Summary
Highlights
In a 48-hour period, $1.3 trillion disappeared from AI-related stocks, with companies like Micron and Broadcom crashing 20%. This event has prompted investors and the financial press to use the word 'bubble,' reminiscent of the dot-com era. The concentration of market value in the top five AI companies, now at 30% of the S&P 500, mirrors levels seen just before the 1999 dot-com crash.
The dot-com crash did not signify the failure of the internet but rather the collapse of overvalued companies. Even successful companies like Amazon saw their stock drop by over 90% during the bust. The key lesson is that while technology can be transformative, irrational valuations can lead to significant wealth destruction. Many companies with poor business models failed, unlike those with genuine infrastructure and customers.
Recent events include a semiconductor sell-off of over $1.3 trillion and a 70% drop in global private equity tech deal value. Bloomberg Opinion has drawn direct comparisons to the dot-com boom, noting companies without actual AI revenue being bid up. Michael Burry, known for predicting the 2008 housing crisis, has posted charts comparing current AI capital expenditures to historical bubbles and accused major tech companies of artificially boosting earnings by extending the estimated lifespan of AI chips.
Mohamed El-Erian, chief economic advisor at Allianz, describes the current situation as a 'rational bubble,' where individual company spending decisions are logical but collectively lead to overbuilding. Scott Galloway, an NYU professor and entrepreneur, predicts AI valuations will crater 50-70% within 24 months. Ray Dalio, founder of Bridgewater Associates, states that all great technological changes produce bubbles, and the AI market is following this pattern despite the technology's real value.
While there are similarities, current AI companies like Nvidia have substantial revenue ($215 billion last fiscal year) and high growth rates (Anthropic growing significantly). This differs from many dot-com era companies that lacked real business models. However, experts like Goldman Sachs' Jim Covello warn that companies must eventually demonstrate profitability. Former CIA adviser Jim Rickards highlights the interconnectedness of AI companies, mirroring the pre-2008 banking system, suggesting that one company's stumble could have widespread repercussions.
The companies with genuine AI technology, revenue, customers, and infrastructure are not the 'pets.com' of today; they will likely survive corrections, much like Amazon did. The real risk lies with companies that merely use 'AI' in their marketing without substantial development or revenue. The market correction will filter out these pretenders, reinforcing that the question for investors is not whether AI is real, but whether the specific companies they invest in are genuinely building it.