Vultr Operates 34 Cloud Regions as AI Data Rules Shape Infrastructure
Chief Marketing Officer Kevin Cochrane said data sovereignty and expected returns influence where companies train and run AI models. Chennai, India, is the provider’s newest region.

Vultr operates 34 cloud data center regions as companies weigh data sovereignty and expected returns when deciding where to train and run artificial intelligence models, Chief Marketing Officer Kevin Cochrane said. The provider’s newest region is in Chennai, India.
Cochrane described a shift from speculative infrastructure spending toward investments tied to established economics and returns. He also pointed to an industry maturing around an open technology stack.
He used a hypothetical global insurer adapting an open-source credit-risk model to illustrate data sovereignty requirements. If the insurer’s data cannot cross national borders, it would need a training cluster in each country, Cochrane said.
Post-training credit-risk models would also vary across the U.K., Germany, Singapore and the United States because each process uses different data, he said.
Cochrane said Vultr built its global cloud footprint to keep cloud-native applications close to users, and that AI inference follows the same principle. AI agents can make repeated requests to reasoning models, he said, so CPUs and GPUs need to operate together near the network edge.