Oaktree Capital's Marks: NVIDIA's 30 times PE "isn't crazy," AI may not be a replica of the internet bubble

On Wednesday, Howard Marks, co-founder of Oak Tree Capital, wrote in a column for the Financial Times that despite concerns about a bubble triggered by the AI investment frenzy, the current situation is fundamentally different from the internet bubble. This seasoned investor, who has experienced multiple bubble cycles, believes that given the extremely unpredictable growth in AI demand, investor behavior undoubtedly has a speculative nature, but should not be simply classified as irrational exuberance.

Marks pointed out that NVIDIA, as a leading AI company, has a forward price-to-earnings ratio of about 30 times, which, although high, is "not crazy" for an outstanding company capable of generating huge profits, and is far lower than the valuation levels of some tech stocks during the 1999 internet bubble. He emphasized that unlike back then, AI products already exist, demand is exploding, and significant revenue is being generated rapidly.

The founder of Oak Tree Capital stated that in recent years, the market and economy have become increasingly dependent on AI, with AI stocks contributing to most of the gains in the S&P 500 index, and industry capital expenditures supporting U.S. GDP growth. However, hundreds of billions of dollars are being invested in the AI race, but the sources of profit and the ultimate beneficiaries remain unclear.

Marks suggests that investors should neither "go all in" and risk destruction nor "completely abstain" and miss out on significant technological advancements; moderate and prudent selective investment seems to be the best strategy.

Cautious Judgments from Bubble Veterans

Marks has personally experienced multiple bubble cycles and studied more historical cases. He stated that people might think that the losses caused by past bubble bursts would prevent the formation of the next bubble, but this has never happened and will never happen. Memory is short-lived, and prudent and natural risk aversion cannot withstand the dream of becoming wealthy through revolutionary technology—especially when "everyone knows" it will change the world.

He pointed out that the key issue is that new things should spark great enthusiasm, but when enthusiasm reaches an irrational level, a bubble will form. During his visits to clients in Asia and the Middle East last month, Marks was frequently asked whether there is a bubble in AI. Although he claims to be neither a technology expert nor a stock market expert, he has tried to make a rational assessment of the topic.

Marks agrees with the research findings of Byrne Hobart and Tobias Huber: investors inject funds into revolutionary opportunity areas, accelerating their progress. Some funds will be directed toward life-changing winning companies, but a large amount of capital will "go up in smoke." Of course, the key is not to become an investor whose wealth is destroyed in the process.

Reasons Why "This Time Is Different"

Marks stated that he has yet to meet anyone who does not believe that AI has the potential to become one of the most significant technological developments in history, reshaping daily life and the global economy. The current extreme enthusiasm is manifested in AI startups raising $1 billion in "seed rounds," sometimes without a clear product.

Although the similarities with past bubbles are inevitable, technology believers would argue that "this time is different." Marks pointed out that almost every bubble has heard these four words, claiming that the current situation is not a bubble. On the other hand, Sir John Templeton, who first alerted him to these four words, quickly pointed out that 20% of the time, the situation is indeed different Supporters argue that comparisons to early bubbles are inappropriate. Reasons include: Unlike the internet/e-commerce bubble, AI products already exist, demand for them is exploding, and they are generating revenue at a rapidly growing rate. Additionally, the valuation ratios of mature players, such as price-to-earnings ratios, are reasonable—NVIDIA's forward price-to-earnings ratio is about 30 times, which, while high for a company capable of generating substantial profits, is not outrageous and is far lower than the trading levels of some telecom, media, and tech stocks during the 1999 internet boom.

Debt Competition: New Risk Points

Skeptics can easily list similarities between today's events and the internet bubble. Most importantly, the ultimate sources of profit from the AI boom and who will capture those profits remain unclear. Hundreds of billions of dollars are being invested in the race for AI leadership, but we do not know who will win or what the outcome will be.

Max points out that so far, most AI investments have been composed of equity capital generated from operating cash flow. However, the winner-takes-all arms race in the AI field is now forcing some companies to take on debt. Debt itself is neither good nor bad; it all depends on the proportion of debt in the capital structure, the quality of the collateralized assets or cash flows, the borrower's alternative sources of liquidity, and whether the lender has sufficient margin of safety. We will see which lenders can maintain discipline in the current exciting environment.

Given the enormous potential of AI and the vast number of unknowns, Max states that almost no one can be certain whether investors are behaving irrationally. Therefore, he advises that no one should go all in without acknowledging the risk of failure. But equally, no one should be completely absent, risking missing out on a great technological leap.

Max believes that taking a moderate stance, combining selectivity and prudence, seems to be the best approach