00:00 Speaker A
We’re introducing Cortex Code, a data-focused coding agent.
00:06 Speaker A
What this means is that Snowflake has a detailed understanding of Snowflake’s various features and what needs to be done for everything from how to ingest data to data pipelines to creating new applications to creating AI agents.
00:20 Speaker A
The power of AI to automate many of these tasks is amazing. we are watching
00:27 Speaker A
You’ll be 5-10 times more productive when completing tasks. Customers like Whoop, who were early adopters of Cortex code, really like that’s the name of the product.
00:41 Speaker A
Obviously, there are a lot of coding agents out there, but having a coding agent that understands Snowflake-specific data, understands all of its features, and is natively integrated is a game-changer. As the CEO, I can take a screenshot of the app and put it into Coco. That’s what we call it within the company. Launch your Streamlit or React app in minutes. This is a complete game changer for us.
01:10 Speaker B
Sridhar, how does this change the business of coding?
01:15 Speaker A
Yes, it allows people to get more done faster.
01:21 Speaker A
I think they can work at a higher level as data analysts rather than writing SQL. They create what is called a semantic view. You can think of this as a description of the data. This means more business users will be able to access this data using products like Snowflake Intelligence. Debugging to pinpoint where the problem is also much easier thanks to the insights that tools like Coco provide. I think it just makes everything go faster and creates more opportunities for all of us, honestly.
01:48 Speaker B
If you and I were having this conversation a year ago, we would probably have asked the same question. But focus more on how companies, technology companies, are monetizing AI.
02:00 Speaker B
This new product makes a lot of sense to me, but how do you sell a product like this?
02:07 Speaker A
Well, first of all, this is a consumable product. So when Snowflake customers use it, they just incur, uh, consumption charges. In other words, you don’t need a large up-front commitment for your product.
02:22 Speaker A
And you’ll see very clear benefits in how quickly your team can get things done. Let me explain with a specific example. Our support team adapted Coco and changed our entire support pipeline.
02:37 Speaker A
What this means in practice is that a problem that would honestly take 5 hours to debug can be debugged in just 15-20 minutes by Coco using a set of specialized skills.
02:51 Speaker A
I believe these productivity gains will occur in every aspect of the data life cycle. And we’re definitely going to help people do more – integrate with AWS, integrate with other systems like Airflow. We really want this to be a tool for working in the data world.
03:05 Speaker B
This is an amazing technology and it’s moving very fast. So what is the role of humans?
03:10 Speaker A
Well, humans make judgments, humans have a context for how things should be done. We can identify what the problem is, but ultimately it’s up to humans to decide how to solve the problem.
03:23 Speaker A
And I’m one of those people who always wants to make sure this coding agent (even Coco) is doing the right thing. I think judgment will continue to be an important factor in the future. Review your work to make sure you’re making good progress. It is also important to understand higher levels of abstraction so that you can develop reusable components called skills that can be developed by others in your team.
03:45 Speaker A
I think it elevates the role of humans and removes a lot of the drudgery that just makes it so difficult to get the job done.
03:51 Speaker B
Another piece of news you’re announcing is that you’re entering into a new partnership with Open AI.
03:58 Speaker B
What exactly is it? What is its value to Snowflake?
04:04 Speaker A
Brian, we’ve always said, there’s no AI strategy without a data strategy. And this partnership represents the next logical step with great AI companies like Open AI working closely with Snowflake.
04:16 Speaker A
Yes, $200 million is a commitment to spend on open AI models. Obviously, that’s a big problem for us. That’s a significant amount.
04:23 Speaker A
But on the other hand, what this means is that we will work closely together to develop AI products that deliver value. Customers like Canva, which is a big customer of both Open AI and Snowflake, will benefit from us working closely together.
04:34 Speaker A
I believe there are many products that create value within the company, and by working together to develop the market, both companies can unlock that value.
04:45 Speaker B
You announced, um, I think it was December, a $200 million deal with Anthropic. You know why it’s so important to use Anthropic and AI.
04:56 Speaker B
You would imagine otherwise, but no one model wins out over time.
05:02 Speaker A
Now, these are famous model manufacturers in the world. We continue to have a very healthy relationship with Anthropic, but they all have their own areas of expertise.
05:14 Speaker A
And more importantly, customers want choice in the models they use. Well, we think these are very important partnerships for us as a data platform.
05:26 Speaker A
Collaboration between data platforms and AI providers unlocks enterprise value.
05:32 Speaker B
This year has started a little strangely, Sridhar. Well, software stocks are really taking a hit, but companies like yours are still here and posting good financial results. The outlook is solid. Where do you think there is a disconnect?
05:44 Speaker A
Well, data platforms are also a little different from pure software stocks. And I think the importance of the data layer will continue to grow. As you know, we
05:51 Speaker A
We acquired an observability company called Observe. The idea is to provide that kind of functionality natively within the platform. Yes, this kind of coding agent, cortical code, etc., makes application development easier.
06:07 Speaker A
But I think big software providers also have lasting value. So I think this effect is a bit of an exaggeration, but data platforms are key to helping businesses, so we feel very good about where we are.
06:21 Speaker A
Unlocking value, we work with Open and Humanity Feed to address that equation.
06:27 Speaker B
It’s still, I mean, do you think this is justice, I mean, like these big companies like Salesforce, these, these SAP, ServiceNow, it’s hard to believe that they’re going to disappear. But when you look at some of the market reactions, what the market is saying makes little sense to me.
06:39 Speaker A
Well, there is a market reaction. As we know, market reactions change over time. We are in a time of unprecedented change, and we believe that the ability to develop new applications is in demand.
06:55 Speaker A
It’s easier than ever. As I mentioned earlier, you can create React and many different types of dashboards just by speaking, asking questions, and making requests in natural language. I think that’s what people are trying to project, but like I said, I feel good about our ability to deliver value to our customers as one data layer that bridges everything.
07:15 Speaker A
Don’t get me wrong. We have great partnerships with companies like SAP and Salesforce. We announced a major partnership with SAP. Now you can add the power of Snowflake to your SAP data and leverage the AI data cloud.
07:29 Speaker A
We will continue to work with these companies to create the next generation of agent applications that you, me, and everyone around the world will eventually use.
07:38 Speaker B
Oh, Sridhar, I was remembering when you and I spoke. I think it was January 2025. It was around the time Deep Seek was announced. Really, it took the market by surprise, but it was a really amazing moment of innovation, and 2025 was just a phenomenal time for AI.
07:53 Speaker B
How do you think AI will be defined in 2026? What will be the next big advancement?
07:58 Speaker A
It’s about providing value.
08:00 Speaker A
As I said, what I’m doing with Snowflake itself, like the support products I talked about earlier, is to help that team do a lot of processing more efficiently. Just before this, I was in a meeting with the SRE team. Can you consider this my job?
08:16 Speaker A
The team is talking about how they can use Cortex code to make things like problem diagnosis much faster. I think it’s all about value creation when deploying these kinds of agents into real enterprise use cases. I always tell people that there are only two things that matter to a company: make more money or spend less money. I think Snowflake customers will be greatly impacted by using products like Snowflake Intelligence.
08:41 Speaker A
and the cortical code that provides that value.
