Goldman Sachs' Kim Posnett has warned that artificial intelligence could generate pockets of speculative excess as investors increasingly favor companies that control their own models, chips and infrastructure.
Goldman Sachs · AI investment outlook
Investors Favor Full-Stack AI, Raising Risk of Speculative Pockets
Kim Posnett warns that capital is flowing toward companies that control their own models, chips and infrastructure, which could push valuations ahead of current revenue and deployment in localized areas.
How the full-stack premium can build
Step 1
Compute and chips
Control of scarce hardware
→
Step 2
Proprietary models
Models owned in-house
→
Step 3
Systems on top
Products built on the stack
→
Risk
Valuations outrun fundamentals
Localized overheating, not a uniform boom
Supporters' view
Controlling scarce inputs such as chips, data center capacity and proprietary models can create pricing power.
Vertical control can add resilience to the business.
Skeptics' view
The same concentration can inflate valuations.
It becomes harder to tell real economic moats apart from companies riding a theme.
Key takeaway
If the preference for infrastructure-centric AI bets hardens, capital could concentrate in a small group of integrated players while other AI startups face tighter funding.
Valuation gap
Whether valuations diverge sharply from underlying fundamentals.
Capital concentration
Whether funding narrows to a relatively small group of vertically integrated players.
Genuine capability
Whether the line between real capability and speculative positioning becomes clearer over coming quarters.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...