 Welcome back to SuperCloud 2, this is Dave Vellante. We're here exploring the intersection of data and analytics in the future of cloud and data. In this segment, we're going to look deeper into the life sciences business with Jesse Cugliotta who leads the healthcare and life sciences industry practice at Snowflake and Nicholas Nick Taylor, who's the executive director of informatics at Iona's Pharmaceuticals. Gentlemen, thanks for coming in the Cube and participating in the program, really appreciate it. Thank you for having us, Dave. You're very welcome. Really try to look at data sharing as a use case and try to understand what's happening in the healthcare industry generally and specifically how Nick thinks about sharing data in a governed fashion, whether tapping the capabilities of multiple clouds is advantageous long-term or presents more challenges than the effort is worth. And to start, Jesse, you lead this industry practice for Snowflake and it's a challenging and vibrant area. It's one that's hyper focused on data privacy. So the first question is, you know, there was a time when healthcare and other regulated industries wouldn't go near the cloud. What are you seeing today in the industry around cloud adoption and specifically multi-cloud adoption? Yeah, for years, I've heard that healthcare and life sciences has been cloud averse, but in spite of all of that, if you look at a lot of aspects of this industry today, they've been running in the cloud for over 10 years now, particularly when you look at CRM technologies or HR or HCM, even clinical technologies like EDC or ETMF. And it's interesting that you mentioned multi-cloud as well because this has always been an underlying reality, especially within life sciences. This industry grows through acquisition where companies are looking to boost their future development pipeline, either by buying up smaller biotechs, they may have like a late or a mid-stage promising candidate. And what typically happens is the larger pharma could then use their commercial muscle and their regulatory experience to move it through approvals and into the market. And I think the last few decades of cheap capital certainly accelerated that trend over the last couple of years. But this typically means that these new combined institutions may have technologies that are running on multiple clouds or multiple cloud strategies in various different regions to your point. And what we've often found is that they're not planning to standardize everything onto a single cloud provider. They're often looking for technologies that embrace this multi-cloud approach and work seamlessly across them. And I think this is a big reason why we here at Snowflake, we've seen such strong momentum and growth across this industry because healthcare and life sciences has actually been one of our fastest growing sectors of the last couple of years. And a big part of that is in fact, we run on not only all three major cloud providers, but individual accounts within each and any one of them, they have the ability to communicate and interoperate with one another like a globally interconnected database. Great, thank you for that setup. And so Nick, tell us more about your role and Iona's pharma, please. Sure, so I've been at Iona's for around five years now. You know, when I joined, it was the IT department was pretty small. There wasn't a lot of warehousing, there wasn't a lot of kind of big data there. We saw an opportunity with Snowflake pretty early on as a provider that would be a lot of benefit for us. You know, because we're small, wanted something that was fairly hands off. You know, I remember the days where you had to get a lot of DBAs in to fine tune your databases, make sure everything was running really, really well. The notion that there's, you know, no indexes to tune, right? There's no, there's very few knobs and dials you can turn on Snowflake. That was appealing that, you know, it just kind of worked. So we found a use case to bring the platform in. We basically used it as a logging replacement, as a splunk kind of replacement with a platform called Elysium Analytics. There's a way to just get it in the door and give us the opportunity to solve a real-world use case, but also to help us start to experiment using Snowflake as a platform. It took us a while to, A, get the funding they get, bring it in, but B, build the momentum behind it. But, you know, as we experimented, we added more data in there. We ran a few more experiments. We piloted in a few more applications. We really saw the power of the platform and now we are becoming a commercial organization. And with that comes a lot of major data sets. And so, you know, we really see Snowflake as being a very important part of our ecology going forward to help us build out our infrastructure. Okay, and you are running, your group runs on Azure, it's kind of monocloud, single cloud, but others within Ionis are using other clouds, but you're not currently collaborating in terms of data sharing. And I wonder if you could talk about how your data needs have evolved over the past decade? I know you came from another highly regulated industry in financial services. So what's changed? You sort of touched on this before. You had these very specialized individuals who were DBAs and could tune databases and the like. So that's evolved. But how has generally your needs evolved? Just kind of make an observation over the last five or seven years. What have you seen? Well, I wasn't in a group that did a lot of warehousing. It was more like online trade capture, but it was very much on-prem. Being in the cloud is very much a dirty word back then. I know that's changed since I've left, but we had major, major teams of everyone who could do everything. As I mentioned, in the pharma organization, there's a lot fewer of us. So the data needs there are very different. We have a lot of SaaS applications. One of the difficulties with bringing a lot of SaaS applications on board is obviously data integration. So making sure the data is the same between them, but one of the big problems is joining the data across those SaaS applications. So one of the benefits, one of the things that we use Snowflake 4 is to basically take data out of these SaaS applications and load them into a warehouse so we can do those joins. So we use technologies like Boomi. We use technologies like 5Tran, like DVT, to bring this data all into one place and start to kind of join that, basically allow us to do run experiments, do analysis, basically find better use for our data that was siloed in the past. Yeah, and just to add on to Nick's point there, that's actually something very common that we're seeing across the industries because a lot of these SaaS applications that you mentioned, Nick, they're from vendors that are trying to build their own ecosystem in the walled garden. And by definition, many of them do not want to integrate with one another. So from a data platform vendors perspective, we see this as a huge opportunity to help organizations like IONUS and others kind of deal with the challenges that Nick is speaking about because if the individual platform vendors are never going to make that part of their strategy, we see it as a great way to add additional value to these customers. This data sharing thing is interesting. There's a lot of walled gardens out there. Oracle's a walled garden. AWS in many ways is a walled garden. Microsoft has its walled garden. You could argue Snowflake is a walled garden. But what we're seeing and the whole reason behind the notion of super cloud is we're creating an abstraction layer where you actually in this case, or this use case can share data in a governed manner. Let's forget about the cross cloud for a moment. I'll come back to that. But I wonder, Nick, if you could talk about how you're sharing data. Again, Snowflake sort of, I look at Snowflake like the app store, Apple, we're going to control everything. We're going to guarantee with data clean rooms and governance and the standards that we've created within that platform. We're going to make sure that it's safe for you to share data in this highly regulated industry. Are you doing that today? And take us through the considerations that you have in that regard. So it's kind of early days for us in Snowflake in general, but certainly in data sharing. We have a couple of examples. So data marketplace, that's a great invention. It's been a small IT shop again, right? The fact that we're able to just bring down terabyte size data sets straight into our Snowflake and run analytics directly on that is huge, right? The fact that we don't have to FTP these massive files around run jobs that may break, being able to just have that on tap is huge for us. We've recently been talking to one of our CRO feeds, CRO organizations about getting their data feeds in. Historically, this is clinical trial data that comes in on an FTP file. We have to process it, take it through the platforms, put it in the warehouse, but one of the CROs that we talked to recently when we were reinvestigating what data opportunities they have, they were a Snowflake customer and we are, I think, the first production customer they have have taken that feed. So they're basically exposing their tables of data that historically came in these FTP files directly into our Snowflake instance. Now, we haven't taken advantage of that. It only actually flipped the switch about three or four weeks ago, but that's pretty big for us again, right? We don't have to worry about maintaining those jobs that take those files in. We don't have to worry about the jobs that take those and shove them on the warehouse. We now have a feed that's directly there that we can use a tool like DBT to push through directly into our model. And then the third avenue that came up actually fairly recently as well was genetics data. So genetics data that's highly, highly regulated. We have to be very careful with that. And we had a conversation with Snowflake about the data white rooms practice. And we see that as a pretty interesting opportunity. We're having one organization run genetic analysis be able to send us those genetic data sets, but then there's another organization that's actually has the, in quotes, metadata around that. So age, ethnicity, location, et cetera. And being able to join those two data sets through some kind of mechanism would be really beneficial to the organization. Being able to build a data white room so we can put that genetic data in a secure place, anonymize it and then share the amalgamated data back out in a way that's able to be joined to the anonymized metadata. That could be pretty huge for us as well. Okay, so this is interesting. So you talked about FTP, which was the common way to share data. And so you basically it's a, I got it, now you take it, do whatever you want with it. Now we're talking, Jesse, about sharing the same copy of live data. How common is that use case in your industry? It's become very common over the last couple of years. And I think a big part of it is having the right technology to do it effectively. You know, as Nick mentioned, historically, this was done by people sending files around. And the challenge with that approach, of course, while there are multiple challenges, one, every time you send a file around Europe, by definition, creating a copy of the data, because you have to pull it out of your system of record, put it into a file, put it on some server where somebody else picks it up. And by definition at that point, you've lost governance. So this creates challenges and general hesitations to doing so. It's not that it hasn't happened, but the other challenge with it is that the data is no longer real time. You know, you're working with a copy of data that was as fresh as at the time at that when that was actually extracted. And that creates limitations in terms of how effective this can be. What we're starting to see now with some of our customers is live sharing of information. And there's two aspects of that that are important. One is that you're not actually physically creating a copy and sending it to someone else. You're actually exposing it from where it exists and allowing another consumer to interact with it from their own account. That could be in another region, someone running in another cloud. So this concept of super cloud or cross cloud could be coming realized here. But the other important aspect of it is that when that other entity is querying your data, they're seeing it in a real time state. And this is particularly important when you think about use cases like supply chain planning where you're leveraging data across various different enterprises. If I'm a manufacturer or if I'm a contract manufacturer and I can see the actual inventory positions of my clients, of my distributors, of the levels of consumption at the pharmacy or the hospital, that gives me a lot of indication as to how my demand profile is changing over time versus working with a static picture that may have been from three weeks ago. And this has become incredibly important as supply chains are becoming more constrained and the ability to plan accurately has never been more important. Yeah, so the race is on to solve these problems. We started with, hey, we're going to simplify database, we're going to put it in the cloud, give virtually infinite resources, separate compute from storage, okay, check, we got that. Now we've moved into sort of data clean rooms and governance and you've got an ecosystem that's forming around this to make it safer to share data and then Nirvana, or at least near-term Nirvana is, we're going to build data applications and we're going to be able to share live data and then you start to get into monetization. Do you see, Nick, in the near future, where I know you've got relationships with, for instance, big farmer like AstraZeneca, do you see a situation where you start sharing data with them, is that in the near-term, is that more long-term? What are the considerations in that regard? I mean, it's something we've been thinking about, we haven't actually addressed that yet. Yeah, I could see situations where, you know, some of these big relationships where we do need to share a lot of data, it would be very nice to be able to just flick a switch and share our data assets across to those organizations, but, you know, that's a ways off for us now. We're mainly looking at bringing data in at the moment. One of the things that we've seen in financial services in particular, and Jesse, I'd love to get your thoughts on this, is companies like Goldman or Capital One or NASDAQ taking their stack, their software, their tooling, actually putting it on the cloud and facing it to their customers and selling that as a new monetization vector as part of their digital or business transformation. Are you seeing that, Jesse, at all in healthcare? Is it happening today or do you see a day when that happens or is it healthier or just too scary to do that? No, we're seeing the early stages of this as well, and I think it's for some of the reasons we talked about earlier, you know, it's a much more secure way to work with a colleague you don't have to copy your data and potentially expose it. And some of the reasons that people have historically copied that data is that they needed to leverage some sort of algorithm or application that a third party was providing. So maybe someone was predicting the ideal location to run a clinical trial for this particular rare disease category where there are only so many patients around the world that may actually be candidates for this disease. So you have to pick the ideal location. Well, sending the data set to do so, you know, would involve a fairly complicated process similar to what Nick was mentioning earlier. If the company who is providing the logic or the algorithm to determine that location could bring that algorithm to you and you run it against your own data, that's a much more ideal and a much safer and more secure way for this industry to actually start to work with some of these partners and vendors. And that's one of the things that we're looking to enable going into this year is that, you know, the whole concept should be bring the logic to your data versus your data to the logic and the underlying sharing mechanisms that we've spoken about are actually what are powering that today. So thank you for that, Jesse. And so Nick, go ahead, please. Yeah, if I could add to that, that's something certainly we've been thinking about. And in fact, we'd started talking to Snowflake about that a couple of years ago, we saw the power there again of the platform to be able to say, well, could we, we were thinking in more of a data share, but could we share our data out to say, AI ML vendor, have them do the analytics and then share the data they result back to us. Now, you know, there's more powerful mechanisms to do that within the Snowflake ecosystem now, but, you know, we probably wouldn't need to have onsite AI ML people, right? Some of that stuff's very sophisticated, expensive resources, hard to find, you know, it's much better for us to find a company that would be able to build those analytics, maintain those analytics for us. And, you know, we saw an opportunity to do that a couple of years ago and we're kind of excited about the opportunity there that we can just basically do it with a no op, right? We share the data right, we have the analytics done, we get the result back and it's just fairly seamless. I mean, I could have a whole nother cube session on this guys, but I mean, I just did a session with Andy Terai at Constellation Research about how difficult it's been for organizations to get ROI because they don't have the expertise in-house. So they want to either outsource it or rely on vendor R&D companies to inject that AI and machine intelligence directly into applications. But my follow up question to you, Nick, is when you think about, because Jesse was talking about, let the data basically stay where it is and bring the compute to that data. If that data lives on different clouds, and maybe it's not your group, but maybe it's other parts of Ionis or maybe it's your partners like AstraZeneca or the AI ML partners, and they're potentially on other clouds or that data is on other clouds. Do you see that, again, coming back to SuperCloud, do you see it as an advantage to be able to have a consistent experience across those clouds? Or is that just kind of get in the way and make things more complex? What's your take on that, Nick? Well, from the vendor, from the client side, it's kind of seamless with Snowflake for us. So we know for a fact that one of the data sets we have at the moment, Compile, which is the large multi-terabyte data set I was talking about, they're on AWS on the East Coast and we're on Azure on the West Coast. And they had to do a few tweaks in the background to make sure the data was pushed over, but from my point of view, the data just exists, right? So for me, I think it's hugely beneficial that Snowflake supports this kind of infrastructure, right? We don't have to jump through hoops to like, okay, well, we'll download it here and then we upload it here. They already have the mechanism in the background to do these multi-cloud shares. So it's not important for us internally at the moment. I could see potentially at some point where we start linking across different groups in the organization that do have maybe Amazon or Google Cloud, but certainly within our providers. And we know for a fact that they're on different services at the moment and it just works. Yeah, and we learned from Benoit Dajaville who came into the studio on August 9th with the first SuperCloud in 2022 that Snowflake uses a single global instance across regions and across clouds. Whether or not you can query across big regions just depends, right? It depends on latency. You might have to make a copy or maybe do some tweaks in the background. But guys, we got to jump. I really appreciate your time. Really thoughtful discussion on the future of data and cloud specifically within healthcare and pharma. Thank you for your time. Thanks for having us. All right, this is Dave Vellante for the CUBE team and my co-host John Furrier. Keep it right there for more action at SuperCloud 2.