When senior researchers and leaders, including brothers Dario and Daniela Amodei, left OpenAI in 2020 to found Anthropic, they weren't just running away from conflict within their organization, they were running toward a new vision for developing artificial intelligence. The core philosophical difference that drove the founding team was the belief that safety, reliability, and robust guardrails should be built into Frontier AI systems from the beginning, rather than being treated as aftermarket or retrofitted product features. This belief system, which elevated prudent scaling and risk mitigation, unexpectedly became Anthropic's decisive commercial advantage in fierce competition from OpenAI, Google, and Meta.
Daniela Amodei, president and co-founder of Anthropic, spoke to CNBC about the company's unique trajectory and affirmed that its founding ethos is rooted in aligning safety with business goals. She said the founding team had a long history of collaboration and shared deep “like-minded values”. This fundamental adjustment resulted in a then-seemingly counterintuitive strategy that prioritized prudence and robustness along with competence. Reflecting on this discrepancy, Amodei said, “At the time, when it seemed so novel, we actually had something like this: a belief that these two things were actually interrelated and that they would go together.” This belief that safety and commercial success are not mutually exclusive, but mutually reinforcing, has established Anthropic's market position.
A key strategic decision based on this philosophical stance was for Anthropic to focus on the enterprise market. While OpenAI gained global attention and consumer adoption with the viral launch of ChatGPT, Anthropic focused its efforts on developing Claude, a family of models tailored for enterprises. The logic was simple. Unlike consumers, business customers demand ironclad reliability, rigorous security, and strict compliance. Anthropic's safety-first DNA naturally gave it an edge in all of these areas. Venture capital partners quickly recognized the long-term value of this approach. Sameer Dholakia of Bessemer Venture Partners articulated the investor's perspective, saying, “We knew Anthropic's focus on safety and trust would work very well with enterprise buyers, and that proved to be true.” Enterprise customers have lower churn risk and higher long-term value compared to volatile consumer markets, providing a stable base for the large investments needed to compete in the AI infrastructure arms race.
This B2B focus has allowed Anthropic to avoid the frenzy of high-combustion consumer competition and focus on deeper integration with Fortune 500 companies, hedge funds, and global institutions such as Novo Nordisk and the Norwegian Sovereign Wealth Fund. The company's revenue growth, from zero to hundreds of millions in just a few years, shows the effectiveness of this strategy. The company's business activities currently account for approximately 85% of its revenue, a stark contrast to OpenAI's consumer-centric usage base. Additionally, Anthropic's enterprise-centric approach has provided a “pure barometer” of real economic value by tracking real business applications rather than consumer novelties, according to CEO Dario Amodei.
But pursuing frontier AI requires incredible computational power, leading to the industry adage that “computing is destiny.” Despite its cautious scaling approach, Anthropic still requires billions of dollars in hardware investments to train and service its increasingly powerful models. Rather than relying solely on a single cloud partner, Anthropic pursued a multi-cloud strategy and secured large-scale compute commitments from all three major hyperscalers: Google, Amazon Web Services (AWS), and Microsoft. These deals collectively result in hundreds of billions of dollars in infrastructure and investments and are critical to the company's survival. Dario Amodei emphasized the importance of these infrastructure decisions and explained the need to anticipate requirements years in advance. “We literally have to decide now… how much compute do we need to buy in early 2024 to deliver the model in early 2027?” These circular deals, where the cloud provider invests in Anthropic and Anthropic spends that money on cloud capacity, are effective strategic co-investments that secure Anthropic’s place in the infrastructure stack.
Beyond commercial success, Anthropic's commitment to safety is demonstrated by aggressive research into existential risks. The company's Red Team, a dedicated internal division tasked with discovering failure modes before customers do, stress-tests models for risky features such as biological exploits, attacks on critical infrastructure, and “agent misalignment,” where an AI system acts maliciously to achieve a goal. Anthropic is unique in publicly disclosing these vulnerabilities, providing transparency that serves both public and commercial missions and building trust with a cautious corporate sector. This research directly addresses the profound implications of highly capable AI for national security. Dario Amodei has been vocal about the urgency of ensuring democratic leadership in this area, arguing that what they are building is “the growth of a unique capacity to have unique national security implications, and democracies need to get there first. This is absolutely essential.” This focus elevates competition from mere market competition to a struggle for the future stability of the global economy and security.
The competition between Anthropic and OpenAI is now reshaping both companies. OpenAI is now rushing to pivot around the funding of tenacious and reliable companies, while Anthropic is learning, through necessity and success, how to operate in the public eye. The company's intentional, long-term focus on responsible scaling and security has not slowed its growth. Rather, it has provided the credibility needed to secure the company's foundational contracts and the large-scale computing contracts that will fuel its continued rise.
