Snowflake Build 2026 - the race to prove AI investments aren't wasted money

Enterprises are pouring money into AI, but many are struggling to show returns. With global AI spending set to hit $2.52 trillion this year - a 44% jump according to Gartner - the pressure is mounting on CIOs to demonstrate value or risk budget cuts.

Snowflake sees an opening. At its Build event for developers in London, the data platform specialist made its pitch that organizations will get better results by using AI-enabled tools from trusted providers rather than building everything from scratch - the hope being that IT leaders will rely on proven partners over experimental in-house projects.

Christian Kleinerman, Snowflake's VP of product, outlined a raft of new tools and an OpenAI partnership designed to help customers move faster. At Build, the company showcased roughly 100 new launches from the quarter, focused on simplifying data workflows from storage through to consumption.

In his keynote, Kleinerman was blunt about the urgency:

If you're thinking that AI is something you'll do in the next fiscal year, next quarter, next month, or next week, you're behind, because companies are getting value from AI today. A lot of what we do at Snowflake to innovate is to help you adopt AI more easily with less friction and to do it faster.

From idea to production, faster

The Build announcements centered on collapsing the time between AI project conception and live deployment. Snowflake's approach: give developers collaborative tools that handle the infrastructure complexity.

The centerpiece is Shared Workspaces, an evolution of the company's existing Workspaces interface. Teams can now collaborate on building production-ready AI services within a unified environment, with role-based security baked in from the start.

Snowflake is also embedding AI into its own tooling. Kleinerman pointed to Cortex Code, now generally available, as a significant example. The data-native coding agent automates enterprise development tasks, trained on Snowflake's own engineering practices and customer feedback:

What coding assistants have done for programming, software development, and software engineering, we're bringing that same paradigm to database management, database operations, and database programming.

The tool handles tasks like building data pipelines, managing governance, and creating access policies - the kind of repetitive work that slows down data teams.

Snowflake also launched Semantic View Autopilot, an AI service that learns from user behavior to maintain accurate business logic. Rather than manual data modeling, the tool taps into Snowflake's query history to suggest relationships and optimize semantic views automatically:

We leverage knowledge that is unique to Snowflake about how you are using your data. We plug into things like query history and make recommendations, saying, 'Hey, these are relationships that should be included.' The goal is to help you accelerate the deployment of AI in your enterprise.

Real deployments, tangible results

The proof, as ever, is in customer implementations. IAG Loyalty is using Snowflake's ML capabilities across its membership lifecycle, powering recommendation and personalization models for Avios members.

Neha Patel, lead machine-learning engineer at IAG Loyalty, said moving their ML platform architecture onto Snowflake eliminated operational headaches while providing a single source of truth for customer data:

Snowflake takes away lots of operational overheads and gives us reliability, which we need to run loyalty experiences at scale. The technology provides several first-class ML capabilities. This gives us a seamless and familiar development experience, as we work with Python a lot, while also keeping model outputs trusted.

At Booking.com, the challenge was different: data specialists were drowning in basic business queries, leaving little time for strategic analytics. Kiran Kodandoor, principal software engineer, said the answer came through Cortex and Snowflake Intelligence, which lets users query enterprise data in natural language.

After a successful proof-of-concept, Booking.com rolled out the agent in phases. The next step is moving from insights to automated action:

We want to take it to the next level, where we want to drive insights to action. We configure the agent to perform specific tasks, and we also have plans to integrate this technology with our internal knowledge base.

The OpenAI factor

Ahead of Build, Snowflake announced a $200 million partnership with OpenAI, bringing advanced model capabilities directly into its platform through co-innovation work. OpenAI models will be accessible within Snowflake Intelligence, creating a tighter integration between enterprise data and frontier AI.

Ashley Kramer, OpenAI's VP of enterprise, framed the partnership around trust rather than technical capability:

We're at a moment where AI is about trust, not the technology. The technology is there, and there's responsibility on OpenAI, Snowflake, and our customers. Our partnership with Snowflake allows us to bring governed data into the models in a secure way so you can be comfortable and trust the technology.

Kramer acknowledged that effective agent deployment can eliminate mundane work, but stressed that explainability and data governance can't be afterthoughts. The challenge, she said, is closing the gap between what's technically possible and what enterprises are actually extracting:

There's a huge value gap between what models can provide and the value that enterprises are extracting. And so, we want to help remove those barriers, to get you to deploy faster.

My take

Boards want ROI from AI. With generative AI in the trough and agentic AI at peak hype level, IT professionals and business leaders must start delivering. Snowflake believes that AI-enabled tools allied to its broad data platform will give organizations a head start. For end users and their tech providers, the race to value delivery is very much on.