Internet companies are also pouring cold water on AI, Airbnb CEO: AI has not brought fundamental changes, this takes time

More and more investors believe that the current AI stock valuation is in a bubble, and the real "killer app" has not yet emerged.

On August 2nd, hedge fund Elliott Management sent its latest letter to investors, stating that Nvidia's stock is in a "bubble" state due to excessive expectations for AI by tech companies. They believe that the large-scale purchases by major tech companies may not continue at the current pace, raising concerns about the sustainability of Nvidia's AI chip sales growth. This is in line with Wall Street's view that the massive investment in AI at present is not yielding sufficient returns.

Airbnb CEO Brian Chesky believes that the evolution of applications takes longer than people imagine, and what the market needs to do is to develop native AI applications that adapt to the model itself, which no company has achieved yet.

Elliott Management: AI Overhyped, Stock Market Valuation in Bubble

Legendary hedge fund Elliott Management warned in its latest investor letter last week that AI is overhyped and many applications are not ready.

Elliott Management stated that it is difficult to accurately determine when the "time" for AI will truly arrive. So far, generative AI has not brought the expected productivity miracles, with practical applications limited to areas such as summarizing meeting notes, generating reports, and assisting in computer programming.

Nevertheless, generative AI has sparked a real arms race among the four supercomputing giants (Meta, Google, AWS, and Microsoft). This year, the four giants plan to spend about $160 billion on data center capital expenditures, with a focus on Nvidia's graphics processing units (GPUs). This spending level is 50% higher than in 2023, nearly three times the spending during 2018-2020, and the market generally believes that this spending level will remain high in the foreseeable future.

Elliott Management is not optimistic about this ultra-high capital expenditure:

"Without finding a 'killer app' or significant investment return, will these supercomputing customers continue to purchase Nvidia's GPUs at record levels? We are skeptical about this and believe that the road ahead may be more challenging. If we set aside the hype and frenzy of AI 'changing human life', what remains is just a software development project that has not yet produced value commensurate with the hype (except for what we believe is a very bubbly stock market valuation)."

Brian Chesky: What We Need to Do is Develop Native AI Applications that Adapt to the Model Itself

Similar to Elliott Management's view, Brian Chesky also warned that the massive investment in AI has not yielded corresponding returns so far. The market should develop native AI applications that adapt to the model itself, but no company has achieved this yet

"One thing we've learned in the 22 months since the launch of ChatGPT is that the changes in applications take longer than people imagine. If I were to think about AI, I might divide it into three layers: chips, models, applications. We have seen a lot of innovation in chips and models, but which application has fundamentally changed because of AI?"

Covello, research director at Goldman Sachs, also expressed a similar view:

"After 18 months since the introduction of generative AI to the world, we have not yet found a truly transformative, let alone cost-effective, application."