competition
NVDA faces increasing put positioning ($8.45B in SMH puts per Karpathy's 13F), suggesting institutional concern about chip margin compression despite continued AI capex spending. (medium conf)
supply_chain
CoWoS/advanced packaging at TSMC remains the binding constraint, preventing NVDA's backlog from translating to proportional revenue growth per CoreWeave Q1 2026 results. (high conf)
competition
Google's Ironwoods custom silicon with model-hardware co-design represents a structural threat to pure-play GPU vendors, with Gemini 3.5 Flash achieving 4x inference speed improvements. (medium conf)
macro
NVDA faces elevated post-earnings volatility and 'long semis' crowded trade status, with Goldman desk citing narrative shift from AI infrastructure to rates. (medium conf)
competition
Diffusion models gaining share over autoregressive architectures may reduce compute-per-task intensity, potentially impacting GPU demand at the chip level. (low conf)
Nvidia is sold out through 2026 and 2027 due to extraordinarily tight compute supply. Agentic AI applications like Clawdbot are driving token demand at Anthropic. The chipmaker has delivered a 66% return over the past year and trades near its 52-week high, with a market cap of $5.4 trillion.
Investors have started returning to Nvidia after extending the semiconductor trade to CPU, memory, optical and semicap stocks over recent months. Nvidia's position as the critical provider of compute equipment is not changing soon. The stock is up 66% over the past year and trades near its 52-week high of $236.54 at $225.32.
Aletheia increased fiscal 2028 revenue and earnings estimates by 32% and 25% respectively to reflect $1 trillion revenue guidance for the company's Blackwell and Rubin chip generations and new products. The semiconductor giant has a market capitalization of $5.46 trillion and trades at $225.32 near its 52-week high of $236.54.
Thesis Impact
NVDA faces mounting competitive and macro headwinds as custom silicon from integrated players and efficiency gains reduce pure GPU dependency.