What Snowflake Taught Me About Moats in Infrastructure Software
A closer look at Snowflake, Databricks, and what makes a durable enterprise data infrastructure business.
In this article 5
Studying Snowflake recently left me with a few strong impressions.
First, this is a product category I haven’t spent much time with before.
In my day-to-day work, I mostly encounter SaaS products for end users or business teams—marketing, operations, and productivity tools. Snowflake operates in enterprise data infrastructure, farther from everyday business workflows and closer to the underlying technology stack.
China’s cloud market has developed differently from markets such as the United States. In China, major cloud providers often offer everything from infrastructure to applications, and many enterprises buy directly from them. In other markets, independent software companies have had more room to grow.
After reading Snowflake’s financial reports and studying its competition with Databricks, five points stood out to me:
1. You don’t have to be the strongest company to build a great business. There will always be a stronger technology competitor. Commercial success depends not just on being technically number one, but on finding your place in the ecosystem, delivering value to customers, and building the right business model.
2. You can succeed even when you compete with giants—if you have your own strengths. Snowflake faces cloud giants such as AWS, Microsoft, and Google, as well as AI data platforms such as Databricks. Yet it has grown into a company with billions of dollars in revenue.
3. The technology itself can be a competitive advantage. I often focus on growth, sales, retention, and business models when studying SaaS. Snowflake reminded me that, for infrastructure products, architecture, the developer ecosystem, and engineering decisions also shape long-term competitiveness.
4. A good business does not necessarily need to go public. Databricks has stayed private even as its revenue has approached Snowflake’s and its growth has been faster.
5. Technical depth matters most for infrastructure software. Marketing alone can’t create a lasting advantage here. Sustained engineering work, a developer ecosystem, and customer trust are what build long-term value.
1. What Is Snowflake? An Enterprise Data Infrastructure Company
Snowflake is a cloud data platform.
It helps companies bring together data from different systems so they can analyze it, run machine-learning workloads, and build AI applications. It is neither a traditional database nor simply a business intelligence tool. It sits at an important point in the enterprise data stack:
Business systems (Salesforce / ERP / apps)
↓
Data storage and governance
↓
Snowflake
↓
Analytics and AI applications (Tableau / Power BI / AI agents)
In the past, enterprise data lived in dozens of separate systems and was hard to analyze together. Snowflake’s value is bringing that data into a Data Cloud, where businesses can query it consistently, analyze operations, and build AI applications.
If the earlier challenge was managing data, the AI-era challenge is putting that data to work.
Snowflake wants to be the data infrastructure for enterprise AI.
2. Why Would a Company Choose Snowflake Instead of Using AWS Directly?
AWS has a full suite of data products of its own. Why would a company still need Snowflake?
Snowflake’s pitch isn’t that its underlying infrastructure is stronger than AWS’s. It’s that Snowflake offers an easier way to manage and use enterprise data across cloud environments.
1. A more polished product experience
AWS provides the underlying building blocks and lets companies assemble them. Snowflake packages complex data infrastructure into a platform businesses can use directly.
2. Multi-cloud support
Large companies rarely use just one cloud provider. Snowflake supports multi-cloud environments, helping businesses avoid lock-in to a single platform.
3. Data sharing
Snowflake’s Data Sharing is an important differentiator. It makes it easier for companies to share data and supports a broader Data Marketplace ecosystem.
4. Separation of compute and storage
Traditional databases often couple compute and storage. Snowflake separates them, allowing different teams to run different workloads against the same data.
5. Enterprise-grade data governance
For large organizations, the key question is not whether they have data. It is who can access which data, and how to meet security and compliance requirements.
3. Snowflake and Databricks: The Cloud Data Era and the AI Era
Snowflake was a winner of the cloud-data era over the past decade. Databricks has been one of the fastest-growing data infrastructure companies during the recent AI wave. The question isn’t necessarily which will replace the other. It’s which platform companies will turn to first for enterprise data in the AI era.
Snowflake was founded in 2012 by a team with backgrounds at Oracle and in database technology. Its main offering was a cloud enterprise data warehouse—an alternative to Oracle and Teradata.
Databricks was founded in 2013 by members of the team behind Apache Spark, one of the world’s most popular large-scale data-processing frameworks. They saw that companies such as Google and Facebook did not lack databases; they needed ways to process enormous amounts of data and train machine-learning models. Apache Spark was built to process that data quickly. Databricks later commercialized the technology and introduced the Lakehouse architecture, combining data warehouse and data lake approaches to support structured analytics and AI/ML workloads.
After 2020, demand for training AI on data and building recommendation systems accelerated, making Databricks’ strengths more apparent.
| Snowflake | Databricks | |
|---|---|---|
| Origins | Data warehouse | Data lake |
| Strength | Enterprise data analytics | AI and machine learning |
| Primary users | Analytics teams and enterprise IT | Data engineers and data scientists |
In the past, businesses asked, “What happened?” Increasingly, they ask, “What should we do next?”
The data needs of the AI era have made Databricks’ strengths more visible. But that does not mean Snowflake has failed. Snowflake was a winner of the past decade’s cloud-data era; Databricks has been a fast-growing infrastructure company in the AI era.
4. Snowflake’s Financial Performance Over the Past Three Years
| Metric | FY2024 | FY2025 | FY2026 |
|---|---|---|---|
| Total revenue | $2.67B | $4.18B | $5.30B |
| Product revenue | $2.57B | $3.98B | $4.47B |
| NRR | 133% | 126% | 125% |
In FY2026, Snowflake reported:
- Non-GAAP operating income of $490 million;
- A non-GAAP operating margin of 10%;
- Operating cash flow of $1.22 billion;
- Free cash flow of $1.12 billion;
- A free cash flow margin of about 24%.
Snowflake had roughly 12,140 paying customers and $4.47 billion in product revenue in FY2026. That works out to about $368,000 in annual product spending per customer. It is only an average, but it shows that Snowflake’s business is not built around SMBs. It primarily serves mid-sized and large enterprises. Its 733 customers spending more than $1 million a year, along with adoption by more than 750 Fortune 2000 companies, make it look much more like enterprise infrastructure software.
5. Future Growth: From Data Warehouse to AI Data Platform
1. Expand from data warehouse to AI data platform
Bring enterprise data into Snowflake, connect it to AI models, and build agents and automated business workflows on top.
2. Grow the Data Cloud ecosystem
Use data exchanges, third-party data, and an application ecosystem to make Snowflake a platform that connects enterprise data.
3. Increase AI usage
Move customers beyond SQL queries toward AI-agent calls, increasing data consumption.
4. Deepen adoption within large accounts
Once Snowflake is inside a large enterprise, bring more departments onto the platform and increase customer spending.
5. Expand internationally
Continue building in Europe and Asia-Pacific, and in industries such as finance, manufacturing, and healthcare.
My biggest takeaway from Snowflake is that commercial success doesn’t require winning every technical comparison. AWS is bigger, and Databricks is growing faster in AI. Yet Snowflake has built an important position in enterprise data infrastructure.
————————————————————
Hi, I’m Renee.
I lead SaaS operations and am a big believer in product-led growth. I share SaaS earnings analysis and growth strategy here. If you work on SaaS, I’d love to talk shop.