Zerodha co-founder Nikhil Kamath and Coinbase CEO Brian Armstrong have warned that the high valuations of blue-chip AI companies like OpenAI and Anthropic face a massive structural threat.
The alarm comes amid growing skepticism among investors, with the two leaders comparing today’s AI craze to the dot-com crash and previous cryptocurrency bubbles.
Why Kamath is selling every AI company today
Speaking on the People by WTF podcast, both leaders drew direct parallels between the current AI boom, the dot-com bust of the 2000s, and record-breaking cryptocurrency market bubbles.
Their common interest is in expensive ownership models that are losing ground to cheaper alternatives.
Kamath has framed risk from a personal and investor perspective. He said that selling every private AI company today could make him money in five years. Comparing the moment to the Internet bubble.
“Like me, as a stock investor, I’m starting to feel at this point that if I took every private AI company and shorted it today, in five years, I might make money…it’s kind of like the ‘dot-com bubble,’” Kamath said. He said.
Follow us on XTo get the latest news as it happens.
The Zerodha co-founder also expects the industry to disintegrate. A market dominated by a few giant American companies It would make way for a self-reliant regional economy built through reverse engineering and rapid local development.
In this view, individual states Stop importing expensive models and build their own. India will run its own local version, with codes and power located locally, and efficient enough for everyday use even if not sophisticated.
“If the world goes in this direction, I see no reason to pay many times what these private companies are paying today,” the Zerodha co-founder said.
What is the 99% cheaper threat described by Armstrong?
Armstrong specifically agreed with this market assessment. He pointed to a stark cost gap between leading elite labs and the open source models that follow directly behind them.
Top-tier labs are spending billions to build the next breakthrough. Open source alternatives, about six months behind, Access the market at a fraction of that price.
The CEO of Coinbase put a number on it. The open source models run about six months behind And up to 99% lower inference cost, so a larger share of workloads can shift towards it.
He drew a clear line between two futures. Elite frontier models remain valuable for highly specialized tasks such as discovering new physics, but ordinary consumers and businesses are becoming extremely price sensitive.
“It makes me a little nervous when I see these valuations growing that fast too. It’s like I’ve seen things like this happen before in cryptocurrencies. They correct, then there’s real value underneath, and then it grows later,” Armstrong noted.
Once standard models run cheaply on everyday commodity hardware, the corporate defenses protecting highly valuable AI companies could disintegrate entirely. This corrosion is at the heart of the warning.
Armstrong closed on a cautious note. Fast-growing valuations make him nervous, mirroring patterns he has seen in cryptocurrencies, where prices correct before real value emerges and growth resumes later.
Subscribe to our YouTubeA channel to watch leaders and journalists provide expert insights.
this post The billionaire investor has just revealed an AI bet that could pay off big in 5 years appeared first on BeInCrypto.
