The AI ​​Hype Hangover — Capital Brief

One of the most frequently repeated cautionary tales in the world of technology investing is the story of telecommunications companies in the late 1990s, in the early days of the World Wide Web.

Back then, U.S. telecom operators, anticipating that internet penetration rates and revenues would skyrocket—and continue to do so indefinitely—significantly expanded their fiber-optic networks and infrastructure, leading to the failure of dozens of companies as the bubble burst.

Now some investors are sounding a similar alarm about artificial intelligence.

The current AI boom has been fueled largely by “pickaxes and shovels” basic work, such as building data centers and training AI models. Actual revenue, productivity improvements, and end-user applications have largely fallen by the wayside, aside from a few standout performers like OpenAIbelongs to ChatGPT.

It was this environment that witnessed the enormous growth Nvidiathat builds the lion’s share of the chips that fill those data centers. Investors are treating it almost like an AI index fund, which is sure to grow as tech companies try to figure out what those much-touted big language models are really good for. (And how much consumers and businesses will pay for that benefit.)

In the macroeconomic report from June Goldman Sachs suggested that estimated spending on AI over the next few years, largely on infrastructure such as chips, data centers and power, would be $US1 trillion ($A1.5 trillion) and would be difficult to exceed with revenue-generating products and overall productivity improvements.