Blackstone President Jon Gray: Current AI Boom Valuation Differs from 2000 Internet Bubble
Jon Gray, President and COO of Blackstone, stated that despite Micron and SK Hynix's stock prices rising approximately 600% and 500% respectively, the current valuation levels are significantly different from those during the 2000 internet bubble. He cited SK Hynix as an example, noting its price-to-earnings ratio is only about 4 times, whereas Cisco's was as high as 150 times back then. Gray believes this difference reflects the market's ongoing doubts about the sustainability of profits for memory chip companies. Investors are concerned that current profit levels are above normal and that companies will face substantial capital expenditures in the future. Gray has repeatedly emphasized in related speeches and interviews that the demand for AI infrastructure is constrained by supply (data centers, power, and chips), rather than the typical oversupply seen in bubbles. He pointed out that while revenues for large model companies are growing rapidly and capital expenditures for hyperscale data centers are significantly increasing, physical and capital constraints make this expansion different from speculative overbuilding. Blackstone has a large presence in data center leasing and AI-related financing, including financing arrangements with companies like Nvidia. Gray mentioned that while the company has not directly invested in memory chip companies, it has observed that their stock prices have surged while valuations remain low, indicating a cautious market attitude towards peak cycle profits. He also warned that if optimistic sentiment persists too long and people believe that "trees can grow to the sky," risks remain. Gray likened the current environment to the early stages of a major technological transformation, rather than a mere valuation bubble. Market Mechanism: Gray publicly compared the low price-to-earnings ratio of memory chips with the high ratio of 2000, driven by the backdrop of expanding AI capital expenditures and rising bubble discussions; funds are flowing from skepticism about cyclical profits to infrastructure and model-related assets still seen as high-return, benefiting data centers and chip supply chain participants that can sustain expansion, while memory companies facing doubts about profit sustainability are under pressure from valuation discounts and capital expenditure burdens.
ABAB AI Insight
As the President of Blackstone, Jon Gray uses the example of SK Hynix's approximately 4 times price-to-earnings ratio compared to Cisco's 150 times to argue that the current AI boom has a different valuation structure than the internet bubble. The significant rise in Micron and SK Hynix's stock prices still maintaining low multiples reflects the market's ongoing doubts about peak profits and future capital expenditures. In terms of capital pathways, Blackstone is investing in AI infrastructure through data center leasing, GPU/TPU financing, and partnerships with Nvidia and Google, directing capital towards supply-constrained areas rather than mere valuation expansion. Gray emphasizes that demand exists while supply (power, data center permits, and chips) is constrained, contrasting with the speculative overbuilding seen in classic bubbles. This is analogous to historical cases where early-stage technological transformations faced infrastructure bottlenecks that constrained expansion speed, or where peak profits in memory cycles were priced at a discount by the market; currently, we are in a phase where AI is transitioning from explosive model revenues to matching large-scale physical and capital infrastructure. Essentially, this represents a valuation reconstruction under a transfer of pricing power: when profits rise significantly but the market still prices at peak cycle discounts, low price-to-earnings ratios reflect skepticism about sustainability, with the mechanism being that supply constraints prevent demand from being quickly met, thereby reducing the risk of valuation collapse caused by oversupply in typical bubbles, allowing capital to flow more towards areas that can resolve bottlenecks.