Would your AI efforts survive if next quarter’s budget were cut in half?
That question gets at a problem behind every business bubble: investment can make progress look real while masking risk.
A recent Wall Street Journal analysis says market bubbles tend to end once capital stops flowing. It opens with the Pets.com sock puppet from January 2000, months before the high-flying company shut down. Another company, Webvan, obtained $1 billion in funding, which at the time seemed astronomical. It also shut down.
Fast forward to today. The Federal Reserve just raised interest rates for the first time since 2023. Capital Economics expects the S&P 500 to drop to 6,500 by the end of 2027, while Morgan Stanley forecasts 8,300 by mid-2027.
I remember the internet bubble clearly from my time in Silicon Valley. Things seem eerily similar now. But we also know that the internet survived, and minted many proven companies today.
So how do you know if you’re in a bubble? And what do you do about it anyway?
Cheap Money Can Create Misleading Proof
Dutch traders in the 1630s purchased contracts for tulip bulbs that were still underground. Rising prices became the proof that tulips had value. It was one of the first modern-day bubbles.
Two hundred years later, British railway promoter George Hudson used capital to pay shareholder dividends, which made his railway lines appear more profitable. Railway shares entered a prolonged, steep decline after the Bank of England increased rates in October 1845.
During the dot-com boom, startups were judged by eyeballs and page views. It was all about getting traffic. Monetization came second, if at all.
Technology suppliers added fuel. Lucent committed roughly $7 billion to finance customers buying its equipment in 2000, compared with $2.3 billion in 1998. The internet bubble burst shortly thereafter.
Cheap capital can pay for the evidence used to defend an investment. Eventually, funding tightens. The Fed increased rates six times from June 1999 through May 2000. By October 2002, the Nasdaq had lost more than 75 percent of its value.
AI Usage Is the New Eyeballs
Nvidia’s financing effort echoes history. The company recently announced it will further backstop financing of its own chips, helping finance demand for its own product.
A similar pattern is appearing inside companies. Meta told engineers this month that token totals and AI adoption dashboards would no longer affect their evaluations. Employees had spent months burning tokens to move up internal leaderboards. Other companies are scaling back AI usage as well given they haven’t realized the promised benefits of the technology.
Usage has become the AI era’s version of eyeballs. It reveals almost nothing about whether the customer received more value.
Why Leaders Like the Wrong Numbers
Large investments create social proof. Once every competitor starts spending, the spending begins to look like validation. And any leader who questions it risks appearing behind the curve.
I have advised more than 30 Fortune 1000 companies over the years. The pattern repeats itself. Leadership teams count licenses and pilots because both numbers can increase quickly and suggest momentum. They show how much “went in” while revealing little about what came out. They are an “input metric” that says little about real results (the output metric).
Amazon provides one of the clearest lessons from the dot-com collapse. In February 2000, the company raised $672 million through convertible bonds. The bubble burst one month later.
The cash gave Amazon time. Jeff Bezos used it to stay afloat while tracking a different set of measures: customer counts, profit per unit, and defect rates. Even as Amazon’s stock plunged, Bezos later said that each measure was improving.
The Wrong Question with the Right Response
Right now, the world is debating whether or not we’re living inside a thinning bubble.
Regardless of the answer, good leadership is about making bets, but also hedging bets at the same time. You want to ensure that you’re prepared for the future, whatever that may be – a boom or bust.
You need a runway. You need fallbacks. You need flexibility. That’s good strategy. And leadership.
Consider these three approaches:
- Measure inputs and outputs. AI “seats” gets people the tools. Reduced costs or customer growth is the reason to use AI in the first place. Measure both.
- Define your critical path investments. If things tanked tomorrow, what AI investments would you keep and what would you scrap. Get rid of the scrap now.
- Protect proven value. Secure longer-term investment, vendor relationships, and customer contracts that are backed by clear measurable value – while the sun still shines.
Bubbles are built on the assumption that more investment proves the investment is worthy. Great leadership is built by taking investment and turning it into customer value. That’s the best bubble hedge.
I publish Leapfrogging the Headlines to help leaders cut through the noise, gain clarity on what’s happening, and make smarter decisions. Subscribe below to get each issue delivered free.
Join 30,000+ other leaders