What does this major AI pullback have in common with the dot-com bubble of the early 2000s?
First, technological promise and investment returns are being conflated. The internet did go on to change the world, but that didn’t make money for everyone who bought internet stocks at their highs in 2000. The same is true of AI: widespread adoption, revenue growth, and strong returns for shareholders are three separate things.
Second, in both cases, infrastructure expanded first, while end demand took longer to materialize. Back then, it was fiber-optic cable, telecom networks, and routers; today, it’s GPUs, data centers, electricity, and cloud computing capacity. Infrastructure providers can book revenue early, but ultimately, end customers across the industry still need to pay enough to cover these investments.
Third, profits upstream can amplify investment cycles downstream. When customers expand, equipment orders and profits can grow quickly; once customers slow their purchases, upstream companies may face both lower earnings expectations and falling valuations. Cisco in 2000 is a reminder that real revenue and industry leadership can still come with an excessively high purchase price.
Fourth, the market has now shifted from “rewarding investment” to “scrutinizing returns.” The recent pullback has centered on AI capital expenditure, free cash flow, and the pace of construction; the November 5 sell-off also reflected concerns about data center policy and construction timelines following the U.S. midterm elections.
#美股走弱，加密货币市场普遍承压
#美股遭遇2020年以来最严重抛售
$QQQ
First, technological promise and investment returns are being conflated. The internet did go on to change the world, but that didn’t make money for everyone who bought internet stocks at their highs in 2000. The same is true of AI: widespread adoption, revenue growth, and strong returns for shareholders are three separate things.
Second, in both cases, infrastructure expanded first, while end demand took longer to materialize. Back then, it was fiber-optic cable, telecom networks, and routers; today, it’s GPUs, data centers, electricity, and cloud computing capacity. Infrastructure providers can book revenue early, but ultimately, end customers across the industry still need to pay enough to cover these investments.
Third, profits upstream can amplify investment cycles downstream. When customers expand, equipment orders and profits can grow quickly; once customers slow their purchases, upstream companies may face both lower earnings expectations and falling valuations. Cisco in 2000 is a reminder that real revenue and industry leadership can still come with an excessively high purchase price.
Fourth, the market has now shifted from “rewarding investment” to “scrutinizing returns.” The recent pullback has centered on AI capital expenditure, free cash flow, and the pace of construction; the November 5 sell-off also reflected concerns about data center policy and construction timelines following the U.S. midterm elections.
#美股走弱，加密货币市场普遍承压
#美股遭遇2020年以来最严重抛售
$QQQ