Last week DigitalOcean held its first investor day since becoming a public company. Management again highlighted positive improvements which are nonetheless impressive, but not so new anymore. The company also provided further details on customers and product growth, but the overall pitch was soft and left me disappointed not to hear anything substantial regarding new product innovation or edge computing. I continue to be impressed by the business turnaround and progress since Yancey Spruill took over as Chief Executive but am underwhelmed at the slow progress around new product and market opportunities. Additionally, given the company only recently hired its first Chief Product officer, I'm eager to see if the product strategy evolves and whether or not the company will dial-up the innovation knob or become comfortable navigating the market with primarily commodity products.
- Company maintained guidance for FY22 and reiterated confidence in reaching $1B revenue target by FY24, growing at 30%+
- Negative implications from Russia/Ukraine War and slowing growth in Europe and Asia
- Will look to utilize M&A as a tool to achieve inorganic growth with the potential for 1-3 deals a year possibly doing larger deals
- Muted focus on product development and innovation as well as a lack of clarity on how the company will approach the large edge computing opportunity it highlighted during the presentation
- Managed Database revenue grew 95% y/y, Managed Kubernetes revenue grew 47% y/y, and App Platform users grew 120% y/y, and expects to grow PaaS revenue to more than 20% in a couple years from low double-digits today.
- Ramping up investment in sales-led GTM, and expects costs to increase from 3% to 10% of revenue
- Customers spending less than $50/mo (categorized as "learners") represent 80% of customer base and 15% of revenue, while customers spending between $50-500/mo (categorized as "builders" represent 10% of customers and 1/3 of revenue, and customers spending $500+/mo (categorized as "scalers") represents 50% of revenue and makes up 10% of customers.
Product Development and Innovation
Earlier this in February, I published some thoughts on the year ahead and highlighted several opportunities across product and market expansion. On product development I highlighted potential opportunities in Compute, Network and Marketplace/Other. For compute, I thought that the most likely product additions could be GPU droplets for AI/ML workloads and support for Microsoft Windows OS, while for Network, I thought a direct connect product, inter-region connectivity between VPCs, and more security features were plausible. And, in Marketplace/Other, I highlighted the opportunity to add virtual networking integrations among a smorgasbord of opportunities in software and cloud infrastructure, as well as building on its Serverless product.
Some of this was touched upon during investor day but a lot wasn't. Let's take a look at commentary around GPU's and the history around the company specifically talking about GPU's.
Product Timeline - Investor Day Presentation
In April 2021, Yancey Spruill responded to an audience question about GPU's at the NY Enterprise Technology Meetup, saying GPU’s are a high priority and actively being worked on. In subsequent events, the talk around supporting AI/ML w/GPU's continued to occur through investor day.
Yancey Spruill, Chief Executive Officer, DigitalOcean @ NY Enterprise Technology Meetup (April 2021)
"We are investing a lot of time, and working with certain customers and hardware vendors—its a top of mind focus area for us in terms of new product categories we're going to invest in, so stay tuned, but we're actively working on a solution for GPU's"
Yancey Spruill, Chief Executive Officer, DigitalOcean @ Canaccord Genuity Growth Conference (August 2021)
And then things that can leverage our database more, AI or ML tools could make sense to help customers make better decisions faster as their customers grow and scale, again, to remove the barriers and the obstacles of customer analytics to help them grow faster.
Yancey Spruill, Chief Executive Officer, DigitalOcean, DigitalOcean’s 4Q21 Earnings Call (February 2022)
Look, there are other areas that we will look to invest in, but it’s important we’ll be adding relevant and sort of targeted capabilities. Given our customer base is early in the journey, we don’t need this long, extended product list. We — but we do need some incremental capabilities, for example, and AI/ML sort of tool is one that comes to mind. So we’re looking to do that.
During investor day Q+A, an analyst from JP Morgan asked: “we’ve heard some focus on AI/ML use cases. Should we expect a GPU offering from DigitalOcean at some point?", prompting the following response from Chief Product Officer, Gabe Monroy who reiterated what we’ve known for a couple years now, noting an increased investment that would need to occur to support GPU's.
Gabriel Monroy, DigitalOcean, Chief Product Officer @ DigitalOcean Investor Day (June 2022)
It's certainly one of the things that we're looking at. GPU is definitely synonymous with AI/ML, I think, as the question points out, also synonymous with things of video streaming, which is encoding/decoding, which is another top use case that we see on the platform today.
It also is a challenging effort because there is a level of capital expenditure required to do GPU right. And so it's an area where we're constantly looking at the ROI and when we think the timing is right for something like that, but it's definitely part of our road map.
So, when are GPU’s coming and what’s the hold up? A few reasons for the delay may include:
- Gabe Monroy joined DigitalOcean in October 2021 as its first Chief Product Officer, so its possible he hit the pause button on some initiatives as he settled into the role and focused on setting priorities, building out the product organization, and updating the product vision as well as roadmap.
- Since Chief Executive Yancey Spruill joined the he company and brought on a new management team, they have focused on improving the financial profile of the business which they talk about frequently, often highlighting their accomplishments in doing so. Considering GPU infrastructure would elevate capital expenditures, it may simply be about timing or even hesitation to make the required investment as the company targets long-term Capex under 20% of revenue. DigitalOcean’s cloud infrastructure is hosted in premium data centers such as Equinix and Digital Realty and that alone comes with higher costs.
- The current macro environment throws another wrench at things with supply chain constraints, high interest rates, Russia’s war in Ukraine, slowing growth in Asia and Europe, and rising energy costs driving higher data center pricing.
- Deal hunting: There are several specialized GPU cloud platforms and its possible the company may be looking to make an acquisition after initially looking to develop the product organically.
Linode launched GPU instances in June 2019 and leverage the NVIDIA Quadro RTX 6000, while Vultr launched GPU's last month and is currently in beta using the NVIDIA A100 with the A40 and A16 coming soon. OVH also offers a GPU product, and all of the major public cloud providers have had GPU instances available as well.
Monroy, speaking in the scripted, pre-recorded video said the following:
The applications of the future will be global, hosted seamlessly across cloud and edge, with great performance from anywhere in the world. They will store large amounts of data to enable machine learning and artificial intelligence, leading to a new generation of digital products that gets smarter the more that they are used.
More usage leads to more data. More data leads to better AI insights. Better AI leads to better digital products and to more customers across the globe, and the cycle continues. We call this the developer flywheel, and it is central to our product strategy.
What Monroy said makes sense, even though it is high-level, these are the trends that DigitalOcean can capitalize on if they have the appetite to make the investments and ramp up the pace of innovation to deliver new product capabilities.
He calls these Application Accelerators, and depicts a flywheel consisting of Geo-distributed applications, Large data sets, and AI/ML insights.
Edge Computing will manifest through several types of infrastructure and platforms with the most relevant and common being what used to be standalone content delivery networks (CDN). CDN's have been in transition for a few years now, evolving into edge platforms where the actual content delivery capability is becoming a function of the platform, and to an extent representing a small percentage of revenue as security services, serverless computing, and other application (PasS) and compute/storage capabilities are developed.
Monroy said that 70% of DigitalOcean's revenue is from customers outside the US and highlighted that IDC is forecasting $274B in World Wide Edge Spending by 2025. Additionally, he goes on to say:
We plan to use inorganic investment to accelerate our edge capabilities and infrastructure. With those new capabilities, we will be poised to deliver a set of products that are deeply integrated across cloud and edge, with DigitalOcean's trademark simplicity and predictable pricing. This DigitalOcean customer uses DigitalOcean to provide an advertising analytics solution that helps companies optimize conversion.
There are too many options in the market to speculate what they could potentially buy, but hey, maybe they will acquire Fastly or StackPath and pull a reverse of Akamai acquiring Linode. There are several smaller CDN platforms out there they can pick up to help accelerate a move into edge computing and enable a more distributed footprint.
A great question on the subject came from James Fish, but with it, an uninspiring answer from Monroy.
James Fish, Senior Analyst @ Piper Sandler @ DigitalOcean Investor Day (June 2022)
Cool. Gabe, maybe have a sip of water because I got a 3-part question for you. So this is from Jim Fish at Piper Sandler. And he asks, please discuss the movement towards edge and edge spending getting to $274 billion? However, how will DigitalOcean, which is more centralized cloud, take advantage of that spend?
And what percentage of that market do you address today versus 5 years from now? Wouldn't you need to spend more on CapEx as a result to keep up with some of the other edge vendors? And finally, how do you think about Akamai and Linode competitively given the combination and announced road map that they shared recently?
Gabriel Monroy, DigitalOcean, Chief Product Officer @ DigitalOcean Investor Day (June 2022)
Well, gosh, that is a mouthful. So just first taking the Akamai/Linode question. There was recently an interview with the CEO of Akamai on, I think it was protocol, where the CEO described Akamai and Linode's combined future as being intertwined with enterprise. And it wasn't the first time they said that. There's really lots of examples of this idea of sort of Akamai and Linode and the history there, sort of taking them away from this SMB direction that is DigitalOcean's core focus and will continue to be our core focus. So I don't really see us overlapping in terms of capability, customer conflicts over time.
In terms of the broader question around Edge, I do think this is a really interesting question. 70% of our customers today are already using CDN capabilities. So we know that there is, in fact, demand for edge solutions within our customer base today. How we go about bringing those capabilities to the platform? Well, I mean anytime we have that challenge, we look at a range of build by partner options, and everything is on the table for us at this point. But I do think we need to figure out how we can move quickly in the space and look at -- is there an opportunity for us to use inorganic growth to really accelerate our path there.
Unpacking Monroy's response doesn't tell us much other than it sounds like the company has a lot of work to do around formulating a strategy and plan to capitalize on the edge computing opportunity it clearly highlighted.
Earlier this year I looked at the potential evolution of DigitalOcean as it related to market expansion, highlighting several locations in emerging markets that could be potential candidates for expansion opportunities. Based on previous management commentary, competitive landscape, and macro factors, those locations were Brazil, Sydney/Melbourne, Australia, Tokyo/Osaka, Japan, Mumbai/Delhi, India, and South Africa.
From that analysis (2022 Outlook: DigitalOcean)
DigitalOcean has talked about expanding to Brazil for almost a decade, when in 2013 its co-founder and then Chief Executive told TechCrunch that Brazil was the most requested location but the company had not yet made a decision. This is something that Yancey Spruill, current Chief Executive has also mentioned at investor events, saying:
We're looking to expand that into certain regions where we're not today, in Asia Pacific as well as South America, Africa. So we will look to globalize.
That also brings Africa and other markets in Asia Pacific into the conversation, which have also been mentioned previously at other events
Recently, in its 1Q22 earnings presentation, Sydney appeared on the map with a footnote indicating the region will open in the fourth quarter this year.
During the Investor Day Q+A, another analyst asked: "What are the opportunities in currently under-penetrated markets, like the Middle East or Africa?
Yancey Spruill, DigitalOcean, Chief Executive Officer @ DigitalOcean Investor Day (June 2022)
I think the beauty of our company is when you look at the 620,000 plus customers, they're spread around the world, are pretty evenly according to -- roughly according to GDP in those regions, Asia, U.S., Europe, Middle East, Africa, et cetera. And I think that speaks to what we enable. We enable people to come to DigitalOcean and get their ideas on the Internet, whether they're just testing them, their learners, figuring things out or they are the future builders or scalers who are actually got an idea into customer stance and they're building a business or scaling a business.
And so our strategy -- it's an interesting business because when you look at revenue, you look at customers, we're already sort of nominally global. And it's not like the typical company a year after the IPO that is 90% U.S. and they have to go build sales to build out a presence in the Rest of the World. We're already in the Rest of the World.
So how we make decisions, for example, we just announced the data center that we're going to launch later this year in Australia. We're already in Australia. We have lots of customers in Australia, but it's a high-growth tech market. We see opportunities to grow even faster potentially than that market.
We're looking at other data centers across Middle East, Africa, South America, where we have presence, we have customers, but we think we can grow faster by adding those capabilities. Jeff mentioned adding sales capabilities in the 3 regions as we look at them. Those are going to help us grow faster in some of these under-penetrated markets.
But it's an interesting question because we're already global. So it raises the bar on marginal investment because we got to be judicious about choosing one versus the other because we have to have conviction if we're going to be able to over-penetrate in the market because we're already there.
Again, we've been hearing the same thing for a while now, so I'm not sure we should expect a more aggressive expansion strategy beyond one or two new locations every few years. The last new market DigitalOcean expanded into was Bangalore, India back in 2016.
Outside of the large public cloud providers, all of DigitalOcean's competitors offer cloud infrastructure in more markets. Akamai/Linode has 11 across North America, Europe, and APAC with the company planning significant expansion in the tens of markets, having indicated at an event they could easily see Linode getting into 100 markets. Vultr is available in 25 markets across North America, South America, Europe, APAC and Australia, while OVH has a presence composed of 32 data centers across 12 markets spanning North America, Europe, APAC and Australia, while Hivelocity is available in 26 markets across North America, Europe, APAC, and Australia. Upcloud is in 12 markets across North America, Europe, and APAC, StackPath is in 35 markets with 60 edge locations, and Cox Edge is currently available in 19 markets in the US with an additional 12 markets planned.
While a lot of what DigitalOcean is doing well can be categorized as the modernization of web hosting—VPS 3.0 even, a larger, more substantial opportunity looms, but the question remains if the company has the appetite to accelerate its product roadmap and make the investments required to capitalize on current market trends or if it is comfortable as commodity infrastructure. And, to be fair, the lower level compute and storage infrastructure at larger cloud providers will also fall into the commodity category, but innovation, services and the ecosystem around them is what brings differentiation.