Snowflake's BUILD London 2026 conference revealed the scale of enterprise AI platform competition. The company launched Cortex functions, integrated notebooks, and agent evaluation tools—features designed to keep enterprise workloads locked into its cloud infrastructure rather than competitors'.
HSBC and Lloyds are among financial institutions now running AI workloads on third-party cloud platforms. This enterprise migration marks a shift from proof-of-concept AI to production deployments requiring industrial-grade infrastructure.
Microsoft Azure, Google Cloud, and AWS each offer competing AI platform suites. Azure provides OpenAI integration and custom silicon. Google Cloud emphasizes TPU access and Vertex AI. AWS counters with Bedrock and proprietary Trainium chips. The race centers on which platform can reduce enterprise AI deployment friction while maintaining vendor lock-in.
Analyst actions reflect confidence in sustained infrastructure demand. Morgan Stanley raised ASML price targets by 40%, betting on continued semiconductor equipment orders for AI chips. Wolfe Research named Nvidia its top AI infrastructure pick, citing enterprise GPU demand durability.
Snowflake's strategy illustrates the competitive dynamics. Cortex functions let enterprises deploy AI models without leaving Snowflake's environment. Notebooks enable data scientists to work natively within the platform. Agent evaluations provide testing frameworks—all features meant to prevent customers from mixing Snowflake data with rival platforms' AI tools.
The financial sector's adoption is particularly significant. Banks face regulatory requirements that make cloud AI platform choices semi-permanent. Once HSBC commits material workloads to a platform, switching costs create multi-year lock-in.
This infrastructure competition differs from previous cloud wars. AI workloads require specialized silicon, high-bandwidth networking, and tight integration between storage and compute. Platforms that bundle these capabilities effectively can command premium pricing and customer retention rates that exceed traditional cloud services.
The analyst upgrades suggest expectations for years of infrastructure buildout, not quarters. ASML's 40% target increase implies prolonged chip manufacturing capacity expansion. Nvidia's top-pick designation reflects belief that GPU demand will outlast current enterprise AI adoption rates.