alternatives.
**Q: How do Armstrong and Kamath see the future of the AI industry?**
A: They predict a shift away from a centralized, American-dominated market controlled by a few giants. Instead, they foresee a fragmented landscape where nations and regions develop their own domestic AI models. This move towards “self-reliant” AI economies is driven by the desire to avoid paying the massive premiums currently asked by private AI firms and by the rapid advancement of open-source technology.
**Q: Are there any examples of this disruption happening now?**
A: Yes, the article highlights the example of China’s Moonshot AI and its Kimi K3 model. This relatively new model reportedly outperformed established leaders like Anthropic’s Fable and OpenAI’s GPT-5.6 Sol in benchmark tests almost overnight, serving as a concrete example of how quickly the technological and market leadership can change, validating the warnings about fragmentation and disruption.
### Conclusion
The warnings from Nikhil Kamath and Brian Armstrong paint a sobering picture of the AI market’s current trajectory. By drawing direct parallels to past speculative bubbles, they argue that the exorbitant valuations of leading AI companies are not based on fundamentals but on unsustainable hype. The dual threats of a potential market correction, as Kamath predicts, and a cost-driven migration to open-source alternatives, as Armstrong warns, suggest a period of significant disruption lies ahead. The future, they suggest, may belong not to a few centralized giants, but to a more distributed and price-conscious ecosystem. The question is not if the bubble will burst, but when, and how dramatically it will reshape the technological landscape.
