You’ve said Snowflake lagged in AI but has now closed the gap. How are you holding on to enterprise customers as hyperscalers, cloud platforms build overlapping AI and data products?
Hyperscalers are our biggest partners. Obviously, they also compete with us. But similar to building a large language model, building a high-quality data platform is a specialised skill, and we are the best in the world when it comes to these kinds of analytic platforms.
So, we collaborate very extensively with them, and as a company, we are not going to go away. There is some competition, but we are confident about our place under the sun.
Snowflake had a stellar listing during the Covid boom but went through a long, rough period. How has it been for you since you took over?
Snowflake’s story began with a belief in the importance of data.
Specifically, the belief in analytic computation—the ability to extract insights from massive datasets at scale, train machine learning models and support deep analysis. That was the first act.
Over the past three years, we’ve focused on strengthening our core analytics layer and adding everything needed to extract real value from it. Our recent launch of Snowflake Intelligence reflects that goal—making enterprise data genuinely accessible. This has given the company strong momentum. But it’s still very early. We’re at the beginning of a long transition, and it will take time to see how it fully unfolds.
Also Read: Snowflake beats revenue estimates, signs $200 million AI deal with Anthropic
How big is AI for Snowflake’s business, and where is it headed?
Our revenue run rate from AI services is about $100 million and our revenue guidance for the year is roughly $4.4 billion, so that’s about 2%—not huge today. What’s more interesting is the number of customers looking at products like Snowflake Intelligence and deciding that Snowflake is a platform they want to bet on.
That’s why AI products matter so much. Where they go is a signal of our ability to innovate, keep up with the market, and help customers extract maximum value from their data.
You have led search advertising at Google. How do you see search ads evolving in the age of AI?
Search, as we know it, is going to change dramatically. Habits are hard to break, but the utility from search tools is increasing sharply. In a world of agents that can browse and navigate, the ability to add value will be quite amazing.
Companies like OpenAI, which now have close to a billion users, are leading that shift. The fact that they don’t have an existing business model to protect gives them more freedom to innovate than the Google search team has. I used to run the search ads team, and I’m very proud of what we built and how strong that team was—and still is.
But hundreds of billions of dollars in revenue is a very real constraint. You can’t make massive changes purely in the name of user value.
Also Read: OpenAI vs Google: ChatGPT’s daily visits fall 22%, Gemini holds steady
What about Google challengers like Perplexity?
I’m less convinced by AI summaries than others are. I don’t see a world where Google can roll out AI summaries for queries that really matter—commercial queries where people are trying to buy things. That’s where the rubber hits the road, because Google makes real money there.
For some categories, like auto insurance, Google earns several dollars per query because competition is so intense. That’s where the real challenge to the existing business model lies. That said, Google has also shown it can innovate.
Models like Gemini took longer than expected to mature.
Was that an execution issue or a structural constraint?
In some ways, it’s the opposite. You have to give them credit for their ability to move quickly. They are world-leading in several areas. Products like Nano Banana are truly remarkable. There aren’t many products like that.
The real question is whether they can continue to do that effectively over the long term, because that’s the pace at which the rest of the ecosystem is moving.
Also Read: Snowflake to acquire Observe, offer its AI observability tools to enterprises
What’s the significance of Snowflake’s acquisition of Observe which just got announced?
The acquisition reflects how central observability has become as customers build AI agents and increasingly complex, data-driven applications. Reliability is no longer just an IT concern; it directly shapes business outcomes. As enterprises operationalise AI, that data foundation becomes indispensable.
Observe operates in the app and system monitoring space, competing with players like Datadog. Why is Snowflake stepping closer to that part of the enterprise stack?
AI-driven systems generate unprecedented volumes of logs, metrics and traces, and legacy observability tools increasingly struggle with cost, scale and context. From Snowflake’s perspective, this makes observability a natural extension of the data platform. This convergence allows enterprises to analyse operational, business and AI data together, which is essential as systems become more distributed, autonomous and central to business outcomes.
Also Read: Snowflake to increase India headcount to over 700 by year-end
Do you see further consolidation in the data stack as AI becomes more ubiquitous?
I do see some consolidation. You can look at the dbt–Fivetran merger as the beginning of that process. When it came to AI, we moved quickly to ensure we had strong products, because we realised how significant it was for us—and for our survival. It’s a very dynamic, evolving space.
How are your Indian customers different from US ones, and which sectors are scaling faster for Snowflake?
We have a large number of Indian customers. Many are digital natives, but we also work with companies such as Bajaj Allianz in insurance, Godrej Capital and Deepak Fertilizers.
India is an interesting market because of its wide spectrum. You have companies with large tech teams and significant on-prem data centre investments, and you also have digital-native companies that rely heavily on cloud providers. We tend to work well with digital natives and companies inclined towards the cloud, because Snowflake runs entirely on the cloud.
What about hiring and expansion in India?
India is a crucial growth engine for Snowflake for several reasons. We’ve invested significantly in our operations here. We have a 500-plus person team in Pune and offices in Delhi and Bengaluru. Many of our partners are based in India as well—about half of our APJ (Asia Pacific and Japan) partners are here.
We’re investing in AI upskilling for our people and partners. We run a large competency centre out of Pune. That reflects the reality that for many multinational customers, such as Morgan Stanley, key technology divisions are now based in India.
How’s India market changed in terms of tech talent and also as a place for bringing in business?
Twenty years ago, India was more back-office work. That’s no longer the case. These are core infrastructure teams, as skilled and empowered as teams in places like New York. Because of that, there’s active collaboration. Some of my best account managers travel to India multiple times a year because their clients are based here.
What will AI change in how enterprises think about systems and employees?
I see a couple of trends emerging clearly in 2026. While 2025 was defined by excitement around agentic AI, 2026 will be defined by control and reliability. We’re moving towards fixed workflows, where AI is given a strict, governed path for critical operations.
That matters because the real world has different consequences for mistakes. If you’re using an agent for MRI screening for cancer, the consequences are vastly worse. Success will be measured not by novelty, but by reliability in high-stakes environments, with human-in-the-loop approvals and fail-safes.