Why bulls have an edge in AI bubble debate: Marty Fridson

Ultimately, the biggest difference between then and now is earnings. At the start of 2000, a number of prominent dotcom companies enjoyed sizable stock market valuations despite never having reported a penny of profit. For example, Akamai Technologies AKAM.O and Amazon AMZN.O — which both had valuations above the index's $25 billion average — reported 1999 net losses of about $59.8 million and $720.0 million, respectively. By contrast, all of today’s hyperscalers are solidly in the black, with net income for 2025 ranging from $450 million to $4.1 billion. Another huge spender in the AI race, chipmaker Nvidia NVDA.O, reported a $72.9 billion profit last year. Judging solely by those numbers, the present AI-fueled tech valuations hardly seem comparable to that period. This doesn’t mean today’s bulls will be vindicated, of course. For starters, the incredibly high rates of profit growth at tech companies in recent quarters could represent a different type of bubble. EPS growth may slow down as elevated borrowing costs and depreciation from the massive data center buildup kick in. Chipmakers could also suffer if “slow AI” turns out to be more than just talk. Aside from that, the Fed could take the wind out of the equity market’s sails more broadly if it has to raise interest rates meaningfully to tame the inflation beast. Ultimately, while some signs of froth are present, many of today’s valuations appear fairly reasonable. Unlike in the dotcom era, prices today still haven’t abandoned all rationality. Investors should also bear in mind that history provides not only examples of pundits who presciently called market tops, but also cases of eminent commentators who raised false — or at least extremely premature — alarms. For example, while accepting an award from the American Enterprise Institute on December 5, 1996, Alan Greenspan said that stock prices were inflated by “irrational exuberance.” Over the next 12 months, the S&P 500 rose by over 40% and didn’t peak until four years after the Federal Reserve chair issued his warning, having risen almost 110%. Similarly, economist Nouriel Roubini derided the March 2009 rebound from the lows of the Global Financial Crisis as a “dead cat bounce.” Instead, that month’s S&P 500 level turned out to be the nadir for the next 10 years. The index ended March 2019 up more than 225%. The reason for recounting these failed prophecies is not to disparage the individuals who made them. Rather, it’s to underscore the monumental challenge in accurately forecasting both direction and timing in markets. That’s something everyone can agree on. A debate is raging over whether today’s high valuations on Wall Street are comparable to the excesses of the 1999-2000 dotcom bubble. The bears may grab the headlines, but the bulls have more data on their side. It’s no wonder that the dotcom bubble still looms large for so many investors. When it burst in August 2000, the S&P 500 plunged by almost 30% in a year, with many high-flying tech stocks ultimately recording losses in excess of 90%. It took the index 6-1/2 years to recover its dotcom peak. Fast forward to this year, and several difficult-to-dismiss observers are arguing that today’s AI-fueled equity rally looks a lot like the dotcom mania. These include asset management firm GMO co-founder Jeremy Grantham, who famously called the top in the 1990s tech craze; Michael Burry, renowned for shorting mortgage-backed securities prior to their 2008-2009 crash; and billionaire hedge fund manager Paul Tudor Jones, who has said that the market could be in store for a “breathtaking” selloff. The list goes on. In the opposite camp, one finds such highly regarded authorities as Oaktree Capital co-founder Howard Marks, economist and author Jeremy Siegel, and famed economic research head Ed Yardeni. None of them could be described as a “permabull.” The fact that such serious people can disagree reflects the thorny issues involved in this debate. These include the unanswerable question of whether technology companies’ massive AI investments will ultimately produce satisfactory returns and the AI safety concerns that have garnered headlines this past week. One thing is clear, however. Simply comparing current pricing metrics with those observed around the start of this century is not enough to settle the matter. The S&P 500’s .SPX current forward price-to-earnings (P/E) multiple is 20.3. While that is certainly well above the 16.8 average seen since 1990, it’s also below the average during the dotcom years and info tech’s 22.5 P/E is less than half its peak of 54.8 in early 2000. Moreover, the S&P’s elevated P/E reading today is partly explained by the index’s historically high concentration in information technology. The sector’s weighting in the S&P 500 was 38% in the second quarter, more than twice its historical average of around 18%. This matters because the tech sector has had a higher multiple than the overall index in 73% of the quarters since the beginning of 1990. A higher concentration in tech will thus logically result in a higher P/E. Consider that if tech’s weighting last quarter had been equal to the historical average since 1990, the S&P 500’s P/E would have been 19.6, or 0.7 percentage point below its actual level.

Why bulls have an edge in AI bubble debate: Marty Fridson

Earnings help distinguish this cycle

Why market warnings can be premature

Bulls and bears clash over AI valuations

Today’s valuations versus the dotcom era

Related News

Shares climb, bonds hold ground and short-term yields edge up after Fed hikes, BoE stays on

Kearny Bank names Yvonne Louro 1st vice president, CRA officer governance and advisory manager

MNTN names ex-T-Mobile executive Mike Katz to board

Hamilton Lane appoints ex-UBS executive Diana Celotto as managing director in London

US stocks-Wall St set to open higher as easing oil prices boost Fed hike relief