Can prior market busts help advisors prepare clients for future declines?
Nearly four years after the public launch of ChatGPT, the artificial-intelligence boom remains one of the most powerful forces in financial markets. Capital continues to pour into data centers, chips, cloud infrastructure, software and the companies building increasingly capable AI models.
At the same time, the bears have not gone away.
Their warnings are familiar. There may not be enough electricity to power the data centers being planned. Demand for AI services may not justify the enormous infrastructure investment. Today’s dominant model developers could be displaced by cheaper or more capable competitors. Higher interest rates could make the industry’s capital requirements increasingly difficult to finance. And, at the extreme end of the debate, increasingly autonomous AI systems introduce risks that are difficult even to quantify.
None of these concerns necessarily means the AI boom is approaching its end.
For advisors, however, the more important question is different: What does history tell us about the way a successful investment boom becomes a sharp market decline?
The answer is less about identifying the exact day a bubble bursts than understanding what happens when the financial mechanism supporting a boom begins to weaken.
That distinction matters.
Booms Don’t Need to Be Fake to Become Vulnerable
The dot-com bubble is often invoked as an argument against today’s AI enthusiasm. But the historical comparison is more nuanced than simply saying, “Technology stocks crashed before, so AI stocks will crash again.”
The internet really did transform the economy. The problem was not that the technology was imaginary. The problem was that investors dramatically accelerated the timetable at which its economic benefits would justify the capital being committed.
The same distinction is important today.
AI can produce genuine productivity gains, create entirely new businesses and reshape existing industries while individual securities become severely overvalued. A transformational technology and a speculative investment cycle can coexist.
That is why advisors should focus less on whether AI is “real” and more on how much future success today’s valuations already require.
History suggests that major booms become particularly vulnerable when they require continuous access to new capital.
A company can lose money for years if investors remain willing to finance its losses. An industry can build enormous capacity if lenders and equity investors believe future cash flows will eventually justify today’s spending.
But the equation changes abruptly when investors stop supplying capital.
That is where the concept of a funding gap becomes useful.
Watch the Language of the Market
During the dot-com era, “burn rate” and “funding gap” became part of the financial vocabulary.
Burn rate described how quickly a money-losing company was consuming cash. The funding gap described the capital it still needed before it could become self-sustaining.
The critical issue was not simply that a company was losing money.
It was that the company needed new investors to continue losing money.
That distinction eventually became decisive.
Pets.com became one of the era’s most memorable examples. In early 2000, the company received a clean audit opinion despite its extraordinary cash requirements. Investors remained willing to provide capital, even though the company’s economics were deeply problematic. The company reportedly sold products for roughly $5 that cost $10 to acquire and another $50 to market.
As long as capital remained available, the model could continue.
Then the market changed.
The Nasdaq fell, investor appetite disappeared, new financing became unavailable and Pets.com shut down in November 2000.
The lesson for advisors isn’t that every heavily funded AI company is another Pets.com.
It is that the availability of capital can conceal weaknesses in an investment thesis for much longer than investors expect—and the disappearance of capital can expose those weaknesses much faster than expected.
AI Has a Different Economic Foundation
There are important differences between today’s AI boom and the dot-com bubble.
Many of today’s leading technology companies are highly profitable, possess substantial cash flows and operate businesses with enormous existing customer bases. AI investment is also being made by companies that already generate significant revenue from cloud computing, software, advertising, semiconductors and other businesses.
That creates a fundamentally different starting point from the speculative internet companies of 1999 and 2000.
But differences don’t eliminate the relevance of history.
They change the question advisors should ask.
Rather than asking, “Is AI a bubble?” advisors can ask:
Which parts of the AI investment ecosystem depend on continuously rising expectations, continuously expanding capital spending or available financing?
That is a much more useful portfolio question.
An established technology company funding AI infrastructure from operating cash flow presents a different financial risk from a highly leveraged developer building facilities based on projected future demand.
A semiconductor manufacturer with diversified customers presents a different risk from a company whose economics depend on one or two enormous AI infrastructure projects.
A profitable AI software company with recurring revenue presents a different risk from a venture-backed business whose primary objective remains raising the next round.
The label “AI” tells an advisor remarkably little about these distinctions.
Capital structure, cash flow, customer concentration, valuation and financing requirements tell considerably more.
The First Warning May Not Be a Price Decline
One of the most important lessons from previous market cycles is that the warning signal isn’t necessarily a falling stock price.
By the time prices are falling sharply, the underlying financing conditions may already have deteriorated.
Advisors should therefore watch the plumbing beneath the headline market.
Is venture funding becoming more selective?
Are investors demanding stronger evidence of revenue and profitability?
Are lenders becoming more restrictive?
Are companies extending their cash runways?
Are capital expenditures being postponed?
Are projects being scaled back?
Are management teams suddenly emphasizing “discipline” in spending?
Does “funding gap” start appearing in financial commentary again?
Does “burn rate” return to common usage?
None of these developments, individually, proves that a market decline is imminent. But collectively, they can reveal that the market’s tolerance for financing future growth is changing.
That is potentially more important than another percentage point of movement in the Nasdaq.
What Advisors Should Do Now
The appropriate response isn’t necessarily to abandon AI exposure.
For many portfolios, that could create a different risk: selling a genuinely transformative technology simply because previous technological revolutions experienced speculative excesses.
Instead, advisors should make the AI exposure more explicit.
First, identify concentration that isn’t obvious on the surface.
A client may believe they own a diversified portfolio because their money is spread among several funds. Yet those funds may share many of the same mega-cap technology holdings.
The same client may also have employer stock, private investments or business interests connected to technology.
The portfolio can therefore be much more dependent on the AI investment cycle than the account statements suggest.
Second, separate technological exposure from financing exposure.
Not every AI-related company has the same vulnerability to a tightening capital environment. Advisors should examine balance sheets, free cash flow, debt maturities and capital expenditure requirements rather than treating the sector as a single trade.
Third, stress-test expectations rather than merely prices.
A conventional portfolio stress test might ask what happens if a major technology stock falls 30%.
A more useful AI-cycle exercise could ask what happens if AI capital expenditures slow substantially, projected revenue growth takes longer to materialize, financing costs rise or investors demand materially higher returns for funding infrastructure.
The resulting portfolio impact may look very different.
Fourth, establish rebalancing rules before volatility arrives.
Clients are much more likely to make emotional decisions when a decline is already underway. A predetermined framework—based on valuation bands, concentration limits, asset-allocation targets or cash-flow requirements—moves the decision from an emotional response to a portfolio-management process.
The Conversation With Clients
The most productive client conversation isn’t, “AI is another bubble.”
It is something more nuanced:
“The technology may be transformative, but transformative technologies can still produce periods of extreme investment losses.”
That distinction gives advisors room to acknowledge both sides of the market.
AI does not have to fail for AI stocks to fall dramatically.
Expected growth could simply arrive more slowly than investors anticipate. Capital spending could peak. Profit margins could disappoint. Competition could intensify. Interest rates could remain higher for longer. Or investors could decide that the future cash flows implied by current valuations deserve a larger discount.
Markets do not require an industry to become worthless before repricing it.
They only require investors to become less willing to pay today’s price for tomorrow’s promise.
History’s Most Useful Lesson
The biggest mistake in thinking about market bubbles is assuming that the only question is whether the underlying technology or economic theme is real.
It can be real and still become overcapitalized.
It can be transformative and still produce terrible investments at the wrong price.
And a company can appear financially viable for years when investors remain willing to bridge its funding gap—only to discover how fragile the model becomes when that funding disappears.
That is why advisors should watch capital flows as closely as market prices.
The return of phrases such as “burn rate” and “funding gap” would not, by itself, announce the end of the AI boom. But their reappearance could signal that investors are beginning to ask a different question: How long can this growth continue without another dollar of outside capital?
That is the question that eventually mattered in the dot-com era.
It may matter again—not because history repeats precisely, but because the financial mechanics of speculative booms have a stubborn tendency to rhyme.
For advisors, the objective isn’t to predict when the music stops.
It is to make sure clients’ portfolios can keep dancing—or simply step away from the floor—when the music eventually changes.
