AI is creating so many millionaires–such as start-ups, investing, huge salaries, VC, consulting, etc. Even if these companies run at losses, so many people are getting rich. At the same time, everyone is now wondering, “Is this the biggest bubble ever?” These bubble arguments sound reasonable and the people who make them are smart and have credentials and industry experience, yet from my own observations, it’s never that simple or obvious. There are so many examples of companies and sectors that were widely predicted to be ‘certain bubbles,’ only for those experts and pundits be wrong.
Conventional wisdom–that AI has to be a bubble that will pop soon and will make everyone poorer as a result–is almost never correct, at least going by my own experience following consumer technology over the past 15 or so years. For every correct prediction of a bubble or crisis (e.g. Theranos or WeWork), there are many other successes that defied the doomsayers, those being Facebook, Uber, Amazon, Tesla and many others. My belief has been that these VCs are smarter or savvier than assumed. They are investing in AI companies at these inflated valuations in anticipation of huge profits down the road from subscriptions and other services, and that they likely will be vindicated if history is any guide.
I predict the huge profits will come, in hindsight making Open AI valued at $500 billion not so irrational, similar to Facebook being valued at $100 billion in 2011, which at the time was also dismissed as unsustainable and a bubble. Now Meta, the parent company of Facebook and Instagram, is worth an astonishing $1.77 trillion dollars, with significant profits to match.
I call this the ‘switch effect’: A consumer technology company after many years of running at large operating losses to build its user base and undercut competitors–at the metaphorical flip of a switch–takes the gas off the spending pedal and suddenly becomes profitable by monetizing its huge user base with ads and subscriptions. Huge losses become billions in profits. Industries that were once speculative or fringe–such as home delivery apps during COVID, social media in 2008, or cloud productivity software in 2017–eventually become thoroughly mainstream.
For example, Uber, which after over a decade of losses, turned around to massive profits:
Uber is expected to see continued profit growth, with analyst forecasts predicting earnings per share to rise from $0.63 in Q2 2025 to $3.49 by the end of 2026, a significant increase. This growth is supported by strong Q2 2025 performance, which saw net income of $1.4 billion, Adjusted EBITDA of $2.1 billion, and a 17% year-over-year rise in Gross Bookings to $46.8 billion. For Q3 2025, adjusted EBITDA is forecast to reach $2.19 billion to $2.29 billion.
Uber stock is up 100% since its IPO in 2019, and worth a record-setting $204 billion market cap. In hindsight, those VCs who paid top dollar for Uber in 2014-2017 when it was widely assumed to be a bubble or doomed to fail due to being unprofitable, are looking pretty good right now. Guess who was right? Yup.
This is why predictions of bubbles tend to be useless. Either it never pops, or it takes longer than everyone expects. There is no actionable advice or insight contained within the belief of something being a bubble. It only amounts to an opinion that things seem to be frothy, but again, so what. All mature industries were at one point frothy. Growth industries are never cheap by valuation metrics. Much of what appears as a bubble is before the switch is flipped; VCs know this. Like with Uber, a similar pattern was seen with Tesla and Amazon: they were widely predicted to be bubbles and unsustainable, and then suddenly began generating huge profits after many years of losses.
There are still people waiting for the ‘social media bubble’ to burst, even as far back as from 2007-2012. I was right, such as in 2015, that it was sustainable. Fast-forward to 2025, and Facebook and related companies are worth over a trillion dollars and generate of tens of billions of dollars of profits annually, comparable to blue-chip companies. It was widely predicted by the media that Facebook’s buyout of Instagram in 2012 and IPO was the ‘high water mark’ of the social media bubble, with Investor Business Daily boldly proclaiming “Facebook Flop Popped Social Media Bubble In 2012”. In hindsight, this prediction was greatly premature. Social media, instead of being a bubble and fading away, became an integral part of almost everyone’s life (and along with it a mental health crisis, but that is a separate matter).
Of course, there are notable failures, like the blood-testing startup Theranos, which went from a $9 billion peak valuation to worthless, but these tend to be the exceptions, that attract tons of media coverage and serve as warnings. However, such failures do not inform us of the future in so far as AI is concerned.
So to answer the original question if AI a bubble, who cares. My general belief has been to just ride the wave. If you’re employee, it doesn’t matter. You’re still being paid, so just keep riding that gravy train (although RSUs complicate matters). For VCs and founders, this is obviously a much more important question due to skin in the game. But even then, if given a choice between a chance at life-changing wealth vs waiting, I think most people will regret waiting. It could all come crashing down tomorrow, or 3-5 years from now we’ll still be waiting and predicting that this will finally be ‘the bubble’.
In conclusion, these pundits who keep predicting bubble or crisis have such poor track records, they should be dismissed on that basis alone. (The ‘broken clock’ analogy is apt.) Similar wrong predictions and warnings of a bubble were voiced about housing prices since 2010. Or the stock market for the past 15 years. Even when the market does eventually correct, as it did in 2000 or 2008, it only gives back some of the gains. History shows you still would have been better off staying invested (market timing tends to be a fool’s errand). The only actionable advice it would seem is to ignore these bubble warnings altogether and let the chips fall where they may.