Orchestrate all the Things - Another globally distributed cloud native SQL database unicorn: Yugabyte raises $188M Series C funding at $1.3B valuation. Featuring CEO Bill Cook, Co-founder Karthik Ranganathan

Episode Date: October 28, 2021

A number of successive funding rounds have given Yugabyte unicorn status, while positioning the company to aim for a big piece of a growing pie in the shape-shifting database market Article published on ZDNet

Welcome to the Orchestrate All the Things podcast. I'm George Amadiotis and we'll be connecting the dots together. A number of successive funding rounds have given Ukebyte unicorn status while positioning the company to aim for a big piece of a growing pie in the shape-shifting database market. I hope you will enjoy the podcast. If you like my work, you can follow Link Data Orchestration on Twitter, LinkedIn, and Facebook. Thank you both for making the time for the conversation

today. And it's been a while since we last talked. And actually, I had to go back and check my records a little bit when was the last time. And I have to confess, I was a little bit confused because, well, you know, with all the funding and the different rounds that you've raised, I thought, so the last time was in 2020, and I thought that was your previous funding as well. But it turns out there was another one in between.

So, yeah, let's start by that. Yeah, I can help clarify that George. We did a, you know, we call it a B1 round at the beginning of this year, really based on the market opportunities. There were some investments we wanted to make that we weren't fully funded for with the previous B round. Namely, we wanted to invest more

in the fully managed service public cloud offering, right? If you think about that kind of the mass market play for you for you go by. And so we wanted to increase the focus in engineering on that effort. And then we also wanted to increase the focus in engineering on that effort. And then we also wanted to increase our go-to-market coverage. We were primarily in the U.S. and had just started in Europe and wanted to establish a presence in Asia, given the market opportunity in front of us and just being able to have global coverage.

So we did raise an additional $47 million at the beginning of the year, which candidly we had thought would be adequate and it was for another year and a half or so before we would start to contemplate our C funding round, which is what this announcement's about. However, kind of the same story. We were ahead of plan on all of our metrics. We were seeing good adoption by large enterprises. And given the market opportunity,

as you know, given the market opportunity, as you know, covering the database market, when a shift starts to happen, you know, it accelerates, you know, and you want to be known as the market leader. And so in our board meeting in, I guess it was towards the end of the summer, we started to talk about,

you know, potentially raising the C round and really just accelerating our growth plans across the company. You know, all aspects go to market, engineering, the rest of it. And we started that process, I guess it's maybe eight weeks ago now, approximately. And Karthik can be honest here, but I think it was about eight weeks ago now, approximately. And Karthik can be honest here, but I think it was about eight weeks ago or so. It was a quick process, yeah. Pretty quick. And, you know, the receptivity, I mean, you know how it goes. You want to, you know, we have a view about the opportunity

and, you know, our position, but, you know, the investors, you know, obviously validate that with writing checks and investing in the company. And, you know, obviously validate that with writing checks and investing in the company. And, you know, it went pretty fast and we were, you know, we got a very strong reception to kind of our position and market opportunity and the investment thesis and, you know, ultimately led to this C round, which you have the details on that Sapphire's leading and you see the raise and the rest of it. So does that fill in the gaps for you, George? Yes, yes, it does.

It does. Thank you. Yeah, I was wondering because I thought, okay, I must have missed something. Because that's quite an amount, I mean, in total that you have raised in a relatively short time. So I guess that speaks to validation that you just mentioned. It seems that investors must like something about what you do. Yeah, and candidly, we could have run the company more conservatively and waited longer to raise and the rest of it. But again, given the market opportunity, given the positioning and the validation we're seeing, you know, for the technology, we felt it was the right thing to do.

And obviously, you want to get the right valuation, the right investors around the table. So that's why we're quite pleased with, you know, if you look at the, you know, the key investors here being Sapphire, who, you know, both Sapphire and Meritech I had worked with before. They were both investors in Greenplum back, you know, a decade or so ago. So, you know, that familiarity helps, right, that they know the people, they know the company, they know they were like our leadership team and what Conant and Karthik and Mikhail had built. And then the two new investors outside of Sapphire and Meritech are really Alkion, who I didn't really know,

but Avi and the team had been tracking UgaByte for a few years. And, you know, they're a big crossover fund, even, you know, funding companies for the long term, you know, that have the potential to go public. And then one of our strategic customers, Wells Fargo, expressed interest in investing as well. So, I mean, from my perspective, it's a fantastic group of investors we have that are in it for the long haul, believe in our company, believe in us as the organization,

the people here, and want to fund us to win and lead the market. So it couldn't be a better scenario from my perspective. Karthik, anything you want to add to that? No, I think you got everything, Bill. I think it really comes down to the trends that are big shifts and really the opportunity in front. As organizations and comes down to the trends that are like, you know, big shifts and, you know, really the opportunity in front. As organizations and enterprises move to the cloud or start with the cloud, it's really that type of an opportunity.

Yeah, yeah, indeed. And, yeah, I guess there's not too many offerings around. If you want something that's, well, both SQL,, SQL Compact, let's say, and globally distributed. You're down to pointing fingers on one hand, basically, I think. Yeah, and it's interesting. I think our competitive positioning versus, you know, Cockroach is, you know, fairly validated in the market now.

If you think about our investment in Postgres and the SQL richness that we have, which large enterprises are looking for as they migrate off of traditional databases, Oracle, DB2, SQL Server, et cetera. So I think that's one positive thing. And then maybe to comment on, similar to what we talked about previously, George, the trends that we're seeing seeing the microservices, right?

Cloud native microservices is certainly driving the need for a distributed transactional database like you go by. And you know, that's a major trend, obviously new apps being born or built in the cloud. And then the third one that we talked about, which is more emerging, but actually we're very excited about, is the whole story around edge, right? If you think about edge and 5G and IoT more broadly,

it's driving a need for data consistency in different tiers, meaning it's not all going to reside in a single public cloud. By definition, it can't. And so that, and if you want a consistent transactional database across all of those tiers and not locked into any particular hyperscaler public cloud offering, to your point, George, there's only a few options and arguably we believe Uga Byte's

probably the strongest offering given our background in history. And maybe Karthik, you could comment, we've made this comment as well, that there was a lot of buzz around the NoSQL players, obviously Karthik being one of the initiators of that when they built Cassandra and HBase when they were at Facebook. But when you get to scale and you get to some of the attributes that people are looking

for in their applications as they grow and scale and they want the transactional consistency, there's a bit of a rejuvenation, if you will, of why SQL really matters and their need for something like UgoByte. Maybe Karthik, you can expound on that a bit. That's another trend that's happening. Yeah, absolutely, yeah. So I think we noticed this back at Facebook when we started building,

like I joined along with my co-founders. We met at Facebook in about 2007. And back then there were no SQL databases. And, you know, the biggest ask of the need of the hour at that time was to both federate queries as well as data. So query processing should be done on multiple nodes and your data should be distributed across nodes. There's really no database that was really available on commodity hardware that could do it. So as Bill said, we ended up building Apache Cassandra. We ended up building a lot of features into HBase

and putting it into production. We were also the team running this as a service for our internal Facebook customers at a pretty massive scale. It was like the scale is pretty comparable to public cloud scale. But from developers, what we heard again and again was they needed features like secondary indexes and more richer database features and joins and other types of queries and so

on and so forth. And what that made us realize was that, yes, we had solved the problem of distributing data and distributing query processing, but the ask one by one, if you kept layering that one after the other, it started going back to SQL, right? So we said at that point, look, we really need a database that supports SQL in order to be generally usable, and it needs to have all the underpinnings of NoSQL. So we really need to hybridize the two sides in order to build a database that can become

something like a default database for cloud native applications. And so fast forward to YugoByte, we made the decision that Postgres was absolutely on fire growing really fast. So we would pick that as the upper half and leave borrowed from the page of Amazon Aurora, which absolutely has been fantastic in its growth and the ability for end users to adopt it because it's really compatible with Postgres, not just at the protocol or wire compatibility level, but also at the features and the semantic level, how it functions, right?

So to that effect, we took the upper half of Postgres and said, let's make all of this feature set available in a cloud native shared nothing distributed world, right? Which is more like Google Spanner, and really hybridized that design into a single offering and made the offering as open source as Postgres itself. And on top of that, made sure that when people want to consume it for serious applications,

they have the ability to consume it as a managed service using Ugo Byte Cloud. So that really, I mean, it's a simple story at the end, but that's really the trajectory that we saw the world headed at and still heading in. And the hypothesis turns out to be true, which is what I guess this round and all of the growth is about.

Yeah, yeah, thank you Karthik. So actually you touched on a couple of points that I'd like to elaborate on. So cloud first and your managed offering, market managed service offering first of all. But before we go there, one last question on the, well, on financials and valuation and all that. So I saw in the press release that your valuation is now something like 1.3 billion. So I guess you're officially a unicorn if that means anything by now, because it's kind of over years maybe. But I was just curious, so how is that valuation derived?

Well, the valuation on, I guess, like if you look at the overall market and the investment, et cetera, right? Like the valuation comes from, you could say the three different buckets, the first bucket being enterprises that are adopting us and the expansions that come thereafter. The second bucket comes from our open source adoption and awareness. So the future potential for people to start running on us.

And then the third bucket comes from the ability to consume YugoByte really easily using a service, et cetera. So I think these are the three in terms of how people consume us and they think about. And then there's three separate trends that are fueling this as Bill was alluding to. The first one being microservices from the large enterprises. Second one being applications that are born in the cloud,

and this could be large enterprises or small companies. The third one being new generation of applications that are like, you know, being built at the edge that require geographic distribution of data built into the database as a first class citizen even to start, right? And the third bucket's early, but it's more common than we often think, right?

If you really think about a shopping experience today, you have storefronts and you have a central e-commerce app in the cloud and they have to work together, right? Similarly, vehicles are becoming smart vehicles, there's wearables, there's like so many other places where this type of distributed substrate starts to live. So to put that all together, right? Like I think this is really the thesis,

but I think Bill, go ahead and- Well, George, are you saying specifically how did the investors pick that valuation? Or are you asking the underpinning that Karthik described? Well, yeah, I mean, the underpinning was indeed elaborated quite eloquently by Karthik. I was wondering, yeah, who came up with that valuation, basically? Well, Sapphire. I mean, Sapphire did.

Sapphire led the round. So they're the ones that put a, you know, a term sheet in and, uh, you know, they set the terms on valuation and, you know, candidly as I think it was in the press release, we were oversubscribed as far as investor interest. So, you know, that's just a, then you're, you know, discussing with each of the investors, how much they want to invest. And we talked to some of our existing investors about taking a little less than their pro rata to make all the numbers work to get the new investors in at the levels they wanted to get in. And

it worked out beautifully. So it's really, and maybe, you know, if you were to ask Sapphire, Jay Das, who's joining our board, you know, I think they would say a couple things, you know, if you were to ask Sapphire, Jay Das, who's joining our board, you know, I think they would say a couple things. You know, one is they believe in the thesis of the company. Then you look at the TAM, you know, of the market opportunity. It's about, what did we determine, $60 billion or so, $65 billion validated from, you know, the normal sources out there of where we play for the transactional database. And so from an investor perspective, to be able to invest in a company that has a large enterprise market opportunity, has enough proof points that there's a thesis that they're going to be the winner. And then you look at the valuations that are in the market of the analytics companies like

Snowflake and others, it's a very prudent investment from their perspective, given the opportunity versus the risk. I'm speaking on behalf of Sapphire and the investors, but I think that's what they would say. Okay. Okay. And if you look at the analysts, like IDC and William Blair and all of these analysts, I mean, you can see clearly from their report that the size for a transactional relational market is actually about twice the size of the analytics SQL market, relational and analytics market. So companies like Snowflake, et cetera, are playing in a market whose TAM is like about half

the total TAM available to transactional relational databases. And I don't know, George, if you saw Jason's from William Blair's report on the database market. They did a report a few weeks ago that was talking about the database market more broadly. You guys mentioned in there. But that gives you a sense of what's happening in the scope. And then there'll be actually a spotlight coming out

specifically on UgoByte later this week as well. Okay, okay. Yeah, that will be interesting. I'll keep an eye out for that. Okay. Okay, nice. So let's move to what I think is probably leading your growth. So the cloud offering. Last time we spoke, it was still, well, I'm not sure if it was actually in your drawing board or it was already in progress, but either way, by now it's already generally available. And you're going to tell me if my assumption is true. Is it indeed the one that's probably driving your growth?

In part. And Karthik to answer that, I think just to ground George, talk about the self-managed and fully managed cloud. Cause we have two flavors that I think are really critical to driving our growth here. Yeah. So I think the last time we spoke,

maybe we'll just start with the, one of the founding thesis of the company, just like we said, we needed a transactional database that had relational features built for the cloud. Similarly, another of the founding thesis was that the world was going to go database as a service, but unlike the analytics space, the OL OLAP space the transactional space people would not be comfortable on day one putting their data into a managed service by a third

party vendor right whoever that is and obviously we were born in 2016 so this thesis is as of 2016 that over the next 10 years or so there'd be like that want to run databases as a service but in in their own account, public clouds, or in a fully managed account. So it could just vary across the two sides. So accordingly, what we did was we built our fully managed DBaaS control plane software as a package software that can ship into, you know, that we can run, or our customers can run inside their VPC and cloud accounts. So that means like so and we started this about six months into the starting the company so this has been our commercial product, the main flagship

commercial product that we've been selling to enterprises and it's become a mainstay of enterprises for them to run the database and the database tier inside their in their environment. So it's something that we can remotely manage if required, but they fully take care of it on their side and run. So they get the control plane software, which takes care of the database, and they just have to take care of the control plane

software, which is a much easier task than dealing with a lot of different things that the database needs to do. So to Bill's point, we built this in such a way that we can use the same thing for our fully managed cloud. And the first offering where customers run it by themselves, it's a self-managed cloud, like YugoByte Cloud, self-managed version. And the one that we run for them in a multi-tenant fashion

is the fully managed YugoByte Cloud. And any feature we build for one site, one group, you automatically get on the other side. So it's absolute reuse of software and features and et cetera, which is a good place for us to be. Because we think people will move back and forth between whether they want to run it as a fully managed service or self-manage it themselves. But with the ease of use of a fully managed service over the next many years still right so from that perspective all of our revenue almost all actually like maybe we have one or two support type contracts but everything else comes from self-managed cloud consumption and it is a

subscription model it is like you know pay for consumption by the number of course so on so forth so it's exactly the way you would expect a cloud offering to work where all of the support is increasingly built into software making it easy for people to manage and operate at scale so on and so forth what we just announced is the fully managed version and as you correctly like you know laid out George like it's it's like the first you know year or so we had it in beta and we were mostly learning about the market appetite and the use case and the access patterns and so on. And just like last month we released that in

GA generally, it's generally available now. So in other words, another way of saying this is we have had like our managed cloud software being used in mission critical use cases at massive scale for like you know a year plus now year two years now almost right and this latest version is the one where everything all the servers reside on the yugo byte accounts and we just expose endpoints for end users to consume so it takes the usability and ease of use to the next level but you know case in point the software and the paradigm has existed for a while. Okay, okay, thank you.

And a related question that I had again, picking up some info bits from the press release, it was mentioned that you had something like 1,200 organizations signing up for your cloud. And I was wondering that sign up equal user in that case, or do some of these organizations have test accounts? Are they all paying customers basically? No, they're not paid customers. There are a lot of test accounts, people building software developers at various stages.

Obviously, a lot of that number comes from the fully managed cloud, which we just opened as generally available, like just a few weeks ago. So these were things, these were organizations that were using our beta tier on our cloud itself for development, for testing development and even exploration

and so on and so forth for a while. And now they're starting to get to a point where they're looking for some features in order to go production. We have a handful of paying customers on the cloud side, but we have a lot more in the pipeline, given we just went generally available and there's still some feature asks from these organizations

in order to go GA. Now, some of that number comes from the self-managed version of the cloud, which is equally easy to use. And when specifically when people need, or when the users have a lot of compliance needs or are going to deploy in like, none of the big three public clouds, which is initially the only places we'll have presence,

like our fully managed cloud today operates on AWS and GCP with support for Azure coming pretty soon, right? But a number of these users want to operate in different clouds where we actually don't have support currently. And so the self-managed cloud is of interest to them as well. And maybe to tie it to the funding round, George, the numbers we've talked about are we're approaching 50 commercial paid customers. And size-wise of talked about are we're approaching 50 commercial paid customers

and size wise of the company, we're about approaching 250 employees. That gives you a sense of where we're at on the journey and the opportunity in front of us. Okay. Yeah, so the other thing I was wondering about, again, with relation to the cloud, you mentioned feature requests from some of your customers, and I was wondering what those features may be, and actually how are they prioritized in your overall queue, let's say?

Yeah, that's a good question right now, because we're right in the middle of that planning process and prioritizing. We had a roadmap, obviously, on the database itself and on the cloud offering. But I'll let Karthik maybe hit the top four or five things that we're focused on. Yeah, absolutely. I think, you know, like Bill said, great, insightful questions, really good questions. So the first set of features has to do with, you know, just catch up on what you would think are, you know, table stakes features for a cloud. But, you know, it's just a prioritization. Like, for example, we have VPC peering, but people want to use private link, right, which is a more advanced form of VPC peering. Or the ability to do multi-region cluster support without having to contact our support right like

they just want to self-service it but it's a harder problem because we don't want like unlike our self-managed offering we don't want this to be super complex but flexible we want this to be very opinionated but simple right so it's really about taking apart the user interface on how that's going to be what are the majority of the users looking for so a little bit of that type of really make it easy to self-serve kind of model that we're working with uh support for azure for example and like you know just some of those like what i would call uh basic features but it's just a long list right so we have to catch up on those uh the second big bucket uh revolves around security and enterprise features like uh they're independent but i mean bucket revolves around security and enterprise features. They're independent, but as in security features and enterprise features independently, we're just

clubbing them into one bucket. So things like we're already signed up for external audits and pen testing has been done. So we've done a bunch of those, but we're working on our SOC 2 compliance, which we're pretty close to getting here. And then PCI compliance is another request. So on the compliance bucket, there's a bunch of stuff. And then enterprises ask for features like BYOK, bring your own encryption keys in order to encrypt data. So even we can't make any sense of the data. We have transparent encryption with key rotation,

but they want to supply their own keys. So you can think of a number of these type of advanced features that enterprises want in order to make the deployment even more secure than it already is. And then the third bucket is a little bit more internal, both in terms of operational improvements, as well as to instrument the metrics on the cloud. Because of the number of users and the sheer volume,

it's very difficult for us to actually go user by user to see if we can help them if they're like there's something we can do better for them is the user experience better so we're working on inverting this to you know have the metrics instrumented where the system tells us how we can help people better and you know give them an even better experience and so on because pretty soon you know a sample may not be as meaningful as like the broad learnings across the market relics so what are the features that people are most looking for or what are the so we want them to tell us by being able to communicate with us through the application so it makes it easier for

them and for us to help them but through the app right so there's that a class of features and then last but not least there's the whole huge ask for data migration services. Our Postgres compatibility naturally lends itself to people being able to migrate from a number of databases with relative ease because the migration is very similar to migrating to Postgres. So data migration service is a huge ask. A lot of people asking about it, so on and so forth. So these are some of the simple ones, like, I mean, there's obviously internal automation and making things more, giving APIs for our clouds,

just like a whole bunch of other stuff that I'm not mentioning here, and not to mention on serverless, like to figure out how developers can consume it. So there's a number of angles here, but yeah, I just giving you a list, so. Yeah, and I would add, George, and if you pull up a level from a broader perspective,

if you think about the database landscape for the last 40 years or so, and how transactional applications have been built and leveraging traditional databases, when you think about the trends that are happening around microservices, cloud native, how do you rethink, and a distributed database like you the byte? How do you think about, you know, what features make sense in this new world? What do you build in the app? What do you put in the database here?

How do you secure across clouds and in any clusters? How do you provide the data privacy and the rest of it you need from security? So from a thought leadership perspective, we take that very seriously. We want to be known as an opinionated leader about how to build apps and build databases and across that whole spectrum that we've described.

And so that's really what's driving our roadmap and our features and being very thoughtful about it and being very collaborative with our customers and our features and being very thoughtful about it and being very collaborative with our customers and our partners on the journey. Thanks. I had the suspicion that what you would be working on would probably be the kind of operational improvements that basically refer to dividing them into those different buckets, which I guess in a way is a sign of maturity, so you're not adding anything spectacular, but just making improvements to what you already have and already works. That's right.

That's right. Okay, so final question, and I think we should be wrapping up soon. So final question. The other thing that is probably noteworthy about, to me at least, about your offering is the fact that you have a number of onboarding paths, I would call them, for developers. So SQL is the obvious one, and then you also have Cassandra query language compatibility and the GraphQL interface as well. So I was wondering if you have any insights as to how those are all working. So, you know, which one is the most popular

and whether the other two, let's say, are also worth keeping around, so to speak. Yeah, I think it's still early days, but between the three that you mentioned, the most popular is SQL. So it's just overwhelmingly popular, I think, followed by GraphQL as number two, and followed by the SQL, because center query language is number three. One of the things that we see is people come in for the SQL support, which is just terrific in our case. But once they understand the trade-offs between SQL and CQL and really the type of application

that they're building, they try to see the best of both worlds that they can get. And different microservices end up going to different sites. So CQL, the way it works for us, may not be the leading acquisition API, but it definitely is a very valuable API. So specifically when it comes to very high performance and very high geographic distributed data, CQL is able to perform really well. And one of the things we're doing strategically is connecting the two sides.

So you can use CQL as a foreign table to our YSQL offering, like Postgres foreign table in some sense. So you will be able to write data on one side and actually join it with data on the other side, right? And so that's been a very heavily asked for feature. So I think we are also tightening the different offerings to make it more coherently appear as a single thing rather than multiple different acquisition points. So I think the point is, where we stand, I think it's becoming clear that we cannot ignore

any of these paths, like they're all super critical to end users for different reasons, right? And yeah, and there's other asks too, there's asks for, can you support other things in addition to GraphQL, right? Can you support an off the top REST API service, for example, or can you like, so there are more asks, so we are, those are some of the things

that we are also thinking about and figuring out how we want to go about it. Okay, thank you. And yeah, it seems like you have a lot of work lined up ahead of you and you're now quite well funded to be able to execute on that. So congratulations again for the funding round and best of luck in your future plans. Thanks, thanks, George.

Thank you, George. I appreciate the coverage. Yeah, thank you. Thanks a lot. I hope you enjoyed the podcast. If you like my work, you can follow Link Data Orchestration on Twitter, LinkedIn, and Facebook.