by Vic Singh
When I look at all the startups I’ve built or backed during my career as a founder and investor, I’ll admit they can seem a bit disconnected, covering everything from social to robotics to genomics to gaming. But as I reflect on my investments, on my own startups, and on how I want to spend my time moving forward, I’ve realized that many of those seemingly disparate dots are connected through a concept that I’m calling foundational [open] ecosystems — and the companies they spawn.
This far-reaching theme provides a wide enough aperture for me to spend time where I believe value will accrue in my areas of passion, but narrow enough to provide more focus as I invest in new companies. (To be clear, this is one of a few investment focus areas for me, but not the only one.) I hope it will be a useful framework that synthesizes a number of existing concepts — as well as a North Star for my work as an investor.
Each piece of this concept — foundational, open, ecosystem — is probably familiar to many readers, but it may be worth a quick refresher before discussing how they come together.
Foundational: A foundational element in technology is one that is essential, fundamental, or basic to the operation or functioning of a particular system or technology. It’s something that provides a basis or starting point for further development or advancement.
Open: In the context of technology, the term “open” typically refers to something that is accessible and available to anyone who wants to use it or interact with it. The concept of openness in technology is often associated with the principles of transparency, collaboration, and inclusivity.
Ecosystem: An ecosystem in technology refers to a system or platform that is made up of interconnected components that work together to achieve a common goal or purpose. An ecosystem may include hardware, software, networks, services, applications and other components that are integrated and interoperable, and that are designed to work together to provide a particular set of functions or capabilities.
Foundational [Open] Ecosystems
Put these three concepts together, and you’ve got a recipe for companies that are building a foundational layer of technology, in an open fashion, that becomes a platform for a larger ecosystem of interoperable hardware and/or software companies to be spawned.
Obviously, the tech industry is buzzing about advances in foundational AI models like OpenAI and StabilityAI. But that’s one of many areas where foundational [open] ecosystems will drive the next wave of technology innovation and large-scale adoption — a wave that I believe will last for decades.
It’s an approach that unites Eniac investments like Toucan (a protocol for on-chaining carbon), Bedrock (which is using autonomous submarines to create an ocean data platform), and Phytoform (a computational genomics company building AI that creates an ecosystem of traits).
[Open] versus closed is an old debate running through the history of technology, with the open source movement winning important victories — free and open source software is estimated to constitute 70–90% of any modern software solution. But what I’m talking about here encompasses more than open source. After all, many open source projects are meant to create standalone products, rather than thriving ecosystems. And while an open source approach can make sense for companies trying to build these ecosystems, it’s far from the only strategy.
We should not treat “open” and “closed” as binary, mutually exclusive outcomes. I intentionally bracketed [open] as a reminder that there is, instead, a spectrum of how open ecosystems can be. Openness and transparency can also be combined with smart governance — I encourage companies not to simply say that they’re open, but to be thoughtful and intentional about how they incorporate openness into their strategies.
This might all seem a bit naive in the context of today’s web, which is dominated by walled gardens owned by giant corporations. But it’s worth remembering that the original Internet was built on open protocols, that web 2.0 was meant to reimagine how users could contribute to social networking and content platforms, and that web3, for all its recent challenges, still offers the hope that users take back ownership through decentralization orchestrated by tokens.
I believe the Internet will be remade by foundational open ecosystems. These ecosystems, on and off the web, can both reimagine old industries and create new ones, in sectors like artificial intelligence, machine learning, data management and analysis, open science and scientific research, open-source software, web3 technologies and decentralized applications, open compute and open social networking amongst many more.
Building and Investing in the Foundational Layer
As investors and operators, how can we recognize companies that will successfully build these ecosystems, or build venture-scale businesses on top of them? One of the biggest challenges and opportunities — where the rubber often meets the road — comes after the creation of the base or foundational layer.
There are several concepts I believe investors and founders need to be mindful of highlighted below and framed in the following diagram.
The Value Spectrum of [Open]
One key question is how open the foundational layer should be from the outset and how open it should be in the long term to accomplish the startup’s vision. It’s perfectly fine for a startup to start as a closed layer with plans to open up over time. Quality control is key at the beginning.
A startup building at the foundational layer should be intentional on how open it wants to be and what core components it plans to expose. Open source startups expose everything at the outset, but there is a spectrum. You can start fairly closed, then expose key parts of the platform. Those parts should be designed to be exposed in a way to build a thriving ecosystem, but also to accrue value back to the foundational layer.
Many foundational platforms will decide that going fully “full stack” — owning the foundational layer as well as the application layer — is the right move. This is completely reasonable and a great way to build a lasting company, but a closed platform does not build an ecosystem. A full stack, closed platform precludes many use cases and applications, potentially limiting value creation for the company.
There another key factor when deciding how open to be and the timing of it — trust. For example, Facebook and Twitter started as fairly closed products, then over time opened up their platform to third-party applications through APIs. The ecosystems, especially the one atop Twitter, started to flourish, but the openness didn’t last: As Twitter saw value accruing outside of its ecosystem, it closed off access to 3rd party developers, and in Facebook’s case duplicated many of the product features built by those developers — leading to a major erosion of trust in the platform.
Not every ecosystem needs to follow Facebook and Twitter’s trajectory of closed to open to closed again. Some that have stayed open include Linux, Apache, TensorFlow, and the Unreal Engine. In this era, we are seeing open protocols like Farcaster and Lens building truly open social networking platforms that may someday form ecosystems bigger than today’s leading social media companies.
Governance of the Ecosystem
Once a startup has made the decision to be open — to whichever degree makes sense — ecosystem governance is paramount.
A full stack, closed, foundational startup may believe it has no issues with quality control and bad actions. But it lacks key democratized governance controls and can face backlash as a centralized entity. Truly open ecosystems can accrue much more value, but they can also get very unwieldy over time.
Governance can play a major role here. Companies can start by opening small parts of the layer that have limited capability but can be exploited. Companies can use different frameworks to govern the community. One way is for the early developers to govern how applications get built and published. As you join a community and build an application, the foundational platform can assign governance rights — similar to what a DAO does in crypto — where each developer has roles in the community for standards adherence and quality control.
Each member of the governing body can provide input into new applications being published and the features they enable. While the concept of governance may sound bureaucratic and restrictive, it’s essential to ensuring the integrity and quality of the ecosystem, especially as growth compounds. Companies that invest in governance early will benefit from a future high quality thriving ecosystem and will likely suffer less backlash.
Opportunities and Pitfalls of Reference Applications
Most foundational open ecosystems can and should develop a reference application to showcase the power and horizontality of the foundational layer of innovation. There’s a needle to thread: This reference application should be broad enough to showcase the power and possibilities of building atop the foundational layer, but also narrow enough to provide obvious and immediate value.
Building a reference application that’s too narrow may result in the foundational layer accruing its value at the reference app layer. The company can get stuck with a small opportunity and never have the opportunity to become an ecosystem. They may also end up competing with highly verticalized applications to win market share.
At my first startup Nearverse, we were building a foundational layer for p2p mesh networking. After getting the core foundational technology working, we wanted to demonstrate its utility, so we built a reference app for media sharing to showcase the power. Unfortunately, we became defined by that product. The time required to build a great user experience and the marketing behind the reference app trapped us in a narrow place.
Once a startup has made the intentional decision to be [open] at the foundational layer, they will likely build tools to help spawn an ecosystem of use cases and applications. The approach to developer tooling should be made carefully. Much like the reference application, building tools for developers can be a huge focus area in and of itself, and and in fact can define the company. The good news is that you will likely accrue more value at the tooling layer than the reference application layer, as tooling is required to spawn a true ecosystem that compounds massive value. Think of tooling as a lower risk but higher reward way of showcasing the power of the underlying foundation — it creates a world of possibilities beyond a reference application.
Building the Community
Once a foundational platform has made the intentional decision to be [open], whether by exposing a part of the layer or all of it, it must focus on building the community. There are a host of resources for building thriving developer and user communities. The key, at the outset, is doing the unscalable things. Publishing content that shows the range of use cases possible on the foundational layer is a key part to building community. Engaging with the community in a non scalable way is an especially valuable ingredient.
When I built my last startup Kanvas, we transitioned from a full stack application company to a developer platform. One of the things we did was create independent Slack channels for the first 100 developers we onboarded. We would engage with them on feature requests and bugs, and we worked closely with them as partners. Developers would interact with one another and they began to cross-pollinate knowledge.
Do these unscalable things at the outset, before building the flywheel of community development and before the community hits critical mass.
Some thoughts on optimizing existing ecosystems
After this was first published, our friend Nima at Polychain provided these additional thoughts on stress testing an open ecosystem, as well as the notion of dynamic ecosystems:
- Another vector to assess the health and security of a platform is by the amount of time in which it has been under use and stress, and therefore how Lindy it ultimately is. By having an ecosystem carefully tap into the open source model and incorporating the model into its incentive structure, it could leverage Linus’ law and effectively scale the time spent on said platform towards anti-fragility.
- It’s well accepted that change is the only constant. A layer can be foundational, open and can aspire to become an ecosystem at one point in time but it’s critical for periodic checks and possible realignment to safely be accounted for. So it begs the question of how we can design systems that are adaptable over time, remain open and attractive to increasing value creation, and modeling what the relationship is between its variable endogenous and exogenous factors to more proactively account for creeping systemic risks. For example, what tools and protocols exist to account for social censorship-resistance? If open governance systems rely on some communication medium to effectively operate, how do we ensure this base capability exists for ecosystems to facilitate coordination and progress?
- I think the need for governance falls into this topic but it’s only part of the picture as is pointed out above by its reliance on some infrastructure. So what else could there be a need for in this bucket that’s not governance?
One critical piece that’s missing from this framework is value accrual and monetization at the foundational layer. There are many business models — deployment fees, usage fees, subscription services, cloud-based fees, protocol fees, revenue sharing, others we haven’t thought of — and they will depend on the type of ecosystem being built. I’ll do my best to unpack how to think about monetization at the foundational layer in a subsequent post.
Building and Investing in Companies Spawned by the Ecosystem
This next section tries to cover some core topics every founder should consider when building atop a foundational [open] ecosystem in order to accrue the most value and build a strong defensible moat.
Here is a high level diagram of how I think about compounding value at the ecosystem and layer built atop foundational [open] ecosystems:
[Open] Access and Commoditization Pitfalls
Depending on the spectrum of [open], developers will have access to a wide-ranging toolset to build applications. Developers by their nature are very creative and will exploit all the tooling offered by the foundational layer. The rub is that this tooling is available to everyone, so what anything one developer builds can easily be replicated by other developers. This is the obvious and crucial question investors ask: How hard is it to build this? And what is truly defensible when everyone has access to the open ecosystem?
Novel Use Cases
One approach that could be compounded with product completeness (below) is to imagine a product that simply could not have existed unless the foundational layer was available to build on. What can I build now that unlocks something truly novel? The mental model here is: What is the best complete product that I could build if I didn’t have access to the foundational layer? If I had to start from scratch? Think of the foundational layer as fast tracking that use case but with product completeness.
Innovating at the Foundational Layer
Another approach to product differentiation is innovating at the foundational layer itself. Depending on how open that layer is, developers should be able to innovate at the foundational layer — fork, build and adapt it — to create lasting defensibility on their own use cases.
One important approach to avoid commoditization without having to go deep into the foundational layer is by simply offering a complete product with classic moats built in — including lock-in, network effects, standards, data, and so forth. I published a post about moats in 2019. Try to utilize the readily available tools of the foundational layer to build what I call a “complete” product. One that provides meaningful value, is sticky and hard to rip out, and provides a complete solution to a user or customer problem. If you have a novel use case and can innovate at the foundational layer, and you have a complete product, you are golden!
Examples of Investments I’ve Made at the Foundational or Ecosystem Layer
- AI — a deep learning platform that sits somewhere between a foundational open ecosystem and the tooling layer enabling custom Large Language Models for enterprise
- Climate — a carbon infrastructure foundational ecosystem exposing its infrastructure layer for carbon-based applications to be built atop
- Community — an open platform where community meets capital built where anyone can find capital and ownership for their projects
- Data — a horizontal geospatial data platform and soon-to-be ecosystem processing synthetic aperture radar data for ground motion
- Gaming — a protocol that provides verifiable randomness to chance-based games built atop their foundational layer
- Genomics — a computational genomics company producing a portfolio of traits using an open source variant of CRISPR
- Infrastructure — an open platform for end to end containerization with a reference application for secure data sharing to showcase the power of the foundational layer
- Labor — an exchange for custom software with autonomous code validation powered by an extensible user-owned foundational layer
- Robotics — a platform that enables robotics data processing exhausting out of ROS, an open robotics platform for developers to bring flexible reprogrammability
I can’t wait to explore and invest in the next foundational [open] ecosystems and the companies they spawn. I believe this will be a decades-long opportunity: More ecosystems will be formed around [open] AI, bio, compute, data, energy, finance, genomics, healthcare, identity, gaming, infra, IoT, media, music, privacy, robotics, science, social, storage, and more.
An additional observation from Nima: “These areas do have overlaps, as they all create vast interoperable networks. For example, a thriving music ecosystem could be complementary to systems and products built in a financial ecosystem to tap into new mechanics unlocked by the primitives each ecosystem offers. The same could be true for other ecosystems mentioned, bringing about a Cambrian explosion of new behaviors and positive-sum feedback loops.”
If these concepts resonate with you and you are a founder or operator building a foundational [open] ecosystem or a company being built atop, please reach out! I’d love to jam and workshop with you on how to navigate this framework: I’m @vicsingh on Twitter and email@example.com
Special thank you to Anthony Ha for helping me tame this beast of a post, and to the founders, investors, and operators who were generous enough with their time to provide feedback: Greg Bildson, Aleksandr Bulkin, Karthik Dinakar, Benjamin Gibbs, Jonny Mack, Garrett Maring, William Pelton, Nima Vaziri, and Arun Venkatadri.