Databricks Data + AI Summit 2026 | Stephen Orban, Databricks

Clips
More from Databricks Data + AI Summit 2026

Stephen Orban

SVP, Product Ecosystem & Partnerships

Databricks

play_circle_outline Databricks Summit: AI Infrastructure at Scale and the Coming Software Revolution

play_circle_outline Databricks Marketplace, Genie Agent Workflows, On‑Prem Storage Integration, and Shipping Apps to Customer Endpoints

play_circle_outline From Best‑of‑Breed to Consolidated Platforms: Unifying Data Across Architectures and Systems

play_circle_outline Deep co‑design partnerships (Databricks and Sigma example)

play_circle_outline Agentic Workflows and Trusted Data: Unity Catalog Governance, Context, Control, and Choice for 24/7 Productivity

Info
Transcript

Stephen Orban, Databricks

Stephen Orban

SVP, Product Ecosystem & Partnerships Databricks

In this interview from Databricks Data + AI Summit 2026, Stephen Orban, senior vice president of product ecosystem and partnerships at Databricks, joins Mike Palmer, chief executive officer of Sigma Computing, to talk with theCUBE's John Furrier about the consolidation of enterprise data stacks and the shift from isolated AI experiments to governed, production-ready intelligence at scale. Palmer argues that the enterprise software market is at an inflection point, with hundreds of standalone SaaS contracts collapsing into platform-led solutions — a shift enab... Read more

explore Keep Exploring

What was the focus and key message of Databricks' Data and AI Summit keynote in San Francisco? add

What did you announce this week to give customers more choice in the partner ecosystem and marketplace? add

What major change have you observed in enterprise software strategy or architecture in the past year or so? add

How does your company approach and differentiate its partnership with Databricks to ensure you tailor Sigma to Databricks’ platform and new technologies so customers get the full benefit? add

What are the main drivers behind the shift toward trusted data, workflow investments, and agentic workflows? add

bolt Powered by CUBE AI

Stephen Orban, Databricks

search

John Furrier

>> Welcome back to theCUBE's livestream here in San Francisco at Databricks's Data and AI Summit. I'm John Furrier, host of theCUBE. Another banner year for Databricks, 31,000 plus attendees, at capacity, busting out at the seams. And really, this show really is a symbol of the AI infrastructure buildup that's enabling a software revolution with hardware, agentic, and data's at the center of it. That's the focus of the keynote, and all the new technology, as the data states have to unify. That's the big message. And there's a variety of other things around it. And what's going to happen is AI will come in, and it's about injecting intelligence into the business, into the process, into the workflows. We have two veterans of the industry here to break it down for me. Stephen Orban, senior vice president, product, ecosystem, and partnerships at Databricks. Mike Palmer, CEO of Sigma, both Cube alumnis. Guys, thanks for coming on theCUBE. Appreciate it.

Stephen Orban

>> John, thanks for having me as always. Mike, thanks for doing this.

Mike Palmer

>> Absolute pleasure.

John Furrier

>> As I said, senior veterans, grizzly veterans, that's me. You guys are much more established senior people. Okay. First thing is, what's happening here? Because I just did a little monologue. We're seeing a transition that's, we've already crossed the threshold. It seems to be like a generational shift around computer architecture, data architecture. And while frothy, highly-engaged developer community, open source, just AI native is here. And so shit's got to get reorganized. This is what I'm seeing here. What's your thoughts, Mike? You're in the middle of it.

Mike Palmer

>> I think you described it well. I mean, often big technology transformations start with hardware, and then they go up from that point. And Databricks has done an outstanding job bringing together kind of the data science spark side, which they were super early with. Bringing on that business side with the warehouse, that was that next layer up. That's where data now sits. Where I think the market is about to embark on a huge transformation, then, is all of the software silos that traditionally have existed in the data center and then later with SaaS. I think a lot of that's about to be re-architected, and I think that rethinking of how the business operates the business, whether it's workflows that are rebuilt in AI, or replacing third-party software, or how we drive productivity, that's the greenfield now. And that's the exciting part to be in this part of the market.

John Furrier

>> There used to be an expression, we used to say in theCUBE all the time, the three pillars of the enterprise, or the holy Trinity, whatever the expression was: storage, networking, and compute. Okay. But now you have a fourth dimension, database. And I put that loosely, data. So now hyper-converge, unconverged, we see Nvidia. Guys, talk about this dynamic. Because on the marketplace you're doing a lot of transactions. There will be a lot more, because when you unlock that database and integrate it into the three other pillars, you now have fourth estate in play here.

Stephen Orban

>> Yeah. Some might call it choice. As Ollie talked about in the keynote, it's control, context, and choice. And so my role as leading the ecosystem is to really make sure our customers have the maximum amount of choice of the solutions that they want to use as they modernize their data and AI estate. I'd say there's probably four or five things we rolled out as it relates to helping customers have the maximum amount of choice for the partner community this week. The first, and I'd say the flagship announcement for us is really now rolling out our marketplace transactable capabilities. And thanks Mike and Sigma for being a part of that launch with us.

Mike Palmer

>> Glad to be there.

Stephen Orban

>> Now our customers who have universal commitments with us can use their pre-committed spend that they have made for Databricks also towards participating third parties like Sigma, so that's one. We also now made it possible for anybody who builds a Databricks app to ship that app to their customer endpoints, so that they can ship their IP to customer data, instead of making their customers ship their data to them. So that's a big unlock. We also announced Genie agent sharing. So not any of our customers who are creating Genie agents on top of Databricks can share that with their customers. And then I'd say the two other things that I'm pretty excited about, one of which I'm sure Sigma will benefit from as well, we announced a storage ecosystem where many of our customers, there's like hundreds of exabytes of data that still sits in on premises environments, you talked about compute, networking, storaging. So that storage that sits with vast MinIO, Qumulo, IBM, et cetera, we're now making those data sets that are sitting on prem available to use in Databricks using our storage ecosystem, because those vendors have implemented our open sharing protocol. So lots of exciting new data for our partners to use and capabilities for them to transact with their customers as well.

John Furrier

>> Yeah. When you mentioned some of those names, Qumulo, VAST, we know those companies. The partnerships that they used to do was, "Hey, we support your platform. We run your software store stuff." In the modern era, the partnership equation's changing. For instance, I was talking to the CEO of Qumulo at a Cisco event, they're like, "I can actually inject an agent, hit their agent, and within seconds, identify threats," and that's what they do. But they're now crossing their lane. They're moving out of their lanes, not staying in your lane. It's like, let's integrate. Mike, so this really hits the partnership. What's it like to be a partner in the modern era, because you got to think differently, got to move fast, not be afraid to do stuff. If Databricks comes out, "We'll evaluate it." No, you got to jump on it, because the speed game is fast, but also integrating. What's your thoughts on this? You've seen the old way, now we're in the modern era.

Mike Palmer

>> The funny thing is, I think it's still the old way in some ways. And I hate to be academic here, but if you go back to Adam Smith, this is a division of labor thing. And if Databricks creates a great asset, Agent Bricks, and we can consume that and not have to replicate it on behalf of a customer, but we can expend it in some way, that's our job. We aren't here to compete, to be redundant. We are here to take the best of what's available, build on top of that, and that is a never-ending cycle. I think the only thing that could be said about the new way versus the old way is that there's no way to build something and have it last for years anymore. All of us working in this space, we have one to two quarters visibility on what we're building. You have to respond to the ecosystem. Integrations are easier than they ever have been, and that just forces you to think, "Okay, what's the next turn of the screw?" or so to speak, on efficiency, on a new capability we can bring? Where did something better than we can do become available to us? And that all benefits customers.

John Furrier

>> What's changed in your mind? If you had to say in this era, in the past year or so, what's been the big change that's different?

Mike Palmer

>> I think there are a couple of changes hitting enterprise. One is, we are exiting a timeframe of what a lot of people call best of breed and we're moving more to platform end-to-end solutions. I think we're going to go away from companies that have hundreds of standalone software contracts. You're going to see those consolidated into 10s, where you're covering a lot more of the enterprise's requirements. They're going to benefit from cost savings, customization, compliance. And I think that's a fundamental shift in architecture that has been brought on by the fact that you can put all of your data, and now your premises data, and your unstructured, and your structured. That capability that Databricks brought unlocked a whole set of permutations, that's really what the change is coming. So I think that when we look forward a year or two years, we're going to see this steady consolidation into platform providers.

John Furrier

>> I like the Adam Smith angle, because I want to just talk about that for a second. Because the old way was we're running our business, IT was a department. IT projects ran a little bit differently, and you're basically getting at, "Hey, is this an asset?" So it's a business view, not an IT view.

Mike Palmer

>> That's right.

John Furrier

>> Which has been around, but now it's becoming much more proliferated. I mean, now I'm the line of business head, I might have the authority or agency to say, "Hey, you know what? I don't want to build that. I want a partner."

Mike Palmer

>> That's right.

John Furrier

>> What's your reaction to that? Do you see that as a key changeover, or is this just business the way it was? Or it's not, IT is still the same, or ...

Mike Palmer

>> IT is definitely not the same. Everyone is now a builder. That is the big economy change, right? This platform consolidation I was referring to, it's because everybody's a builder now, and that means people sitting in HR and marketing. One of the biggest demand signals that we are responding to is that you can vibe code applications, but what IT teams don't want is to relinquish the control and responsibility they still have for security, and governance, and reliability, right? Everyone wants freedom, but freedom requires a certain amount of structure. So we keep coming back to the combination of a Sigma and a Databricks gives absolute choice to that end user, while still making sure that that choice doesn't result in the sort of performance or security exposures that data leaving the Databricks ecosystem might cause.

John Furrier

>> Stephen, talk about this trend. Because based on that logic, you could argue that there's going to be a lot more partners. Because now if everyone's a builder, do you look at that as a tsunami coming, or you look at just like, continue the relationships? How are you thinking about the onboarding of partners? Because you have super partners, you have, I can self-serve. I just want to get the content.

Stephen Orban

>> I mean, first of all, I look at it as a big opportunity, but at the end of the day, my job is really to make sure our customers have any solution that they want to run really well on Databricks. And I think Mike, if I'm being frank, was being a little bit modest with the sort of investment that he's made in the partnership. I think what's changing in partnerships is it's no longer good enough to just be the same partner you are on one platform with another. And so because of the investments that Mike and his team have made to really differentiate how Sigma works on Databricks, versus, let's say, and of course he works with other data platforms as well, but he's really making the best of Databricks available inside of Sigma, and that really sets him apart and sets Sigma apart, and the solution to both our customers, and arguably just as importantly, our sales teams too. Because our sales teams know if they bring Sigma in to help a customer solve some big problem for them, a lot of the best parts of Databricks are going to come with it. Whereas some partners will say, "Well, I have to work the same on Databricks, and this data platform, on that data platform, and the other, and these guys don't do that."

John Furrier

>> So you're saying don't do a rinse and repeat because you're running a partner program. You're saying tailor it to the core competency or asset, what you said-

Mike Palmer

>> I'd go even further. I would say I want to be the first adopter of a new technology that Databricks brings to the market. I want to work most closely to make sure that we have our common eye on what the customer benefit is, so that if we make changes to make accommodation or best use of something Databricks has brought to the market, we are going to do that. We are not going to treat Databricks like anybody else, because a Databricks customer is not like anybody else. So we pride ourselves on that. We think about some of the innovation around Unity Catalog. I've already mentioned Agent Bricks. I mean, these are all areas where we feel like the combination in particular of what Databricks is bringing to the table gives us the ability to construct value that we may not be able to bring to anybody else.

John Furrier

>> You're building a company, you're scaling up. I mean, Steve, we talked about this before many times on theCUBE, and now it's more fashionable because it's relevant, is this co-design philosophy. It trickles down into the business level, technical level. So it really is a business relationship. So it's not just a partnership, it's teamwork, it's partnership, it's teamship. I think that word's been kicked around.

Mike Palmer

>> Teamship, all right.

John Furrier

>> Talk about that, because you're committed. You got money involved, you got personnel, you're betting the future, you're building products.

Mike Palmer

>> We're mutually committed. And Databricks is an investor in Sigma as well. Thank you for that, by the way. But this is, again, all about customers. We do this not for each other. We do this for customers who get value to pay. We think about customers like DraftKings, whose business is real time data. If you haven't seen DraftKings, you haven't seen one of the myriad sporting events that have been so exciting. By the way, in the last month, I can't avoid-

John Furrier

>> won two-nothing. I have tickets, but-

Mike Palmer

>> Ah....

John Furrier

>> Dave Vellante's going and I had to give up my ticket to come and cover Databricks. See how committed I am to the cause?

Mike Palmer

>> By the way, my side hustle is going to be creating a t-shirt that says, "Don't tell me the score." But having said that-

John Furrier

>> Exactly.

Mike Palmer

>> I watched ... I'm looking.

John Furrier

>> Live sports. Never record live sports.

Mike Palmer

>> working.

John Furrier

>> When your neighbors are cheering up, they won.

Mike Palmer

>> Exactly. But if I think about a DraftKings, their job is to take all of that excitement and move the line in real time. Their whole business is based on how well do they understand the right now data to offer the best line. And they chose Databricks and Sigma to make sure that that data is available to a combination of their data science team, and the people who have to get involved to make that a real thing.

John Furrier

>> You bring up a great point, and to kind of just go into the weeds a little bit on the enterprise is that you see big companies, innovators like Uber. They had to do a lot of things right on a lot of things. Column their database, real time, people, places and things. DraftKings, again, another one. There's going to be more. So this is becoming the shift or the mainstreaming of real time, data warehouse, data engineering, transactional business. That's what I think what makes this show unique, is that's software, that's data, that-

Mike Palmer

>> -...

John Furrier

>> so take us through some other examples that you're working with, because this is going to be more of the same, I think. You're going to want to be more like Uber, more like DraftKings, more like Palantir with ontologies. You're starting to see the pattern.

Mike Palmer

>> Well, I-

John Furrier

>> Well, I won't say Palantir, but ontologies, I mean-

Mike Palmer

>> a longer conversation, I'll kick down the road, but I think-

John Furrier

>> Don't go down that rabbit hole.

Mike Palmer

>> But I'll tell you, one of the customers I'll be engaged with even just later this evening, DraftKings people can sort of feel, because they watch the TV. But JP Morgan is making massive investments in Databricks to enable their team in Chase, their team in corporate finance, their team in wealth ... They are trying to rethink the bank. This is a bank that's been around, if you read a great book called 1929, you would be hearing about really the origins of the House of Morgan. But this is still, I think, the highest public capitalized bank in the world, who is trying to think about the next 50 years. They are putting their people onto live data. They're building portfolio models. They're doing compliance. They're doing data distribution to their teams. I mean, these are really massive security, compliance, performance, and yet end user enablement transformational efforts. So you can find it at the high end of the banking market. You can find it in the consumer market. Everybody is looking forward to this new data stack that's going to take them into the next day.

John Furrier

>> , in your opinion, because again, if you look at this era we're in now, is it gen AI? Is it agents? Is it all the above? How would you quantify or categorize the driving force?

Mike Palmer

>> I have a three tier thing, and Stephen can help me, because he's working in the ecosystem and he sees different angles on this. I talk about three areas of investment. There's no point in doing automations if your data is not accurate and trusted. So we think about, yes, your data has to be there, and it's great to see now some of the data can also be connected, but not there. How you build semantics on that data, so we think about Unity Catalog and Sigma data models on top of Unity Catalog. If you have built a data infrastructure you can trust, that, I always think about that as your house foundation. You can start building rooms and floors that go sky-high. So that is investment number one, trusted data. Two is workflows. I would say enterprise is a team sport. It's not an individual's game. So how do you build workflows that reflect competitive advantage in your business, that are cost-effective and scalable, that aren't aggregating 150 SaaS products? And that is the second area of investment. Agents is the most recent one. I think agentic workflows where you have individual productivity gains, where you have shared agents, where my productivity can be shared with you and you can take advantage of that. And now full workflows with agents operating in the background all the time, you now have a workforce that you're orchestrating that is a real time, 24-hour a day workforce. This is a massive change. We've always been constrained by the eight, 12, 15-hour, whatever workday. There's no constraint there anymore. So these are really big changes to rebuild our businesses on.

John Furrier

>> What's your take on the drivers?

Stephen Orban

>> Yeah. I mean, just to say it a little bit differently, I'd say context control and choice, and the context is all about making sure that your data is well-established, centralized in one place so that it can be then governed with the control on top of it. And I think that's actually probably the key reason over the last few years that most of our large enterprise customers have went wall to wall with Databricks, because they're able to catalog and govern who has access to what when. And now that agents have an insatiable appetite to get at this data, and then be used in tools like Sigma, we have this Unity AI capability that can also govern the agents and what models can be accessed under what circumstances.

John Furrier

>> When the shift left happened with cloud native many years ago, that turbocharged cloud native DevOps, DevSecOps, governance has that same kind of pattern. People who build governance into the CI/CD pipelining, get it right, have unlimited upside potential, to your point, get that foundation right. Mike and Stephen, I'd love to get your thoughts on this last question, because this is more of a little holistic organizational question. You run a company, you're building out, scaling up. What organizational elements change or disappear with agentic workers? Because if you get that governance right, you have a lot of productivity up and down the stack. Every layer of the stack has efficiencies, and revenue potential, or outcome-specific potential. What's the biggest areas you see that you're monitoring or have a visibility on that's going to change? Doesn't mean jobs are going to get lost. So like organizational management, you mentioned workflows, it's going to start going after things that are going to be easier taking friction out. What's your thoughts on changing organization?

Mike Palmer

>> Oh, first of all, there is nowhere in my organization that is not changing. If I have someone working in my company that says they are not changing, they probably are being changed faster than they realize.

John Furrier

>> They'll be changed out.

Mike Palmer

>> Because there's just no part of the business where you can't get some efficiency, where the entire ground-up workflow is going to change. That I think is pretty much assumed by everybody. We don't have to live in the world of zero-sum, where people have to be worried about this. We are a hundred percent year-over-year growth company. We need more all of the time, and the only way to scale that is to get each individual to do more, and AI is definitely fueling that. I think what we are most looking forward to is not necessarily just the conversation about how does one person do 150%. I think the question is, what thing were we not able to do at all a year ago, that is now going to be created for the first time because of these capabilities? That's where the true successful company is going to originate. It's that net new service that was not possible before, and we're at the very, very beginning, in my opinion, of entering that market.

John Furrier

>> Awesome.

Stephen Orban

>> Yeah. I'm just finding that curious people want to learn these new tools and want to put them to work, and they're actually able to get a lot more done. And at least me personally, I'm actually more excited to do some of the work that I do, because now, I'm an engineer by trade, but I haven't written production software in many years in my career, I suppose. But now-

John Furrier

>> You can....

Stephen Orban

>> I can. And I can make changes to Databricks. I can build apps with Sigma. I can use any of these vibe coding tools, and as long as I'm sitting on top of a well-governed data infrastructure like Databricks, I'm kind of free to do whatever that those governance policies allow me to do, and it's a lot more things that we're able to do now that, like Mike said, we were never able to do before.

John Furrier

>> The new mantra is demo, not memo.

Stephen Orban

>> Demos not memos. We love that.

John Furrier

>> Because you can actually, you can code it. On that book, 1929, there's a section where they talk about him walking down to Goldman Sachs, down Broad Street. I just got an apartment in that building. They turned it into condos.

Stephen Orban

>> Nice, all right.

John Furrier

>> I love that book. Great call out. Guys, thanks so much for coming on theCUBE. Very engaging, very insightful. Thanks for sharing.

Mike Palmer

>> Absolute pleasure.

John Furrier

>> As always, Stephen, thanks for having-

Stephen Orban

>> Good see you....

John Furrier

>> I'm John Furrier, host of theCUBE. We are live on the floor, Databricks' Data and AI Summit. The world is changing, the productivity organizations. Be curious and lean in. Be a partner. Move fast. That's the way the world's going. Of course, accuracy with the data, if you don't have the data right, it won't work. More coverage after this short break.