Some markets depend on shared infrastructure, but no participant can own it without the others refusing to use it. Cloud providers all needed the same way to run containers across a fleet of machines. Banks needed a payment network that reached past their own customers. Manufacturers needed a safety mark that buyers would believe, which ruled out any mark a manufacturer owned.
The reason is ordinary commercial self-defense. Build your product on a competitor’s platform and that competitor sets the rules you operate under, watches what you’re doing while you do it, and can change the terms whenever changing them pays. So you refuse, even though you can see that the shared layer would grow the whole market and your own business with it. Everyone else in the market does the same arithmetic and refuses too.
Either the common ground never gets built and everybody keeps paying for the fragmentation, or an interested party builds it and the others quietly route around it. The second happens more often, and it costs more. One company controls the shared system, so everyone else spends engineering time building workarounds to reduce their dependence on it. Imagine one cloud provider owns the software everyone uses to manage containers. Its rivals may support that software, but they will also build adapters, alternatives, and backup systems so the owner cannot dictate how their customers operate.
A better product cannot solve this trust problem. More features, lower prices, or stronger performance do not change who controls the system or what that owner might do later. Participants will choose the organization whose rules and governance prevent it from favoring itself, even if another product is technically better.
In this kind of market, giving up control is what makes the shared system valuable.
That cuts against most of what B2B strategy has been telling operators for the past several years, which is to own the platform, hold the data, and convert a partner network into a captive audience. I’ve argued here before that the classic fortresses are eroding: switching costs, proprietary technology, and data advantages don’t hold the way they used to. Neutrality is the strange exception. It’s a moat made entirely of powers you give away, and it holds because you can’t take them back.
Why the better product loses this one
You can build the best shared platform and still fail to become the standard. Choosing infrastructure is a long-term commitment. Before a large customer depends on your platform, they will ask what happens if your interests stop aligning with theirs. If you can change the rules in your own favor, they have a reason not to trust you, no matter how good your software is. They may choose a weaker platform governed by an organization that cannot favor itself.
Google worked this out in public. On 21 July 2015, the day Kubernetes reached 1.0, Google contributed the project as the seed of the new Cloud Native Computing Foundation under the Linux Foundation, with a technical oversight committee and founding members that included Cisco, Goldman Sachs, IBM, Red Hat and VMware.
While Kubernetes belonged to Google, adopting it meant taking a dependency on Google, and every one of Google’s competitors could run that calculation. Under the foundation, companies that would never have standardized on a Google product standardized on the foundation’s. CNCF’s 2025 Annual Cloud Native Survey, published in January 2026, found that 82 percent of container users were running Kubernetes in production.
Google gave up the ability to steer the project alone. What it got back was an industry standard it helped design, running on hardware it sells, staffed by engineers it had trained.
Trust comes from giving up control
Most companies promise to be neutral but stop short of giving up the control needed to prove it. They write the blog post about their commitment to openness, publish a set of principles, and expect the market to treat that as neutrality. It doesn’t, and the reason is easy to see from the outside: the promise costs nothing to make, and the incentive to break it sits exactly where it always sat, with you. Everyone you’re asking to trust the promise can also see the quarter in which breaking it would be worth more than keeping it.
What counts is a structure that makes your own betrayal expensive or impossible. Move the trademarks and the governance to a body you don’t control. Give the participants real votes, including votes that can go against you. Publish guarantees of openness that actually bind you. Every one of those is costly, and the cost is the entire point, because a commitment that would hurt to break is the only kind anybody believes.
Bank of America learned this the hard way with BankAmericard. It had licensed the card to other banks since 1966, and every licensee was a competitor of the owner, which meant every bank was building its card business on rules controlled by a competitor. In 1970, Bank of America gave up control. The issuing banks formed National BankAmericard Inc., a Delaware non-stock corporation the members owned jointly, with Dee Hock running it, and it took the name Visa in 1977. Once no member could tilt the rules against another, the network was finally usable by all of them. Visa reported net revenue of $35.9 billion for the fiscal year ended 30 September 2024.
Bank of America traded ownership of a licensing program for membership in the network the whole industry would end up running on. That trade looks obvious now. It did not look obvious in 1970 to the executives who had to sign it.
The ground sits in very few hands
Safeguards such as independent governance, shared voting rights, and separate ownership of trademarks may seem excessive. But shared infrastructure often depends on a small group of maintainers who decide what gets built, accepted, and released. If one company employs those maintainers or can overrule them, that company still has control, no matter what the governance documents say. The structure has to be strong enough to prevent a handful of people from becoming a back door to control.
Nadia Eghbal’s Working in Public (2020) treats open-source code as public infrastructure and cites a study of GitHub projects finding that in more than 85 percent of them, under 5 percent of the developers accounted for 95 percent of the code and the conversation around it. Her argument is that the scarce resource keeping shared infrastructure alive is the attention of a small set of maintainers.
Buyers care about who controls the technology they depend on. If a few maintainers keep an important project running, buyers will want to know who employs them and who has the power to overrule them. They will also ask basic questions: Who owns the trademark? What happens to the project if its sponsor is acquired? Clear, verifiable answers to questions like these are what make a project safe to adopt and a company worthy of trust.
I’ve spent time on this at the Linux Foundation and in commercial open source. The technical case is usually easy to make: neutral governance can bring in more contributors, partners, and customers. The hard part is deciding what the company must give up to make that neutrality believable. Will it transfer the trademark? Give competitors a vote? Let independent maintainers reject a change that would help its commercial product? Promise that it cannot change the license later? Each concession closes off options the company may want in the future. That is why the hardest conversations are usually internal, with executives being asked to surrender control before the market will trust them with it.
Where the money comes from
Neutrality is valuable because no participant gets special treatment. But that rules out the easiest ways to make money, which usually involve giving your own product an advantage. The workable model is to keep the shared system neutral and sell services around it.
Underwriters Laboratories (UL) has held that separation from the beginning. William Henry Merrill founded the Underwriters’ Electrical Bureau in 1894, paid by insurers who needed to know whether wiring would burn a building down, and the organization has never made the products it tests.
The standards live in a nonprofit, UL Standards & Engagement, which has published nearly 1,700 standards and guidance documents. The commercial arm, UL Solutions, went public on 16 April 2024 at $28 a share, and the nonprofit sold 19.4 percent of it for $1.03 billion net and kept the rest. The commercial company makes money by testing products, while the nonprofit controls the standards manufacturers must meet.
Software companies can use the same model. Keep the shared project neutral, then make money by selling support, certification, hosting, or tools around it. The model breaks when the company uses its control of the project to favor its own products. That might mean putting a feature in the paid product before adding it to the open-source project, or requiring customers to use the company’s tools to earn a certification. Each decision may look small, but together they show that the company can change the rules for its own benefit. Once partners and customers see that, they start looking for a project that no single vendor can control.
What it costs to hold
Choosing neutrality means turning down some easy revenue. When sales are below target, someone will suggest a small change that favors your product and seems unlikely to draw attention.
That pressure never goes away. Every quarter, someone will propose a change that helps the commercial business but weakens the neutral project. Someone else has to say no. The lost revenue is easy to measure; the trust that decision protects is not.
Trust also takes years to build and can disappear after one self-serving decision. If the company takes control of the governance or gives its own product special treatment, participants will see that the promise of neutrality no longer holds. They may leave much faster than they joined.
Neutrality only pays off when competitors need a shared standard, network, or meeting place. If your customers do not need to work with one another, giving up control adds cost without creating much value. So neutrality is first a decision about how the market needs to work, not about how to promote the company. It belongs in the same conversation as whether you’re building the market or inheriting someone else’s.
Give up control to gain the market
Neutrality becomes credible when a company gives up the power to change the rules on its own. Competitors need to know that a new CEO, board vote, or acquisition cannot turn shared infrastructure into one company’s advantage.
Kubernetes, Visa, and UL show what that choice can produce. Their governance gave rivals a shared system they could use without depending on a competitor’s goodwill. Wider participation created more value than any founding organization could have created through sole control.
When a market needs common infrastructure, participants choose the organization whose rules protect them from the owner as well as from one another. Giving up unilateral control is the price of becoming the standard.