Snowflake CEO：AI贡献半数营收增长，上调全年指引

[music] Bloomberg Audio Studios, podcasts, radio, news.

Let's speak to Snowflake CEO Shria Ramaswami. You've had three straight quarters of accelerating um product revenue growth and it seems a strange question. Good morning to you. Welcome to the show. How much of that is AI?

Hi Ed, good morning. Yeah, indeed this quarter was pretty amazing for us. $1.49 $49 billion uh in product revenue of 37% yearonear. Uh we are doing this while also expanding operating margin by 400 basis points to 15%. About half of the beat the over performance came from AI. And what's nice about this moment is that AI is actually driving a nice feedback loop because more customers are bringing data onto Snowflake using our AI products Koko and co-work to get things done which in turn is driving consumption on the platform which sets up this very nice flywheel of people wanting to do more with Snowflake and uh it's an effect that's pretty profound.

It also gives us confidence about where we are headed. That's why we raised the full year guidance by 5% to 36% for the year.

What is Coco?

Uh Koko is a coding agent for data. It comes as part of uh Snowflake. Uh I've used it to write software as well. But what it specializes in is any job that you want to do with Snowflake goes 5 to 10 times faster using Coco.

Yeah, I was trying to I I know the answer. I was reading so much because the surprise that jumps out off the page, right, is the momentum. 2,000 accounts added in that quarter. The the the overall usage is really high in a market that is competitive. There are many coding agents out there. Why why do you think that Koko's standing out, Shreida?

Because it's the easiest sell for Snowflake as a company. I literally go to our customers and say if your data team uses Coco, you you can make things go five to 10 times faster and we publish benchmarks on this. We're good at what we do compared to off-the-shelf models and even harnesses from the foundation labs. Coco is just a lot more effective on Snowflake jobs because it understands so much about Snowflake and that's really the magic.

Um, plus we don't have things like per user subscriptions and a lot of the other products do. Coco's included for free with uh Snowflake. It's a consumption model and uh it just makes this a very very compelling product for every data team on the planet.

That business model is interesting. So if you added 2,000 um accounts in one quarter, where does that show up or translate into revenue growth? uh in first of all it shows up in AI revenue because people that are using cocoa consume tokens that's the first solder effect but the bigger effect is that they are doing all kinds of things on snowflake they are bringing more data into snowflake they are optimizing and debugging their pipelines they are setting up things like uh interactive tables that can power dashboards they're creating agents that go into co-work and can get deployed to business users it's the second and third order effect of Google that's most interesting for Snowflake

remaining performance obligations RPO came in below at least Wall Street's consensus estimates and expectations how would you ask the market to interpret that that figure and and or or what kind of emphasis would you put on it

I wouldn't put a lot of emphasis on it we are seeing this trend where the bigger renewals are getting pushed out by a quarter to the latter part of uh this year um We are uh pretty confident that RPO will increase especially in Q4 by a lot. Uh revenue is the leading indicator that uh drives everything and as you know that's very very healthy.

We talked about Coco in the context of the other coding platforms that are out there. Generally speaking, a lot of the the analysts ask about data bricks, right? And data bicks is talking about, you know, it's 80% growth. It crossing 7 billion in in ARR. Where are they doing well against you? Where do you feel you're doing well against them?

It's a big market. First of all, our strength has always been in the context of a cohesive, easy to use, trusted data platform. Their heritage is more that of catering to developers, wide open code bases, sprawl of data and uh not as much attention paid to things like governance and disaster recovery.

And uh Snowflake is the place where enterprises have put their most valuable data, the one that they use to run their business, which is what sets us up super well to move upstream. Yes, technology is important, but I have so many meaningful conversations with our customers about how we can transform how their sales teams operate or how I can how we can help them optimize their supply chains. I would say that we are leveraging AI to do the kinds of things that we honestly would not have imagined possible 3 years ago.

I'm very happy with where we are headed. Snowflake is becoming a lot more than a tech provider that's largely invisible. We are able to drive real outcomes for our customers.

Sha, I like that. You know, AI is clearly driving more usage of Snowflake. So, slightly differently, what can a customer do today with their data?

Yep.

Within Snowflake that they weren't able to do even a year ago.

Yeah. If you look at uh customers like United Rentals or uh Indeed, which is among the world's uh biggest job sites, Indeed has rolled out Coco and Co-work uh to their employees in over 60 countries and 28 languages. uh ready access to data, ready access to reasoning, ready access to being able to do anything else that the data team wants to do on top of Snowflake that simply would not have been possible a year or two ago.

That represent the frontier of what is possible with AI and data today.

Snowflake CEO Shria Ramaswami, the stock up 20% on track for its best day since May, highest level since December 2021. Thank you so much.