"In economic terms, a bubble is a bet that turned out to be too big, leaving you with more supply than demand. What makes the question of the AI bubble so tricky to answer, is mismatched timelines between the breakneck pace of AI software development and the slow crawl of constructing and powering a datacenter. Because these data centers take years to build, a lot will inevitably change between now and when they come online."
"The supply chain that powers AI services is so complex and fluid that it's hard to have any clarity on how much supply we'll need a few years from now. It isn't simply a matter of how much people will be using AI in 2028, but how they'll be using it, and whether we'll have any breakthroughs in energy, semiconductor design or power transmission in the meantime."
"When a bet is this big, there are lots of ways it can go wrong - and AI bets are getting very big indeed. Last week, Reuters reported an Oracle-linked data center campus in New Mexico has drawn as much as $18 billion in credit from a consortium of 20 banks. Oracle has already contracted $300 billion in cloud services to Open AI, and the companies have joined with Softbank to build $500 billion in total AI infrastructure as part of the "Stargate" project."
An economic bubble means making a bet that yields more supply than demand. Datacenter construction and power provisioning take years, creating a mismatch with rapid AI software progress. The supply chain for AI services is complex and fluid, making multi-year capacity forecasts uncertain. Demand growth depends on how AI will be used and on potential breakthroughs in energy, semiconductors, or power transmission. Major firms have made extremely large infrastructure commitments, increasing the risk of oversupply. Survey data show widespread AI adoption but limited deep use, leaving substantial uncertainty about future capacity needs.
Read at TechCrunch
Unable to calculate read time
Collection
[
|
...
]