娇色导航

Two vendor names keep popping up when CIOs describe how they launched successful AI projects.

The two companies, Databricks and Snowflake, started from different market positions and technical perspectives, with Databricks focused more on unstructured data processing and real-time analytics, while Snowflake has concentrated on abstracting and simplifying data warehousing in the cloud.

But as both vendors have grown, they’ve become direct competitors in the emerging market of , which have become valuable tools for standing up and supporting AI and other automation projects.

Research firm Gartner defines platforms within the category as integrated sets of code-based libraries and low-code tooling. Offering wide-ranging functionality, these platforms enable data scientists to collaborate with IT and business leaders throughout the data science lifecycle, including business understanding, data access and preparation, model creation, and sharing of insights.

The generative AI explosion of the past three years has greatly expanded the data platform market, with enterprises across many industries turning to these platforms to help transform key business processes and operations.

AI on steroids

Databricks was focused on machine learning and traditional AI before the recent gen AI explosion, which accelerated the data platforms market like a “steroid shot,” says , Databricks’ co-founder and vice president of engineering. This is because organizations have begun to realize true value and ROI comes from organizing and mining their own data, he says.

“The message is starting to get out in terms of your own data and how the most successful AI projects are building on your own internal data, not using some ChatGPT answers mined from the internet someplace,” Wendell says.

While Databricks and Snowflake have plenty of competition in the space, the two vendors have found niches that continue to earn them new customers, says , CEO at AI-driven audio company Datavault AI.

Snowflake excels in ease of use, security, and structured data analytics at scale, Bradley says, while Databricks is a developer-first environment with strong machine learning an AI tooling and support for multimodal workloads.

“What’s compelling about both companies is how they’ve moved up the stack to become more than just storage or processing platforms — they’re now integrated AI orchestration environments,” Bradley adds. “Enterprises want platforms that not only store and query data but also enable collaborative model training, data versioning, and production-scale inference.”

Cloud interoperability

The two vendors also win points for being cloud agnostic, adds , CTO at data and IT consulting service provider Resultant. Both run on each of the three major cloud providers, he notes.

The move of both companies to a platform approach is also a selling point, he says. “Both have expanded to be end-to-end platforms that allow for data storage, analysis, and visualization,” he explains. “This simplifies security, patching, and operation of the platform.”

Ease of use is a big advantage for these data analytics platforms, adds , COO and chief data and analytics officer at finance industry SaaS provider and Snowflake customer TS Imagine.

“It’s like this one-stop shop,” he says. “I don’t have to move around to many different platforms and technologies because I have everything in one place; I can do SQL, I can do big data with trillions of rows, I can do fast queries, and all of the LLMs run natively there.”

With Snowflake, Bodenski doesn’t need a database administrator or a systems administrator, he says. “My data engineers and my data science self-service themselves, and it’s near zero maintenance,” he adds.

The platform approach, along with tech agnosticism, is key to success for both Databricks and Snowflake, says , an AI and automation analyst at IDC.

“They’re both trying to offer that unified data platform for their end user to make it as easy as possible to use for complex tasks, like AI, and they’re all trying to with that separation of storage and compute,” he says. “You’re seeing not just these two vendors, but a lot of other vendors, basically being storage agnostic, and it’s wherever your data is, whatever kind of data, we’re going to connect it together.”

Under this platform approach, the Databricks Data Intelligence Platform includes business intelligence, data warehousing, AI and data science functionality, real-time analytics, and orchestration. Snowflake’s platform allows customers to build data pipelines, run analytics, create and deploy AI and machine learning tools, and share live data across clouds and organizations.

An evolving marketplace

In addition to the platform approach, simplicity is a key ingredient of Snowflake’s success, says , the company’s CEO. Two of the company’s cofounders came from Oracle and saw how complicated systems could quickly bog down customers, he says.

“Even before AI, part of the product philosophy was that technology should be simple and easy to use,” Ramaswamy says. “They set about creating a product that didn’t come with a huge number of knobs that had to be continuously tuned. They believe in self-tuning systems.”

Integration with other products, including multiple AI models, is part of that philosophy, he adds. “We attach a lot of importance to product cohesion and product simplicity, meaning that we pay a lot of attention to every part of our product working with other things,” Ramaswamy adds.

Databricks embraces a data analytics platform approach, but the company also recognizes that customers demand interoperability with AI models and other tools, Databricks’ Wendell says.

“The key to interoperability is that the most critical interfaces need to be open source because, first of all, it’s a lock-in issue” he adds. “No one wants to lock their storage into proprietary format, and we can’t convince customers to bet on us if the core things are not open.”

The future looks bright for both Databricks and Snowflake, but the data science and machine learning platform market is far from settled, with competition coming from several corners, including the cloud providers themselves, with Google offering its BigQuery product and Microsoft pushing its Fabric tool, Resultant’s Bolles notes. Several other smaller vendors also compete in the space.

The market is evolving rapidly, he says, and user organizations will turn to vendors to ensure their data governance and cataloging are ready to support the next evolutions of AI.

The next big evolution in generative AI is self-service business intelligence, Bolles adds. “This puts the power to interrogate the data and build data visualizations in the hands of the end user rather than requiring a BI engineer to create a visualization,” he says.

IDC’s Pratt sees partnerships as key to growth in the data science and machine learning platform space, with best-of-breed databases and best-of-breed data governance tools working together to provide a full-fledged data and analytics solution.

“The platform guys obviously have a little bit of a head start, and it’s easier to incorporate internally gen AI, and now agentic AI, into those systems, because it’s all under one vendor,” he says.