“I don't really care if there's a bubble. Bubbles pop, corrections happen, and the people who need to worry most are hedge fund managers and anyone uniquely overexposed to AI. What I care about is that the infrastructure we are building right now, especially around energy, is real and compounding. A utility CEO told me they had planned for 30–40% more demand this decade and now realize they need to prepare for 1,000%.”

“Bubbles are measured by stock prices and fear, not by real use. I look at API calls, and those keep climbing because companies like Slash run a $150 million ARR business with one accountant and AI replaces ten SDRs with one.”

“The real part is the infrastructure and the narrow use cases that already pay for themselves; the bubble is the $700–800 billion pile of money chasing impossible promises and million-dollar AI salaries that will collapse a lot of companies in two or three years.”

“I spent 25 years watching emerging tech booms, and I’m convinced the AI bubble will burst: plenty of overhyped startups will hit the dust. But the real bubble isn’t AI, it’s the outdated mindset still clinging to the past. The companies that treat AI as a platform for transformation are pulling ahead; the ones calling it a fad are the ones actually standing in a bubble.”

“I lived through the dotcom bubble and burst, and this time the leverage is different. In 2000, Cisco financed fiber networks before the demand showed up; today, Nvidia-backed GPU clouds are building capacity that has already been sold under take-or-pay contracts to investment-grade buyers. That makes the demand real, at least for now.”

“I don't think AI itself is the bubble. The bubble, if there is one, is in the solutions and industries where people slap AI on top of broken processes. What gives me conviction is the physical footprint: data centers, energy demand, and construction are booming, not just slide decks. That's infrastructure, not hype.”

“The science is real, but the economics around it are a bubble. I spent years at Meta trying to do more with less compute, and the current playbook of scaling inefficient systems to sell more GPUs is the opposite of good ML. Most VCs are parroting AI to move LP dollars, not because they understand where value actually accrues.”

“AI assets are in a bubble, but the AI market isn't. I lived through the dot-com bust and the same pattern is playing out: a rush of capital chasing half-formed ideas at the crest, then the real value gets built in the next wave.”

“AI is in a huge hype bubble, but that doesn't mean it won't change the world; most of the tokens companies are burning produce nothing, while the ones that actually learn how to embed it in operations will compound every month. I felt the same thing at 15 when people told me the internet was a scam, and I feel it again now: the money is real, the froth is real, and the split between the two is about to get obvious.”