DeepAI

ALT Q. You and your Anthropic cofounders started discussing commercialization as early as 2021; is that correct? MR. GOOCH: Speculation. THE DEPONENT: When we formed Anthropic, we very seriously considered making Anthropic a nonprofit organization. And several of the cofounders advocated strongly that we should become a nonprofit organization in our pursuit of safe and beneficial AI. However, in those discussions, we also speculated upon what kinds of resources might be necessary in order to have the positive impact that we hoped to have. And many considerations were at play, but one of them was that it wouldn't be possible to have the positive impact we hoped to have without some type of customer base or commercial activity. At the time, the most popular reason for -- well, I think there were many different views, but one of the popular reasons advocating

ALT for not being a nonprofit was to have customers so that the benefits of AI could be disseminated, and so that we could get information from users about how AI was being used in order to understand how AI was useful and improve it. I mean, at that time, I would recall that this was several years before ChatGPT or -- or any now-familiar AI system. There was a great deal uncertainty about what form AI would take. So it's the case that we considered the possibility of commercialization at that time. But the overall view of the founders in 2021, I would say, was strongly against commercialization, and towards the idea that we would effectively never pursue it, and that we might be funded primarily by philanthropists in perpetuity. Q. So Anthropic was not founded as a nonprofit, correct? MR. GOOCH: Asked and answered. THE DEPONENT: I already answered that question. Q. Anthropic was not found as a nonprofit, correct? MR. GOOCH: Asked and answered. THE DEPONENT: I answered that question.

ALT Q. (By Mr. Hailey) What's the answer? MR. GOOCH: Asked and answered. THE DEPONENT: Anthropic was founded as a public benefit corporation. Q. (By Mr. Hailey) You and your Anthropic founders started discussing commercialization as early as 2021, correct? MR. GOOCH: Misstates testimony. THE DEPONENT: When we were founding Anthropic and considering the possibilities, we very seriously considered founding Anthropic as a nonprofit. And several of the founders had a strong preference initially for creating a nonprofit organization at the time. When we were engaged in these discussions, there were many different discussions of possible situations. The principal consideration at the time, I believe kind of the overarching consideration, was a question of how much resources would be required to have the positive impact on AI safety and beneficial deployment of AI as we hoped to have. And some argued that the total amount of resources we would need would be small enough that

ALT we could easily succeed as a nonprofit organization and never pursue any commercial motive, nor raise funds from commercially motivated investors. However, we also were open to the possibility that the level of resources we would need or the form it would take when we were trying to deliver the potential benefits of AI to the world would require us to have customers. And, thus, we compromised and founded Anthropic as a public benefit corporation to give us the flexibility to pursue either avenue. Q. (By Mr. Hailey) You sell commercial AI products now, right? A. Anthropic, as of November 2025, has commercial deployments of Claude. Q. And you could give those AI products to customers for free, right? A. I don't understand the question. Q. You could provide AI models to customers for free? A. I think that just as you could provide your legal counsel to all our clients for free, we could, in the same sense, I suppose, provide AI for free. (The last three answers are in bold.)