We’ve long believed that the future of enterprise IT lies in simplifying complexity – and that the most powerful solutions often come from founders who’ve lived the pain firsthand.
That’s why we backed cloud management platform emma early on. From our first meeting with founder Dmitry Panenkov, it’s been clear that he and his team have a unique understanding of the challenges large enterprises face as they navigate the messy realities of on-prem, hybrid, and multi-cloud infrastructure.
Since leading their Seed in 2023 and joining the $17M Series A in late 2024, we’ve seen emma evolve into one of the most promising platforms in cloud management – a space we know well, having backed category leaders like Datadog.
We sat down with Dmitry to hear how the company got started, why early missteps helped define their product and how they’re building toward a unified cloud standard in the age of AI.
Hi Dmitry, let’s start at the beginning. What problem did you see in the cloud infrastructure space that led you to start emma?
I used to work for a large American company that produces networking switches. We had customers all over the world including big players like PayPal, SAP, Leroy Merlin. These were large enterprises with traditional on-prem data centers full of hardware, and our switches interconnected them.
At the same time, those companies were starting to adopt cloud providers like AWS and Azure. Google Cloud was just starting to emerge. Some customers came to us and said, “You interconnect our data centers – can you help connect our on-premise infrastructure with the cloud, or at least help us manage data transfers between the two?”
The answer was no. Our company focused on switches and the software that orchestrates them – we didn’t touch cloud.
But that question stuck with me. If these massive enterprises were struggling with cloud connectivity and management, there had to be a bigger opportunity. I started digging in and reading analyst reports from Gartner, IDC, and others. That was my first mistake: while these analysts describe the future, they don’t say when it will become painful enough that customers are willing to buy.
Still, I believed in the opportunity. So, I quit my job and started building. That was my second mistake – never quit your job before getting confirmation from the market that there is a true product-market fit!
What went wrong?
I hired engineers, built the first prototype, and went back to the same companies I’d previously sold to. Their response? “We’d never use this – you need to rebuild everything.”
So we did. We spent another three years rebuilding from scratch. That process taught us that the real challenge was simplifying how organizations interacted with increasingly fragmented environments – cloud, hybrid, on-prem. emma was born from that complexity.
Was it just a timing issue, or did the market need more education?
Both. Even today, when we talk to enterprises, they often say, “We’re not doing multi-cloud.” But then they’ll say, “Well, we use Azure… but we were frustrated with Microsoft, so we spun up a Google Cloud environment too. And in some regions we still run on-prem data centers using VMware. But no, it’s not multi-cloud.”
Of course, it is. They’re using Azure, Google, VMware which are three different environments, managed by 16 engineers using 30 tools. That’s the definition of complexity. So, yes, part of our job is helping customers realize they’re already in a multi-cloud world… they just haven’t realized it yet.
Once that realization hits, are they quick to act? Or is it still a slow process?
Depends on the customer. Legacy enterprises take longer as you have to walk them through their infrastructure: on-prem servers, underused hardware, virtual machines, legacy apps. We help them modernize – not just by dumping everything into the cloud, but by transitioning to microservices where it makes sense.
One big caveat here is that microservices aren’t a one-size-fits-all solution. But when it works, we help them scale those workloads into the cloud using the cheapest available instances. And we make the entire process visible through a clean, unified interface.
With cloud-native companies, it’s faster because they understand the problem. But they have their own challenge: repatriation. Some are pulling workloads back from the cloud to on-prem. They love the managed services cloud providers offer, but on-prem they have to build everything from scratch. Without emma, that’s nearly impossible. We replicate the managed services experience on-prem, making repatriation as easy as migration.
What did early validation look like? When did you know emma was working?
One of our first customers went from a $2,000–3,000 monthly cloud bill to $150,000 within a year. That was our lightbulb moment. We realized that once a customer starts using emma – even for a small part of their infrastructure – they tend to expand rapidly.
But we also hit a wall. We’d hear things like, “Your product is amazing. The POC was incredible. You’ve built something more advanced than we expected… but we can’t move forward.”
That’s the worst-case scenario. No one’s saying anything is wrong, but no one’s buying. Eventually, we realized it was about credibility. Big enterprises didn’t believe we could onboard and support them. That mismatch between our stage and their size was a hard lesson.
When we finally landed a large enterprise, deployed at scale, and proved ourselves is when things started to click.
You’re working with a lot of legacy companies. How do you build trust and credibility with CIOs and technical execs?
At my previous company, salespeople had to be technical. We didn’t use pre-sales engineers. I remember interviewing with the CTO and he asked about TCP/IP, BGP sessions, deep networking stuff. That stuck with me.
At emma, we’ve built the same culture. Our sales team understands the technology. When a CIO asks how workloads move across clouds or how our networking backbone works, we don’t defer to someone else or take the homework with us. We always have the answers to their questions. That way, our future customers realize that they are in good hands and that we are a team of highly skilled professionals.
And our marketing helps too – we show up and talk at analyst events like Gartner and IDC and that kind of visibility matters in enterprise sales.
Let’s shift gears to AI. Everyone’s on the hunt for more and cheaper access to GPUs. How is emma supporting that?
First, we interconnect providers that offer GPU access via API. That lets customers build logical clusters using the same GPUs across different providers. With today’s shortages, that’s huge.
Second, we have our own limited GPU servers – an alpha program, really – available to a small number of customers. We built a full software stack for running and training models, storing outputs, and scaling across GPUs.We’ve also partnered with Nvidia (we’re part of their Inception program), and we’ve teamed up with Hewlett Packard and Supermicro which gives us access to the latest platforms in case we need them. We see more and more AI companies using our services, especially startups, who do not have enough engineers in and their companies prefer to use something simple to use. We include 10 hours of DevOps support each month to support their endeavours.
You’re at the intersection of DevOps, cloud, and enterprise IT. How do you stay ahead?
Honestly? Like many entrepreneurs would tell you, success often comes down to luck. And in our case, the biggest stroke of luck was that we were completely wrong at the start!
We began by building a physical networking backbone to interconnect service providers. That’s not something cloud management platforms typically do. But as it turned out, that architecture, and the algorithms we developed on top of it, became the only viable way to overcome the inherent limitations of cloud service providers.
Nobody knew it at the time – not even us! But it’s now the core of what sets us apart.
That backbone makes our platform uniquely connected and comprehensive. It’s why even much larger companies, including one of the biggest players in observability, have reached out to us. They told us: “We can give customers insights, but we can’t help them act on those insights. You can.”
That’s our differentiator. And we’re not standing still. We’re constantly building, adding services and connections. So even if someone wanted to compete with us today, it would be nearly impossible to catch up. We’re moving very, very fast.
Final question, what does success look like for emma over the next 3–5 years?
I think history tends to repeat itself. Around 25 or 30 years ago, the internet was just a collection of disconnected campus networks. Then someone invented a protocol to unify them and we got the internet.
Then came Netscape, which gave us the first real browser. Before that, even checking your email meant using a command line. Netscape created a layer of abstraction that made the internet usable for everyone.
Then came the next layer of abstraction: the cloud. We went from physical servers to managed services and applications. But in a way, we’ve circled back – we now have another set of disconnected “campus networks,” just with different names: AWS, Azure, Google.
Our vision is to build the next unifying standard – something like what the TCP/IP protocol did for the internet – but for cloud. Because at the end of the day, customers don’t care about infrastructure. They care about their applications, their users, and their revenue. When you watch YouTube on your TV, you never ask how many providers were involved in delivering that stream. You just want it to work.
That’s what we’re building: a layer that equalizes cloud providers and gives users the freedom to deploy and scale applications however they need – without ever having to think about what’s underneath.
If you want to learn more about what Dmitry and his team are building at emma, you can check out the website here.
