Harry Stebbings is betting a trillion-dollar American open-source AI company will exist

At least one American company, valued in the hundreds of billions or possibly at a trillion dollars, will be built around American-first open-source artificial intelligence. That is the call Harry Stebbings, the venture capitalist and founder of 20VC, has put on record. It is a specific, falsifiable bet, and its implications cut across competitive strategy, regulation, and the geopolitics of AI development.

The valuation threshold Stebbings names deserves to be taken seriously on its own terms. The largest proprietary AI companies have demonstrated that frontier model development can sustain enormous enterprise value. Stebbings is arguing that open-source-first companies can reach a comparable ceiling, and that the American national framing of such a company is a strategic feature rather than incidental branding. Those are two distinct claims bundled into one forecast, and both require scrutiny.

The national framing matters because it ties a commercial prediction to a geopolitical premise. Stebbings is not describing an open-source company in the generic sense. He is describing one oriented specifically around American origin and American interest. That framing implies a world in which the provenance of an AI model carries commercial and regulatory weight, in which buyers, partners, and governments treat “American-first” as a meaningful differentiator rather than a marketing phrase. Whether that premise holds depends on policy environments, procurement decisions, and competitive dynamics that are not yet settled.

I believe that we're going to have at least one massive, you know, multiundred billion if not trillion dollar American company focused on American first open source. Harry Stebbings

For a multi-hundred-billion-dollar open-source-first company to materialize, several things would need to be true at the same time. The open-source model would need to sustain a business at frontier capability, not just in adjacent tooling or infrastructure built on top of someone else’s weights. The company would need to find a durable revenue model without the margin protection that a closed, proprietary model affords. Regulatory conditions would need to permit open-weight distribution at scale rather than restricting it on national security grounds. And the company would need to capture enough of the developer and enterprise market to justify a valuation that very few technology companies of any kind have ever reached. None of those conditions is guaranteed. Several are actively contested.

Stebbings is not hedging the call. The language is direct: at least one company, at least multi-hundred billion dollars, explicitly focused on this combination of American origin and open-source orientation. The precision is what makes the forecast useful. A vaguer prediction, one that said simply “open source will matter,” would be untestable and therefore uninteresting to track. This one has a shape a reader can return to and evaluate over time.

The risk on the other side of the prediction is worth naming. If Stebbings is wrong, the more plausible conclusion is not that open source loses to proprietary approaches outright. It is that the value open source creates gets distributed across many smaller companies, infrastructure providers, and developer ecosystems rather than concentrated in a single institution of generational scale. That distribution outcome would still represent a significant shift in how AI gets built and deployed, but it would not produce the landmark company Stebbings is forecasting. The difference matters for how investors and builders allocate attention and capital across the next several years.

The call is on record. What rides on it being right or wrong extends well beyond any single company’s fortunes. If Stebbings is correct, it suggests that the open-source model can sustain not just products but institutions of the scale historically associated with dominant proprietary platforms. That would represent a meaningful shift in how large-scale technology value gets created and captured. If the forecast misses, the question is whether it missed on the scale, the national framing, the open-source premise, or some combination of all three. Each failure mode points toward a different conclusion about where the AI industry is heading. The clock is running, and the bet is clear enough that the answer, when it arrives, will be unambiguous.