RockstarMarkets
All news
Markets · Narrative··Updated 2d ago
Part of: AI Capex

AI infrastructure capex narrative remains intact despite extreme valuations

Data center and AI infrastructure stocks including CoreWeave, Nvidia, and related suppliers are rallying on sustained corporate capex plans from cloud giants. CoreWeave's CEO warns Nvidia must expand capacity or risk losing customers to AMD, signaling competitive pressure alongside secular demand growth.

R
Rocky AI · RockstarMarkets desk
Synthesised from 8 wires · 53 mentions in the last 24h
Sentiment
+60
Momentum
80
Mentions · 24h
53
Articles · 24h
49
Affected sectors
Related markets

Key facts

  • Cerebras raises IPO price range to $150-$160 per share from $115-$125; demand surges
  • CoreWeave CEO: Nvidia must expand capacity or risk losing customers to AMD
  • Wells Fargo revises CoreWeave stock price targets higher in 2026
  • Nvidia's circular investment model (funding CoreWeave, Iren, NBIS) criticized as artificial demand
  • Data center capex visibility extends through 2027 but faces margin compression risk if deceleration occurs

What's happening

The AI infrastructure capex story remains the dominant narrative driving equity markets, with data center operators, network equipment makers, and foundry-adjacent names continuing to attract capital flows despite already extreme valuations. CoreWeave CEO statement that Nvidia must expand AI capacity or risk losing customers to AMD has struck a nerve with investors, simultaneously signaling both the health of demand and emerging competitive risk. Cerebras is raising its IPO price range to $150-$160 per share (from $115-$125 previously) and increasing the share count marketed to 30 million, a sign that institutional demand for AI infrastructure plays remains robust even at elevated entry multiples.

The narrative centers on a 'Goldilocks' scenario: capex spending by AMZN, MSFT, GOOGL, and other cloud giants will sustain through 2027, supporting semiconductor and equipment maker margins indefinitely. However, this thesis faces several pressure points. First, some traders openly acknowledge that Nvidia's circular investment pattern (where Nvidia funds startups like CoreWeave, Iren, and others that then buy Nvidia chips) creates a closed-loop that may not be sustainable at scale. Second, the capex cycle could peak faster than consensus expects if AI model training efficiency improves, reducing the need for exponential increases in compute capacity. Third, AMD's emergence as a viable alternative accelerates if supply constraints ease.

Nvidia's dominance in GPU supply remains total, but the CoreWeave relationship has drawn criticism as potentially artificial demand inflation. Jim Cramer pushed back on the 'too late' thesis for AI winners, arguing data center growth remains in early innings despite recent stock moves. Wells Fargo revised CoreWeave price targets higher, reflecting the structural strength in data center utilization. Yet beneath the surface, some technical analysts note that semiconductor momentum has already hit 147% above 200-week moving averages at RSI extremes, suggesting near-term upside is increasingly priced in.

The key risk is visibility cliff: if 2026 capex growth disappoints or 2027 guidance shows deceleration, the multiple expansion that has driven valuations will reverse sharply. Companies like Broadcom, which supplies switches and optical components to data centers, could see earnings compression if network infrastructure spending normalises. Conversely, if AI training requirements accelerate (larger models, more inference compute), capex could overshoot consensus and lift all boats further.

What to watch next

  • 01NVDA, AMZN, MSFT capex guidance for 2026-2027: any moderation could trigger repricing
  • 02Cerebras IPO settlement and trading: price discovery will signal institutional appetite for AI infrastructure
  • 03AMD server chip share gains: if mix shifts away from Nvidia, capex per unit could decline
Mention velocity · last 24 hours
Coverage from these sources
Previously on this story

Related coverage

More about $NVDA

Topic hub
AI Capex: Who's Spending, Who's Earning, and What's at Risk

Tracking AI infrastructure capex — hyperscaler spend, data center buildouts, memory demand and the margin compression risk.