As the generative AI market continues to heat up, competition between OpenAI and Anthropic has further intensified. Recently, in an investor memo, OpenAI stated that Anthropic "operates on a significantly smaller scale curve" and argued its growth is constrained by computing resources. According to the memo, OpenAI plans to build approximately 30 gigawatts of compute capacity by 2030, while Anthropic's projected compute level as of late 2027 stands at around 7 to 8 gigawatts. OpenAI emphasized that even using the upper bound of Anthropic's projections, its own expansion pace remains "significantly ahead, with the gap widening." Both companies are currently undergoing rapid expansion, with combined valuations exceeding $1 trillion, making them widely regarded as potential IPO candidates. While competing for enterprise markets and developer ecosystems, both face pressure from well-funded tech giants such as Google and Meta. Founded in 2021 by CEO Dario Amodei along with several former OpenAI researchers, Anthropic recently unveiled a new model and launched Project Glasswing, a cybersecurity initiative made accessible to select corporate clients, thereby strengthening its competitiveness in the enterprise sector. Addressing divergent strategies regarding compute infrastructure, OpenAI stressed in the memo that its approach relies on synergistic progress through continuous iteration of hardware infrastructure alongside algorithmic optimization to enhance model capabilities and reduce cost per unit of intelligence. The company believes this "compounding advantage" will further drive product adoption and support scaling services to larger user bases. Meanwhile, Anthropic responded that it is increasing compute investments through collaborations with Google and Broadcom to meet surging market demand.