CoreWeave deal signals Nvidia must expand capacity or lose customers to AMD rivals
Nvidia's investment in CoreWeave infrastructure competitor IREN has sparked warnings from CEOs that the chipmaker must rapidly expand AI cluster capacity or risk losing customers to AMD and other rivals. The deal underscores intensifying competition for compute resources and highlights infrastructure supply constraints as a near-term bottleneck.
RKey facts
- Nvidia invested in CoreWeave (IREN), a direct infrastructure competitor
- CoreWeave CEO: IREN would not exist without Nvidia backing
- AMD MI300 supply improving; hyperscalers now using AMD as second source
- Data center Nvidia wait lists extend 12+ months; customers diversifying
- Wells Fargo notes AMD MI300 pricing pressure on Nvidia margin expansion
What's happening
The revelation that Nvidia has backed and is strategically invested in CoreWeave (IREN), a direct infrastructure competitor, has triggered CEO commentary that Nvidia faces mounting pressure to expand its own capacity buildout or risk customer defection. CoreWeave's CEO stated bluntly that if Nvidia did not help CoreWeave and similar infrastructure companies exist, they would face extinction; the inverse implication is that Nvidia's own capacity constraints are forcing customers toward third-party buildouts. This dynamic has accelerated interest in pure-play infrastructure names like IREN and others competing for Nvidia GPU allocations.
AMD is explicitly positioning itself as a Nvidia capacity alternative, particularly for hyperscalers building alternative clusters. Wells Fargo and other analysts have noted that AMD's MI300 GPU supply is improving and pricing is competitive, creating a viable second-source option for large cloud providers. This bifurcation reduces Nvidia's monopsony power and may pressure gross margins if customers gain leverage through alternative sourcing.
The infrastructure buildout cycle is accelerating faster than chip supply can support. Data center operators like Lambda, Crusoe Energy, and others are deploying capital to grab GPU allocations early, but many are hitting Nvidia wait lists that extend 12+ months. This scarcity is driving infrastructure companies to diversify into AMD and other accelerators, reducing Nvidia's bargaining power. Analysts covering the space have raised price targets on infrastructure names while moderating expectations for Nvidia's data center margin expansion.
Bears counter that Nvidia's lead in software ecosystems (CUDA, libraries, pretrained models) remains formidable and that AMD's ecosystem maturity lags. They also note that Nvidia's own data center capex is ramping substantially, suggesting confidence in capacity to meet demand. However, the structural dynamic of Nvidia being forced to fund and support competing infrastructure companies is a subtle bearish signal that capacity scarcity is eroding its negotiating leverage.
What to watch next
- BloombergNvidia Partner Hon Hai Profit Jumps After AI Fuels Server Sales
Nvidia Corp.’s major server assembly partner Hon Hai Precision Industry Co. reported a stronger-than-expected increase in quarterly profit, highlighting sustained spending on hardware essential for AI.
59m ago - CNBC Top NewsU.S. clears H200 chip sales to 10 China firms as Nvidia CEO looks for breakthrough
Before U.S. export curbs tightened, Nvidia commanded about 95% of China's advanced chip market.
1h ago - Yahoo FinanceStock Market Today: Nasdaq 100 Rises Despite Hot PPI, Nvidia Hits Record High13h ago
- Yahoo FinanceWhy Nvidia Bulls Are Suddenly Watching Nebius Ahead Of NVDA Earnings14h ago
- Yahoo FinanceNVIDIA Corporation (NVDA): One of the Best AI Stocks Poised for Robust Growth on Strategic Partnerships14h ago
- Yahoo FinanceAlphabet Inc. (GOOGL) Poised to Usurp Nvidia as Valuable Company on AI Boom14h ago
- Yahoo FinanceAlibaba Group Holding Limited (BABA) Refutes NVIDIA AI Chips Smuggling Claims14h ago
- MarketWatchWhat CEOs from Tesla, Nvidia and over a dozen other companies hope to gain by joining Trump in China
A key part of President Donald Trump’s visit to China this week is his effort to bring along with him at least 17 executives from prominent U.S. companies. Here’s what it means for them.
15h ago
Related coverage
- Tech CEOs Cite Severe Memory Constraints in Earnings; $MU Trading at 7x P/ETech & AI··0 mentions
- Memory Constraint Crisis: MSFT, META, GOOGL, AMZN, AAPL All Cite Supply LimitsTech & AI··0 mentions
- Alphabet Cuts AI Memory Use by 6x With TurboQuant; Gemini Efficiency GainsTech & AI··0 mentions
- NVIDIA Call Buying Hits $249M as Jensen Huang Joins Trump to China; NVDA at Record $5.5TTech & AI··0 mentions
More about $NVDA
- Alphabet Cuts AI Memory Use by 6x With TurboQuant; Gemini Efficiency Gains·Tech & AI
- Trump-Xi Beijing Summit Includes NVDA, Tesla, Apple CEOs; Trade Normalization Hopes·Tech & AI
- Magnificent 7 Call Buying Surges; NVDA, TSLA, AAPL Drive $249M Premium·Equities US
- AI Memory Shortage Sustains Capex Cycle; Chip Stocks Trade at Discount·Tech & AI
- Trump Brings Jensen Huang, Elon Musk, Tim Cook to Beijing; Tech Lobby Seeks China Deal·Tech & AI
Live coverage of the AI semiconductor cycle — NVDA, AVGO, AMD, ASML, memory demand, capex run rates and overbought signals.