NetApp bets on Novus to fix enterprise AI’s data bottleneck

NetApp CEO George Kurian has staked the company’s AI strategy on Novus, a new architecture designed to help enterprises move AI workloads from pilot to production. The pitch appears to center on data challenges, which have been a recurring theme in recent industry discussions.

Novus may represent an attempt to unify storage, compute, and governance tools, though details remain sparse. The focus on safety risks and compliance could indicate an appeal to regulated industries, where AI adoption has historically been slower but is now gaining momentum. Whether Novus succeeds in this space may depend on how well it competes with emerging startups offering specialized hardware solutions.

The timing aligns with a wave of activity in AI infrastructure. Over the past month, StartupReader’s coverage has highlighted moves like Together AI’s Saudi data center deal and Mistral’s €3 billion raise, suggesting a shift toward securing compute resources or pivoting to services. NetApp, as a legacy vendor, has existing enterprise relationships but must prove it can innovate as quickly as startups unburdened by technical debt.

What remains unclear is how Novus will stand out from existing tools. The company’s messaging mentions addressing safety risks but doesn’t specify whether this involves encryption, access controls, or other features. Adoption may hinge on whether Novus can work alongside popular enterprise platforms or if it will require customers to adopt NetApp’s ecosystem exclusively.

The bigger question is whether any single vendor can dominate the AI data layer. Hyperscalers are already embedding storage and governance into their AI platforms, while startups are tackling the problem from different angles. NetApp’s approach assumes enterprises will prefer a neutral provider, but that neutrality will only matter if the performance meets expectations. If enterprises validate it, NetApp may have identified a real need—but if not, the architecture could struggle to differentiate itself.

Sources: siliconangle.com