Like most people outside of the financial industry, you may be wondering: What is Snowflake?
Essentially, Snowflake is a cloud-based artificial intelligence powered data platform designed to integrate enterprise and operations data at scale to break down data siloes and enable enterprise-wide analytics and allow users to collaborate securely with customers and partners.
In early June 2024, I was invited to attend Snowflake’s Data Cloud Summit in San Francisco. I was curious to learn about what Snowflake’s data cloud services could offer to the manufacturing industries, especially considering that cloud data storage and analytics services are offered not just from the likes of Amazon, Microsoft and Google but also numerous automation technology suppliers. And seeing several well-known manufacturers—from ABB, Honeywell and Siemens to Sub-Zero, Toyota, 3M and ExxonMobil—are partners and/or users of Snowflake’s products, only served to pique my interest in this technology provider that I hadn’t heard much about before.
Following are excerpts from my interview with Tim Long, Snowflake’s global head of manufacturing, to help the Automation World audience understand what Snowflake is and what they offer to the manufacturing industries.
David Greenfield, Automation World: For manufacturers who are already using cloud and artificial intelligence analytics with AWS, Microsoft Azure or Google Cloud, can you explain where Snowflake sits with them since they're also all Snowflake partners?
Tim Long, Snowflake: Snowflake is a managed service that runs on all three of those hyperscalers. We provide a different experience in that we're really trying to give you everything that you need in one product, rather than having to pick out all the parts that you need to put together your own solution. Snowflake is also one of the biggest customers of each of these cloud providers and they're also partners of ours. So, we work together in one ecosystem to provide a managed service experience.
David Greenfield, Automation World: Many automation technology suppliers also provide cloud data management and AI services. Where does Snowflake fit with those? Would it be more competitive or does it fit along with those like it does with AWS, Microsoft and Google?
Tim Long, Snowflake: It is very complementary. We're building partnerships with many of these suppliers. Our role is to, first and foremost, provide the data foundation for our customers where they can integrate not just their OT data, but also you blend that together with traditional IT data, which is a stronghold for Snowflake as we were early to the cloud. And OT is, in many cases, still sorting out the role of the cloud for operations data. We’re building partnerships that really leverage Snowflake’s strengths as an AI and data product while our partners leverage those strengths to deliver services to their customers.
David Greenfield, Automation World: A decade ago, when you would talk to manufacturers about moving their operations data to the cloud, it was not in their plans. It was very much like 20 years ago, when most manufacturers were wary of Ethernet use of on the plant floor. Given the manufacturing industry’s move toward the cloud in recent years, what should manufacturing technology end users—from the executive level on through to plant and engineering management—know about Snowflake that sets you apart in this really crowded market of cloud providers?
Tim Long, Snowflake: Snowflake is focusing on enterprise AI. Our strategy is to make it incredibly simple and fast for our customers to bring AI to their data. That's really our focus right now. We've been known for a long time for our strengths in delivering data products of all kinds and many manufacturers use us to integrate multiple ERP systems acquired through mergers and acquisitions. But we don't just stop with data integration. We're continuing to innovate in AI capabilities, native to Snowflake, that make it easy for customers—once they bring their data to Snowflake—to get the answers they’re looking for in a very efficient way.
To build these advanced models with AI, you have to have the scale of the cloud. And our model is a consumption-based model that doesn’t require users to license the product, they just pay for what they use. That allows them to spin up an insane amount of resources for a particular problem and then shut it off and only pay for what they use. If a manufacturer was trying to do this on premises, they would have to size up their infrastructure to be able to handle the maximum peak workloads. But this kind of massive compute environment wouldn’t be busy all the time. So, the manufacturer would have to deal with depreciating those assets that are not fully used. Snowflake’s managed service consumption model allows customers have better control over their costs while always having access to the resources they need when they need them.
David Greenfield, Automation World: Walking through the exhibit halls here at the Snowflake Data Cloud Summit event, it seems like a lot of exhibitors here offer a lot of the same things that Snowflake does—from data governance to AI analytics. Can you clarify how they fit with Snowflake, because it seems a little bit like the peak dot-com days where there were a lot of companies doing a lot of the same things?
Tim Long, Snowflake: Think of Snowflake as a platform. Our AI Data Cloud is a platform and it's a natural place for our partners to build on top of that. Some of our customers may choose to build their own solution using our technology. Others may choose one of our partners who have a special capability built for Snowflake. We have so many partners that have different strengths and capabilities. Part of my role at Snowflake is to build the right partnerships that really lean in and build on Snowflake’s strengths to help our customers get the most value from their data.
David Greenfield, Automation World: Since Snowflake started out in the financial industry, how long has the company been focusing on manufacturing and what instigated Snowflake’s expansion into the manufacturing industry?
Tim Long, Snowflake: Financial services continues to be a great industry for Snowflake and our largest in terms of contribution to revenue. But manufacturing is growing incredibly quickly, and it's an exciting place to be. I believe I was the first employee at Snowflake to have manufacturing in my job title. That was about three years ago. We've since built a great team of industry experts, whether it's industry principals that are part of my team or field CTOs who have manufacturing and supply chain expertise, and a sales team that's focused on the industry.
David Greenfield, Automation World: The presentations by Sub-Zero, 3M and Toyota here at the Snowflake event were very interesting in how the companies are connecting enterprise and operations data with Snowflake. Can you highlight some other manufacturing applications where operations data is being analyzed to improve production operations?
Tim Long, Snowflake: Yes, I'll give you a great example of a semiconductor manufacturer where their manufacturing is outsourced using a foundry model. And their collaboration with the foundry is critical to having very short-cycle decision making. As the wafers move through the fab, they need to be able to make decisions in near real time on how to route and process that wafer. To do this, they collaborate closely with their foundry. Using data in Snowflake, the chip designer will recommend next actions in the manufacturing flow. They're actually bringing in data from the foundry’s shop floor and they're running predictive models in Snowflake and then they're making recommendations based on the results to direct the next step for that wafer in its production process. It's like production workflow and collaboration across two different companies.
David Greenfield, Automation World: One term I've heard a good bit here at the event is “end-to-end.” Can you define what that means for Snowflake because, in manufacturing, end-to-end can mean a lot of different things?
Tim Long, Snowflake: This idea of destroying the data silo has been Snowflake’s mission for a long time. And business problems these days are so much more complicated than any one specific function. They really span throughout the business. Integrating data end to end is one thing that we talk about as it relates to our AI capabilities. We say we have end-to-end capabilities there because we have all the tools you need to develop a solution that is deployed and managed inside Snowflake as a full lifecycle solution with no additional infrastructure required while keeping your business logic, your intellectual property and data all under one secure governance roof.
David Greenfield, Automation World: Is there anything else that manufacturers should know about Snowflake since, for most of them, Snowflake is likely a new name?
Tim Long, Snowflake: I would say the most important thing for manufacturers is to think of Snowflake as the “easy button” for integrating data IT and OT (operations technology) with the ability to connect data in one governed and secure environment that gives you all the horsepower you need to do predictive yield modeling, root cause analysis and optimization of factory networks. All of these capabilities can be developed and delivered in Snowflake faster than ever.
