NVIDIA CEO Jensen Huang criticized the annual infrastructure planning cycles of major cloud providers, including Amazon, Microsoft, Google, and Meta. He stated that these rigid processes are ineffective in the rapidly evolving AI market, where plans made 12 to 18 months in advance are often inaccurate.
Hyperscalers are expected to increase their capital expenditures from $800 billion in 2026 to approximately $1.3 trillion in 2027, marking a growth of roughly 63%. This rapid expansion creates significant pressure on supply chains, power availability, and construction timelines.
Huang proposed the emergence of 'neoclouds'—smaller, AI-native cloud providers focused on regional needs and speed. He suggested that the market could eventually require between 50 and 1,000 such providers to complement existing hyperscale infrastructure.
These smaller operators are intended to handle specific tasks like regional inference, edge deployments, and specialized training. According to Huang, the AI-native cloud segment is currently expanding at a faster rate than the traditional hyperscaler market.
