Build a Great Business on Open Source without Selling Your Soul

Recorded: March 5 @ 08:00 am PT
Presenters: Robert Hodges

In this 25-minute FOSDEM 2025 Community Devroom talk, Robert Hodges draws on nearly two decades of open-source business experience to answer a question that trips up even talented developers: how do you turn a thriving open-source project into a profitable, sustainable company without abandoning the community that made it great?

Using three real-world case studies, Percona (pure services), DBeaver (open core), and Altinity (managed cloud), Robert walks through the business fundamentals of gross margin, the three commercial models that actually work, and what to genuinely stress about: marketing, sales, and financial management, not licensing. He closes with a frank discussion of VC funding traps, the temptation to re-license, and how foundations and alternative funding sources can help you avoid making choices you will regret.

Here are the slides:

Key Moments (Timestamps)

Key moments generated with AI assistance.

• 00:41 – Introduction and shout-outs to Percona’s Peter Zaitsev and DBeaver’s Tatiana Kropenia
• 02:02 – The core question: how to build a business that makes money and gives back
• 02:21 – Setting the stage: your successful open-source project
• 03:21 – Why big, well-funded companies failed to profit from open source (HashiCorp et al.)
• 04:51 – Basic business theory: the P&L, gross margin, and why it matters
• 07:28 – Model 1: Pure services on open source: education, consulting, enterprise support
• 09:00 – Case study: Percona: 24×7 support, managed services, builds, Percona XtraBackup
• 12:13 – Model 2: Open core: community edition plus proprietary extensions
• 14:01 – Case study: DBeaver: 10 million users, enterprise edition, giving back through community
• 16:30 – Tailscale as another open core example at scale
• 17:01 – Model 3: Managed cloud services on open source
• 18:52 – Case study: Altinity: Altinity.Cloud, Kubernetes Operator, contributions to ClickHouse®
• 20:21 – Licensing: don’t stress too much; pick one that supports the four freedoms and stick with it
• 21:34 – What you should stress about: marketing, sales, financial management
• 23:15 – VC funding traps: the 100:1 return expectation and why it leads to re-licensing
• 25:16 – Alternatives: put software in a foundation; customer-funded growth; revenue-based financing
• 26:23 – Closing: building businesses is a way to protect open source, pay it forward

Talk Transcript

[00:41] – Introduction

[Introduced by the session chair as “Robert Hodges, CEO of Altinity, who will speak on building a great business on open source without selling your soul.”]

Robert: Can you guys hear me okay? Great. It’s wonderful to be here. I want to start with a huge shout-out to FOSDEM. This is just an absolutely awesome conference. I was here for the first time last year, and it’s totally energizing. If you’re into open source, this is the place to be.

I also want to start with a shout-out to two people without whom this talk would not have been possible. That’s Tatiana Kropenia, she runs DBeaver, one of the companies I’ll be talking about. And Peter Zaitsev, who is right here, a friend and someone I look up to as an icon of how to build a successful open-source business. We’ll also be talking about his company.

What we’re going to be talking about is how you build open-source businesses that also give back to their communities. I’m going to give you three examples and then talk about some general business issues. These are not the only ways to build successful businesses, but they’re three that worked. Bear that in mind as we jump in.

[02:21] – Your Project Is Great. Now What?

Robert: Let’s talk about you. You have a wonderful open-source project out on GitHub. You’ve got hundreds of contributors. You have thousands of stars, and the stars are growing by leaps and bounds every day. You have community members who love you.

Moreover, you have a nice Apache 2.0 license. Corporate lawyers don’t feel love the same way we do, but when they see an Apache 2.0 license, their hearts swell with joy. So what you have here is a project that is successful and poised to begin maybe building a business. And you start to think: I can do this for a living. It’s not just a side project, it’s not just something fun.

So the first thing you might want to do is look at other people who had successful projects and turned them into businesses. Let’s look at a few.

Here are some businesses that are probably familiar to all of you: companies you may well know, founded by very smart people, backed by a lot of money. What they wanted to do was build businesses on products that were basically open source. So how did it go?

Not well, actually. I picked those companies because, to the best of my knowledge, not one of them has ever been profitable. And along the way, in their search for profit, they all re-licensed. I’m not making a moral judgment here. As we’ll see in this talk, part of the reason they ended up re-licensing is that they were facing some bad math.

Building a successful project and turning it into a successful business is a very hard step. So if you’re going to do it, you have to take a step back and think about how businesses are built. What I’m going to do for the rest of this talk is answer two questions. First: how do we build an open-source business that actually makes money, so it’s stable, it persists, and you make a decent living from it? Second: how do we make sure it gives back to the community that helped build the project in the first place?

[04:51] – Business Theory: The P&L and Gross Margin

Robert: To begin with, I’m going to talk about some business theory. It’s really not that bad. This is about all you need to know to start an open-source business.

Here’s your project. You’re going to offer some kind of product or service related to it, maybe a combination of both, and you’ll collect money. That’s called revenue.

Picture a simple profit-and-loss statement. Money comes in, say 10 euros. Whatever good or service you’re selling, you have to pay to deliver it. If it’s support and it costs 2 euros to deliver, that’s gone. Next, you have operating expenses: the overhead of running a business, including marketing, sales, development, and administrative people like me. Then there’s a special category for things like taxes, interest, and stock-based compensation. This is actually the real reason why a company like MongoDB is not truly profitable: they’re giving away a lot of stock, and they don’t want to count that when they’re reporting how much money they make.

Take all that out, and whatever’s left is profit. That’s real money you can stick in the bank, either spend it or save it.

The important concept here is gross margin. This is the percent of your revenue left over after you pay for the service you sold. It’s a way of judging how efficiently you can deliver a product or service, and we’ll use it to compare different business models.

[07:28] – Model 1: Pure Services on Open Source

Robert: Model number one: take a pure open-source project and offer services on top of it. The software is completely open source, no proprietary extensions, nothing. What kinds of services can you offer? Education, because open-source projects don’t always have the best documentation. Feature development, because people want things added to the project. Consulting on how to set it up and use it in real applications. And enterprise support, which is the most interesting for many purposes.

Enterprise support is where you sell a service contract. If something breaks, you help fix it. If they’re deploying a new application, you tell them how. If they’re upgrading, and so forth. Why is this good? Very high gross margins. In our business, we look for about 80% or more. That means 10 euros come in, 2 euros to deliver the service, and 8 euros stay with the company. That’s really good: it means you can pay for the other expenses and build a company that actually makes money.

Enterprise support is basically insurance. For the most part, it doesn’t cost much to deliver. But when something really breaks, say you’re running an e-commerce site on MySQL, or better still a stock trading system on an open-source database, and that database goes down, the market is down. That’s when everybody who’s giving support wakes up, helps you get it back online, and everybody’s happy. That’s why the customer finds it a good deal.

Case study: Percona. One of the best examples of a company that does this really well is Percona. The main Percona services are 24×7 enterprise support, managed services where they actually log in and help you run things, consulting on specific projects, and a database monitoring product. They started doing MySQL: Peter and Vadim, who founded it, are performance experts in MySQL. They’ve since added support for PostgreSQL, MongoDB, and other databases. All of it is completely open source.

How does Percona give back? For years the Percona blog was daily reading for every MySQL user. They also maintain builds, a fork of MySQL with production-system fixes that they’ve maintained for many years. A really important product they built is Percona XtraBackup, a backup and restore tool for MySQL databases. This is a boring, difficult problem that everybody needs solved. By solving it and providing it as open-source software, Percona delivered a major benefit to the MySQL community. They also run conferences and have done many other things.

That’s model number one. Red Hat, by the way, is the most prominent example of this model at scale: they do builds they then sell you, but it’s fundamentally a services model on open-source software.

[12:13] – Model 2: Open Core

Robert: The next model is what we call open core. Some of you have heard of this and probably have mixed feelings about it. Here’s how it works when done right.

You have an open-source product that’s fully usable: maybe it doesn’t do everything, but it’s something you can actually run on production systems. Then you have proprietary extensions in a separate build. Those extensions are things that are useful to companies that want more than basic features. For example: single sign-on, because username and password are fine for one developer, but a bank with hundreds of systems needs SSO integration with Okta or Auth0. Enterprise integrations with proprietary systems. Performance monitoring and analytics. High availability and disaster recovery features.

If you do this well, it has very high gross margins, typically 80% or higher.

Case study: DBeaver. One of the best examples I can think of is DBeaver. If you haven’t tried it, go do so. It’s a visual database tool that lets you connect to any database. With the community version you can edit data, submit queries, do simple automation, and edit your schema. It works great with ClickHouse, which is the database I use.

Then they created DBeaver Pro, the enterprise closed-source build. It’s reasonably priced. They’ve been incredibly creative at thinking of things useful for large businesses: single sign-on, Oracle support, NoSQL databases, data design tools, and task scheduling. A bunch of features that enterprises running many databases find useful. This has enabled them to build a profitable business. Tatiana mentioned the last time I talked to her that they had something like 10 million people using their software, of whom roughly half are enterprise users.

How does DBeaver give back? The biggest thing is simply offering the community edition. That’s a real gift. The project has 41,000 GitHub stars and hundreds of contributors. One particularly interesting thing: because DBeaver is open source, if you’re building your own open-source database, you can submit a PR with a DBeaver connector, get it merged, and DBeaver will support your database. They’re giving back to other open-source databases. That’s a really elegant form of community contribution.

Another example at scale is Tailscale. Tailscale does something similar: the free version is an MIT-licensed open-source project that lets you access services from anywhere. There’s an enterprise version that is very popular. There are some very large companies using this model.

[17:01] – Model 3: Managed Cloud Services

Robert: The third model, and the one my company uses, is to build cloud services. You have some open-source software that runs in the cloud that people want available as a service they can just connect to. You offer support and cloud management: upgrades, performance, security, and so on. That management layer is proprietary; the thing it’s managing is open source.

Gross margins on this model can look like 50% when you’re running it in your own account, because you also have to pay the cloud provider. But if you run it in the user’s account, a popular bring-your-own-cloud model, the gross margins are very good. It’s also very sticky: when you’re running something for somebody, they’re not going to leave unless you screw up or they hit financial problems, and they have a tendency to use more of it over time.

These managed services are also good vehicles for very high-quality, proactive support. Because you’re operating the system, and if you have deep expertise in both the software and the applications, you can detect problems and help solve them long before users ever notice them.

Case study: Altinity. We have a product called Altinity.Cloud that manages ClickHouse across four public clouds, including Hetzner and the three usual suspects. How it works: we stand up Kubernetes whenever people want managed databases, and then the management software pushes open-source components into that Kubernetes cluster and takes care of remote management. It can run in our accounts or in the user’s accounts.

One interesting thing: because the whole stack is fully open source, customers can disconnect whenever they want. They usually don’t, if something’s working you don’t want to mess with it, but that ability to leave is a big value for customers.

How does Altinity give back? We contribute PRs to upstream ClickHouse: we’ve done hundreds of them. We built the Altinity Kubernetes Operator for ClickHouse, the most popular way to run ClickHouse on Kubernetes, and probably the most successful open-source software we’ve developed. We also produce Altinity Stable Builds for ClickHouse, hardened long-term-support builds, as well as Altinity Backup for ClickHouse and many other ecosystem tools.

[20:21] – Licensing: Don’t Stress About It

Robert: Everybody stresses about open-source licenses, especially when you’re building a business. You think: if I don’t pick the right license, I’m going to get ripped off. How much should you stress? As a practical matter, in my experience, not much. It really doesn’t make that much difference in your ability to make money.

My advice: pick a license that supports the four freedoms. That’s why we’re into open source: we believe in sharing the software, granting rights to use it, understand it, modify it, distribute it. Pick one that supports those, and stick with it. Don’t change it. Don’t mess with people.

Yes, Apache 2.0 is more popular with corporate lawyers. GPL v3 has other advantages, even if many corporations don’t like it. But in the end, these differences don’t matter that much. There are bigger things to worry about.

[21:34] – What You Should Stress About

Robert: These are not related to code, which is why they’re hard, because we’re all geeks.

Marketing is the first one. When you’re starting out, you don’t have to stress too much: giving back to the community is already getting your story out, showing you can be trusted and produce good software. But as your company grows, it gets harder. We work in very competitive markets. It’s expensive to tell people you have a product, expensive to find the people interested in it, and it takes special skills. You’ll generally need to hire people who have those skills.

Sales is the second. Asking for money is a real art: knowing how much to ask, knowing when to ask. These are great skills. Unless you’re very good at doing it yourself, you’ll need to hire people who are.

Financial management is the third. Managing money is critical. We do everything on spreadsheets, but we’re very careful. Our company has been blessed by two people who are very gifted at this. You have to be interested in things like cash flow: how much money do I have in the bank, what are my revenue projections? Otherwise, you’ll end up laying people off right when things are looking up, or conversely spending too much money because you thought you had more.

[23:15] – VC Funding Traps

Robert: Now let’s talk about the VC funding trap, which is where a lot of the re-licensing comes from.

A friend of mine who runs a business intelligence startup said his investors told him they expected a hundred times their investment back. That’s typical for venture capital. The way they make their money is that they bet on a bunch of companies, most of which fail. The few that succeed have to generate those enormous returns to cover everything else. It’s the math of the model.

Now, if you have an open-source project and you take VC funding, you’re going to be on the hook for those returns too. And there are ways you could conceivably achieve them: get very big very fast, get acquired for a huge sum, or get profitable quickly. But the way some companies have chosen is to re-license, essentially turning open-source software into proprietary software. That’s because if you could charge proprietary software prices, you’d have much higher revenues per customer, and maybe you could get to that 100x return.

The problem is it doesn’t work very well. Users and community members feel betrayed. Forks get created. And you’ve traded away the community good will that was one of the most valuable things you had.

[25:16] – Alternatives to VC Funding

Robert: So what can you do instead?

Put the software in a foundation. If your software is really taking off, this is a way to avoid that choice entirely, just remove it from the table. Putting the code in a foundation ensures the project remains open regardless of what happens to the company. There are some great foundations: the Apache Software Foundation, the Linux Foundation, and the CNCF are three widely known examples.

Use other funding sources. VCs are not the only way to get money. The all-time best way: get your customers to pay for it. Then you know you’re building something they really want. There’s also creative financing. We’ve used revenue-based financing, where funding is based on the amount of revenue coming in. Or you can skip early-stage and go straight to late-stage funding, which comes with better terms and less pressure.

[26:23] – Closing: Build Businesses to Protect Open Source

Robert: If you’re going to build a business, look at companies that have done this, then dive in.

One thing I want to say: if you believe in open source and you believe the foundations of computing should be open-source services available to everyone, building businesses is a good way to pitch in and help. Profitable companies are one of the best protections for open-source projects and the communities that depend on them. So please go ahead and try it.

If you want advice, these are ways to find me. All of us did this through the help of people who were paying it forward. We’ll do the same for you. If you want to do this, come talk to us. I’d be happy to help. Thank you.

FAQ Section

Q: What are the three commercial models that work for open-source businesses?

Robert identifies three proven models. Pure services means selling education, consulting, enterprise support, and managed services on top of fully open-source software, with gross margins of around 80% because support is essentially insurance. Percona is the best example. Open core means maintaining a fully usable open-source community edition alongside a proprietary enterprise edition with additional features such as SSO and enterprise integrations, also with very high gross margins. DBeaver is the best example. Managed cloud services means operating open-source software as a cloud service, with the management layer proprietary and the underlying software open source. Altinity with Altinity.Cloud for ClickHouse is the primary example.

Q: Why did large, well-funded open-source companies like HashiCorp fail to become profitable?

Robert argues they faced bad math: to achieve the 100:1 returns that VC investors need, an open-source company would have to charge proprietary-software prices, which is fundamentally at odds with the open-source model. The pressure from investors to achieve those returns caused several companies to re-license or restrict their software in ways that damaged community trust. The root problem is structural: VC economics require enormous exits that open-source business models rarely produce.

Q: How much should you stress about choosing an open-source license?

Not much, according to Robert. The choice of license has very little practical effect on your ability to make money. His advice is to pick a license that supports the four freedoms (use, study, modify, distribute), such as Apache 2.0 or GPL v3, stick with it, and never change it. What matters far more to the success of the business is marketing, sales, and financial management.

Q: What is the most important thing Altinity does to give back to the open-source community?

Altinity contributes on multiple fronts. The most significant project is the Altinity Kubernetes Operator for ClickHouse, the most popular way to run ClickHouse on Kubernetes and probably Altinity’s most successful open-source contribution. Altinity also contributes hundreds of PRs to upstream ClickHouse, produces Altinity Stable Builds with long-term support, and maintains Altinity Backup for ClickHouse and other ecosystem tools. See the Altinity open-source software overview for more detail on these contributions.

Q: What are the alternatives to VC funding for open-source businesses?

Robert recommends three approaches. The first is putting the software in a foundation such as the Apache Software Foundation, the Linux Foundation, or the CNCF, which removes the question of re-licensing from the table entirely by ensuring the project remains open regardless of what happens to the company. The second is customer-funded growth: letting customers pay for development, which confirms you’re building something people genuinely want. The third is creative financing such as revenue-based financing, where repayment is tied to incoming revenue rather than fixed schedules, or going directly to late-stage funding with better terms and less pressure than early-stage VC.

© Altinity, Inc. All rights reserved. Altinity®, Altinity.Cloud®, and Altinity Stable® are registered trademarks of Altinity, Inc. ClickHouse® is a registered trademark of ClickHouse, Inc.; Altinity is not affiliated with or associated with ClickHouse, Inc.