Nvidia's Jensen Huang Projects $3T-$4T AI Spending by 2030 Amid Supply Constraints

During the Goldman Sachs Communacopia + Technology Conference, Nvidia CEO Jensen Huang reiterated his prediction that global spending on AI infrastructure will reach between $3 trillion and $4 trillion annually by the end of the decade. This forecast mirrors his previous statement from the same conference a year prior.

Huang attributed the anticipated growth to a significant increase in demand for specialized hardware, driven by the stagnation of Moore's Law, which traditionally allowed for consistent improvements in chip performance. Nvidia's fiscal Q2 2027 report revealed a revenue of $96.2 billion, marking a 106% increase year-over-year, with $89 billion generated from data center operations alone.

However, Huang noted that supply is struggling to keep pace with this soaring demand, particularly in memory chip availability, which he identified as a key bottleneck for Nvidia's product shipments. The pricing of Nvidia's GPUs reflects this demand, with Hopper-generation chips priced around $18,000 each, and the newer Blackwell architecture costing approximately $25,000. The upcoming Vera Rubin platform is expected to be priced around $40,000.

Hyperscalers, the large-scale data center operators, are projected to increase their capital expenditures significantly, from about $800 billion in 2026 to $1.3 trillion in 2027, representing a year-over-year growth of around 60%. Nvidia is also collaborating with major financial institutions to channel over $500 billion into AI data center initiatives, with the company forecasting a revenue growth of approximately 70% for fiscal 2028, although analysts predict a more conservative growth of around 45%.

This environment of supply constraints presents investment opportunities within the semiconductor sector, particularly for companies that produce high-bandwidth memory and advanced packaging solutions, as they may benefit from the increased capital attracted by these bottlenecks.