People respond to the rise in memory and chip prices as if a natural resource is running out. That is a mistake, and I would bet a fair bit on it rectifying itself surprisingly soon.
4:20 AM · Jun 26, 2026 · 1.1K Views
Sentiment
Users support the claim that chip price spikes will self-correct via market signals rather than scarcity because markets have historically resolved such issues effectively.
Pos
100.0%
Neg
0.0%
1 comments with sentiment.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Posts from X
Most Activity
Most Activity
VIEWS176BOOKMARKS1
Anders Sandberg@anderssandberg
It is awkward for many industries, chip supply chains are slower & more inelastic than we would wish (building new chip fabs is expensive/hard, manufacturing process pipeline is long), bullwhip effects, not a lot of obvious solutions. That is typically what markets solve best.
Anders Sandberg@anderssandberg
People respond to the rise in memory and chip prices as if a natural resource is running out. That is a mistake, and I would bet a fair bit on it rectifying itself surprisingly soon.
10hViews 176Likes 4Bookmarks 1
LIKES5RETWEETS1REPLIES1
Anders Sandberg@anderssandberg
That is why I say "surprisingly soon". Saying that a market will sort something out often sounds like magical thinking: you cannot explain how it will do it, while all the problems are obvious. Yet, empirically this works really well.
Anders Sandberg@anderssandberg
It is awkward for many industries, chip supply chains are slower & more inelastic than we would wish (building new chip fabs is expensive/hard, manufacturing process pipeline is long), bullwhip effects, not a lot of obvious solutions. That is typically what markets solve best.
10hViews 104Likes 5Bookmarks 0
Anders Sandberg@anderssandberg
Most stories about market failures are about externalities or weird lack of incentives, not lack of ability to solve. There are key tough bottlenecks (please circumvent the Kroll process for titanium!) but when you can route around them things turn very flexible.
10hViews 165Likes 2
Anders Sandberg@anderssandberg
Also, there is a Jevons issue here: even if there is an AI bubble burst by the time the new manufacturing is churning out chips (as many fear), those will be useful for a lot of things. I expect the Wright/Moore/Sahal law to apply. https://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0052669
10hViews 168Likes 1
#M U G A I N B I O@smchatter1
@anderssandberg CXMT building more capacity 2026 H2 to produce 2027 H1. Korean cartel had previously explained it would not do this in near term for obvious profit reasons.
9h