When Giants Fight, Does the Neutral Party in the Middle Win?
Deep thoughts on AI and aspirations —— ByteDance Deep Thinking Circle
In January, Anthropic did something that seemed self-serving: they blocked OpenCode. Any request with “open code” in the system prompt was rejected, effectively cutting off users who accessed their Claude subscription through OpenCode.
The move backfired spectacularly. The blockade inadvertently placed OpenCode on the same tier as Claude Code—even people who’d never used it noticed that Claude Code felt threatened enough to take action against it. Days later, OpenAI announced official support for OpenCode. Users flooded in.
According to OpenCode’s team in interviews, the product went from zero to approximately 13 million monthly active users in one year, processing around 7 trillion tokens daily, with annualized revenue approaching $60 million. As a reference point, they mentioned OpenRouter’s entire platform processes about 6 trillion tokens daily. These figures come from the company itself, so take them with a grain of salt—but the growth trajectory is undeniably striking.
In an era where everyone says “the model is the product,” every model provider has launched their own coding tool, and star products like Cursor are already out there. How does an open-source coding agent grow in the gap between Claude Code and Codex?
Winning Wasn’t About the Product—It Was About Position
The obvious explanation would be a superior product. But OpenCode co-founder Dax Raad’s answer is precisely the opposite, and brutally honest: most of their success came from getting the positioning right. He even said that if the product quality were half as good, the outcome would probably be similar.
It stings to hear, but it’s honest. What they built isn’t fundamentally different from Claude Code or Codex in terms of functionality—they’re all coding agents that help you write code. What’s truly different is the position they chose to occupy: the open-source position.
Their thesis was that while there were plenty of coding agents on the market, no one was explicitly claiming “we’re the open-source option,” and that position was extremely valuable. The reasoning was simple: historically, almost all developer tools eventually migrate to open source. Databases are open source, compilers shifted from proprietary to open source—look at any tech stack and the major components are open source.
So when they launched, they claimed support for over 70 models and providers, and built an open-source database called models.Dev to catalog information about various models and providers. These moves had one clear purpose: occupy that position and attract every user who didn’t want to be locked into a single model. Even well-funded large companies use them, citing the desire to avoid vendor lock-in to a specific model or framework.
Why the Giants Can’t Crush It
Why is this “neutral” position so valuable? You need to understand today’s model competition landscape first.
OpenCode’s core thesis is: competition between models will only intensify, with closed-source providers battling each other and open-source models catching up, all while prices face downward pressure. We’ll examine this thesis later. First, consider its corollary.
For anyone or any company using models to get work done, the biggest risk is betting on the wrong horse and getting locked in. So the market needs a neutral layer that doesn’t bind to any model and allows switching at any time. And this neutral layer is useful to every model provider. It gradually becomes industry infrastructure—no giant will proactively destroy it, because doing so doesn’t benefit them either.
Competition doesn’t hurt it; it amplifies it. Consider that blockade incident: after the block sparked user outrage, OpenCode immediately messaged OpenAI, essentially saying: tomorrow everyone will be angry at Anthropic, you can take the opposite stance and officially support OpenCode, winning PR points. OpenAI confirmed the next day and announced official support that same day.
The team distilled this playbook into one sentence: choose a temporary “villain,” unite all competitors, and jointly oppose it. The giants each have their own calculations, but as long as they’re competing with each other, they’ll compete to endorse the neutral party. This is leverage even small companies can use.
This position has an added benefit: it makes trying new models nearly costless. When a new model launches, users can immediately switch to test it within this neutral tool and switch back if unsatisfied. Providers wanting to win users with new models actually depend on this intermediary layer to reach users. The fiercer the competition, the more comfortable this position becomes. In the founder’s words, every time one model tries to beat another, they benefit.
The Hidden Premise Behind This Position
But don’t rush to treat this playbook as a universal formula. This position’s viability rests on one premise: competition must persist.
If one day the model market becomes winner-take-all with only one dominant player, or several providers join forces to close off APIs and focus only on their own applications, then users’ need for “switchability” disappears, and the neutral layer becomes useless.
This isn’t fearmongering. OpenCode’s co-founder himself discussed the risk of this path being cut off: there’s structural tension within labs—product teams want to lock users in by bundling models with products, while sales teams must open APIs to meet revenue targets. As long as this internal struggle exists, providers find it difficult to go fully closed. But “difficult” doesn’t mean “impossible.” The day some lab puts revenue targets aside and commits to locking down its ecosystem, this neutral position gets undercut.
In other words, the entire value of the neutral position bets on a variable it can’t control: multipolar competition must continue. Once this premise fails, this position depreciates rapidly.
Surviving Is One Thing, Making Money Is Another
Having traffic isn’t enough—products in the gap must monetize. OpenCode’s approach is worth dissecting.
It doesn’t make money from ads or tool subscriptions, but treats free tiers as customer acquisition cost. The founder’s take is that the free tier is the new CAC, and tokens are advertising. Users first experience the value of coding agents through free credits, then become willing to pay.
Actual revenue comes from two streams. One is inference aggregation: many users can’t manage accounts, rate limits, and integrations across various model providers, so OpenCode provides a unified entry point, aggregating multiple inference services for a fee. The team disclosed this business reached nearly $40 million annualized in about eight months, with healthy margins. The other is enterprise control panels: companies with thousands of engineers can’t have everyone configure their own API keys—they need unified management of models, permissions, and budgets.
Notice that both revenue streams grow from the same position: because you’re neutral and don’t lock in, users trust you with the entry point; because you aggregate all models, you’re qualified to charge this “access fee.” Position isn’t just a moat—it’s the basis for charging. This is far more concrete than simply “building goodwill through open-source freemium.”
Another Kind of Moat
This offers a different kind of insight.
Most people understand moats as technology, product, brand, network effects—built through accumulation and advantage. But the OpenCode case demonstrates another type: positional moat. Its value comes from occupying an ecological niche, not from being better than others.
Positional moats fail differently than technical moats. Technical moats fail when they’re surpassed; positional moats fail when the battlefield itself shifts. The former you can see and chase; the latter is hard to defend against because your value derives from others’ opposition, not your own accumulation.
So I prefer reading this as a conditional pattern, not a success formula. When choosing direction, don’t rush to compete on who has stronger technology—first ask: will this market remain multipolar long-term? If yes, the neutral position is worth seizing; if it’s destined for monopoly, the neutral position is a trap.
When looking at any project labeled “open-source, neutral, compatible with all models,” ask: what premise does its value rest on? Does that premise still hold?
The fiercer the giants fight, the more valuable the middle ground becomes. That’s only half right. The complete statement is: this ground only stays valuable as long as they keep fighting.