Snowflake CEO: ‘I’m Not In The Business Of Selling AI. I’m In The Business Of Creating Value.’
Mutual value creation is the key to success, Snowflake CEO Sridhar Ramaswamy tells CRN. And that, he believes, means that the vendor, partner and customer should always come away with money-making opportunities.
Snowflake CEO Sridhar Ramaswamy’s philosophy on partnerships was forged back in his days building the ads and commerce business at Google.
The common denominator between working with agencies and publishers on the ads platform and working with solution providers is the need for mutual value creation, Ramaswamy told CRN in a recent interview. The vendor, the partner and the customer should come away with money-making opportunities.
That idea is intrinsic to how Ramaswami wants Snowflake and its channel partners to talk about AI.
“All of us should start with value” in customer conversations, he said. “’I’m not in the business of selling AI. I’m in the business of creating value.”
Ramaswamy is proud of the explosion of the Bozeman, Mont.-based data, AI and cloud vendor’s partner ecosystem—which now numbers more than 14,200 global partners, up 22 percent year over year and 24 times the 600 partners Snowflake had in 2022.
Snowflake channel partners see big opportunities to leverage data platforms to serve new fundamental needs in industries that have become data-heavy, as well as to unlock cutting-edge AI use cases.
Alejandro Laplana, co-founder and CEO of Austin, Texas-based solution provider Shokworks, which has been a Snowflake partner for about two years, said that the vendor’s tools have been key for data cleansing work with his midmarket customers across oil and gas, biotechnology, life sciences, wealth technology and other verticals.
His company is creating digital twins of customer businesses and leveraging data and AI for automating workflows and predictive maintenance.
“We do see the potential for transformation through the data layer, through the application layer, to be really spectacular,” he said.
Read on for more of what the Snowflake CEO had to say on the vendor’s fast-growing partner ecosystem.
On the services side, our partnership strategy comes from the fact that any customer that wants to get something meaningful done with Snowflake has to do a set of work with their internal systems.
Where does the data come from? How is it being processed today? How do we change that to use Snowflake? How do we set up the database, the schemas, the permissions in Snowflake so that it is truly enterprise-ready?
Previously, it used to be, how do we plumb it out to, let’s say, a dashboarding tool. And increasingly, [it’s becoming] how do we plumb this out to things like Snowflake Intelligence so that the data becomes consumable?
Our partners have always played a key role in helping realize the value of Snowflake, hence the growth.
This is going to be a really interesting year, not only because our partners can help customers get more value from their data using AI and agents, but they can dramatically change the process by which they do their work as well.
A single person’s ability to pull together systems end to end, it’s night and day. It’s like 100X more productive. It’s a wild time.
What’s the vibe like at Snowflake at this time when you begin fiscal 2026?
It’s one of anticipation and opportunity. What I mean by that is we had a pretty good year. The company is doing well.
But like with many other places in tech, we also know that we are on the cusp of a big set of changes. And so I think everyone, including me, feels [like], ‘Gosh, there’s a lot of opportunity, but also a lot of hard work in realizing that opportunity.’ I think it is that combination. It’s the opportunity plus the understanding that there’s a lot of change in how we operate, how we do things.
If you look back at how did we process information at the beginning of last year versus this year, you’re like, ‘Wow, that’s a big, big change.’ Similarly, how do you use Snowflake as a customer, as a user?
What I did at the beginning of last year versus now—I look up customers in real time on the fly as I’m talking to them trying to find out something about that particular aspect of our relationship.
Aerospace came up and somebody was telling me, ‘Hey, we would love to see if you work with aerospace companies because there might be a partnership there.’ [I was] completely unprepared for the question. No problem. [I pulled up] Snowflake Intelligence. ‘Here’s a list of everybody that we work with in the aerospace industry.’
Is it OK that this customer now knows you used Snowflake Intelligence?
I told them I was doing it. To me, I think, it’s the beginning. I think there’s lots more to come.
This is why even things like the Nvidia acqui-hire-ization of Groq [in December, when Groq signed a nonexclusive licensing agreement with Nvidia for Groq’s inference technology and, as part of the agreement, the company’s founder, president and other team members joined Nvidia] is so exciting.
There is attention that is going to be paid to inference time speed. People fundamentally also underestimate how much speed can drive usage. If something is really fast, you’re going to use it more.
These AI agents are very good, but sometimes they can take 30, 40, 50, 60 seconds. Super excited by, ‘Can you bring that down to two seconds? One second?’ Because then usage explodes. The kind of things that you can do with it explodes.
There’s a lot of both anticipation and realization that big changes are afoot.
How should Snowflake solution providers talk to their customers about AI in 2026?
All of us should start with value. I’m not in the business of selling AI. I’m in the business of creating value. And something like our sales agent, which I show to our customers all the time, comes with value that every CEO, every CIO, instantly understands.
It is literally every piece of information that you want to know about your customers at your fingertips. I’ve shown it to CEOs who are also Snowflake customers. Literally, my parlor trick is, when I’m talking to a CEO, ‘Let me see what we have about you.’ What’s the consumption? What are the use cases deployed? What are new use cases we are working on? What are recent meetings we have had? Did you have any support cases recently? All of it comes as part of that single report. I just turn around and show it to them, and they go, ‘I want that.’
It should always be accompanied by [a look at] what’s the value you’re actually creating? Are you saving people time?
We look at things like how many questions are being asked of the sales data agent? And we know that each of these would have meant somebody went to a dashboard, had to navigate to the right table, go look at something. Instead, they asked a question.
That kind of focus on ROI and value is part of every conversation.
I start every AI conversation either with, ‘What are the business problems you want solved?’ Or, ‘Give us an afternoon and we’ll give you a series of prototypes of what is possible with AI so you can begin to get a feel for what’s the value it can deliver.’
Sometimes, AI apparently can solve every problem known to mankind. And so a reasonable customer will be like, ‘But how do I know what ones I should have solved?’ So I think that focus on utility is important.
Do you still see a hodgepodge of where customers are on their AI journeys, some far along, some just starting?
To the extent that there are well-established, functional archetypes for how information is consumed.
Your sales team, they want to know, ‘What's my pipeline? What is my consumption revenue looking like?’ There’s a standard set of questions that is part of the sales vocabulary.
Similarly, when it comes to [human resources], it is questions like, ‘How is our hiring doing?’ Or even things like, … ‘What’s the performance distribution last quarter in terms of how we rated people internally?’ Or other questions about the health of the internal team.
So the concept of what should an HR agent do: relatively well-understood. Similarly, on the finance side, we have an agent that gives me information for things like what is our GAAP revenue … things like book revenues, a whole bunch of different kinds of revenues in every complicated company.
And so I think these archetypes then serve as the basis for, “Aha, OK, I know what a finance agent should do. I know what an HR agent should do. I know what a sales agent should do.”
What would you like to see from Snowflake partners in 2026 and beyond?
It’s an exciting time for partners. It’s a time of lots of change, primarily driven by things like coding agents.
The word ‘agent’ truly is a little buzzy. And ‘10X’ can similarly be buzzy. But I genuinely think that we can get things done 10X, even faster, compared to what could be done before. And this is born out of practical use of AI.
[AI users are] not just coming and telling you stuff that my manager told me, and in turn, they had no clue, and they went and talked to somebody. … I see productive people within my team come back personally and say, ‘Hey, this is what I did in this many hours.’ Truly magical.
What this offers for partners is a chance, for example, to focus on outcomes. This is a point of constant friction. … Partners, reasonable people, don’t like taking risks.
They’re like, ‘I’ll charge you by the hour because I can’t really estimate how much this is going to take.’ But if there is a massive compression of how much time it takes to get something done, you get a chance to switch to a model based on outcome.
We are doing this with our own services team. I tell them, ‘I want more and more of your projects to charge based on outcomes.’ Because we now have the capacity to [do this] thanks to this dramatic reduction.
Innovative things like that, in how they operate, will drive increased adoption. Honestly, I think they can get a lot more work done this year than ever before. And at Snowflake, we are very focused on creating: what are the right tools? What are agentic tools that make the process of setting up agents on Snowflake faster? It’s very ‘Matrix,’ but I think we are living it.
Is Snowflake helping its solution providers adopt an outcome-based business model?
Yes. All the tools that we create, the coding agents we create, the skills that we create, we work closely with partners.
We’re starting with the true-blue partners because they work exclusively with us. We want this to roll out more broadly. We have tools like SnowConvert, [an AI-powered migration and modernization tool], that we make freely available to all our partners. We see this very much as a collaboration.
We see our partners as amplifying our strengths and building the best tools that there are. Having them use those tools and helping them use those tools is very much part of how we operate.
With Senior Vice President of Channels and Alliances Amy Kodl (pictured) now in the channel chief role, what are your goals and hopes for her and the team?
Every use case that we win, every customer that we win, typically has a partner. I want that to be a robust, tied-at-the-hip conversation from the very beginning.
The more Amy and team bring partners along the entire selling journey, [the better] so that by the time we win a use case, there’s a partner that's going and implementing it—sometimes it might be an internal services partner, it doesn’t matter. I think the more we are holistic in our thinking about how do we compress time to value with partners, the better we are going to be.
That’s what I look at from the partnership team’s perspective, how do we have a really tight relationship with these partners?
We also have signed a bunch of, essentially, operating agreements with our partners. I want more of our partners to be making $1 billion off Snowflake every year.
If you think about the rough math, if you’re at like $4.5 billion and growing, call it, roughly like in the 25 percent to 30 percent range, you’re generating $1 billion more of revenue, at least, every year.
Typical ratios between services revenue and consumption revenue is like 5-to-1 to 10-to-1. So there is $5 billion of money to be made implementing new revenue on top of Snowflake. That’s the opportunity for our partners. It’s that $5 billion to $10 billion.
What makes a valuable partner in the AI era?
It comes back down to value creation. The partners that think purely in terms of, ‘I have capacity and my job is to fill hours and bill by the hours’ are going to have trouble succeeding in today’s world.
The more partners go, ‘It is actually a blessing that I can get the work done in a fifth of the time because it means I can be more definitive about the value that I create, create it faster and do a lot more than what I was able to do before,’ I think those are the partners that are going to succeed.
My general sort of folksy advice about business is you want to be part of rapidly growing things. Revenue growth solves all known problems, as Eric Schmidt at Google famously used to say. [Schmidt served as CEO of Google from 2001 to 2011 and executive chairman from 2011 to 2015. Ramaswamy worked at Google for about 16 years, leaving his role as senior vice president of ads and commerce in 2018].
This is a moment where opportunities like that absolutely exist. But I think being flexible, being outcome-driven, is what is going to drive great results for partners.
Over your career, how closely have you worked with the channel and partners, or is this more new for you?
I used to run Google Ads. And we had a lot of partners. All the agencies were partners because they worked with customers, they worked with us, they worked with publishers. And working with partners in that sense is not new. There are similar principles.
When I talk to partners, when I talk to people that say they want to be partners—for example, a lot of startups or new companies will want to be partners. I just tell them partnership is about value creation. We can be partners if we can jointly create value for our customer and for each other. You need to make money. I need to make money. The customer needs to make money.
And so having simple principles like that can be hugely, hugely clarifying. This is part of the reason, by the way, that we are partnering a whole lot better with people that, let’s face it, [we] struggled to partner with before. Our Microsoft partnership is in a much better place. Our [Google Cloud Platform] partnership, a big announcement came out. [In January, the companies agreed to enable Google Cloud’s Gemini 3 natively within Snowflake Cortex AI, bringing Google’s proprietary large language models (LLMs) to Snowflake’s secure, governed data environment].
SAP, another partnership. Workday, we have a very good partnership. All of these are based on how do we create win-win situations. With some of these folks, absolutely, we’ll compete in some areas. That’s OK. It’s a very big world.
Snowflake recently acquired Observe. How is that a win for your partners?
Observability is a data problem. Massive amounts of data. Structured data. Unstructured data.
It’s been a very difficult place. And our core thesis is that a data platform that specializes in storage, in efficient access of data, can be a great foundation for an observability platform.
Because Observe has always worked with Snowflake, the integration is actually going to be incredibly easy. And it’s not only going to be a world-leading product. With things like an AI SRE [site reliability engineering] agent built into how it operates, it’s also going to be an incredibly efficient product.
There is going to be a lot of opportunity for partners, and especially in the world of agents. These new software paradigms are throwing out tons more data. I think there’s a lot of work to be done on AI observability as well.
Will Snowflake keep acquiring companies, or will we see more organic innovation?
Both of them. We want to keep growing.
But at the kind of numbers that we are in, we have to be selective about who we acquire.
But definitely lots of interest in both the tuck-ins and the meaningfully large acquisitions like Observe.
How about Snowflake’s future? The company is still best suited as a public one? Would you entertain an acquisition offer?
Our goal is to create an iconic company. We think, especially in today’s world, there’s this massive opportunity not only to be where we are, but to grow significantly faster.
That’s an aspiration, clearly. And so we—this goes for me, it goes for the management team—we are focused very much on being that iconic company.
Anything I’m not asking you about? Do you want to see Snowflake add more partners to its ecosystem?
The main point that I want to hit is it is a time of big change for Snowflake and partners. AI brings us closer to helping our customers realize value. And both partners and Snowflake need to seize this opportunity together.
No [partner count goals for the ecosystem]. There are common-sense principles you need to have when it comes to partnerships. But beyond that, it’s a little bit of you let it play out.
What are your predictions for AI in 2026? How about a safe one, a medium one and a spicy one?
In terms of a safe AI prediction. I think agentic systems as we know them will become more and more commonplace for different kinds of applications, like to gather data—for example, a sales agent, which we are actively using. Those will become more and more commonplace.
But I think you’ll also see one-stop-shop kind of applications where you can both read and write data back to different kinds of source systems. Honestly, I think that’s a pretty safe prediction.
There’ll be issues with things like making sure that security is well-solved. There are things like that to get through, but that’s one thing that I feel pretty confident about.
A medium prediction would be—less a prediction than a conjecture—that open models will catch up to closed models. Not clear [that this will happen] because the research labs always have had this gap between them and the best open models. But this year might be the year when open models get to be really good and broadly used.
Spicy prediction? Not sure. AGI [artificial general intelligence, AI that is as smart or smarter than humans]? (Ramaswamy laughs). Then we all go on vacation.
This is shaping up to be a really cool year in terms of how much value we can create for our customers.
Not only great products like Snowflake Intelligence, but also things like Cortex Code [an AI-powered coding assistant]. That’s going to make the process of using Snowflake to create all of these great products on Snowflake just a whole lot easier.
I’m very excited about how we can both embrace these, learn from each other, but focus on the big thing of delivering more value to customers faster.