OpenAI and Anthropic have entered not only the competition over artificial intelligence (AI) model performance but also a race for credit ratings. This is because simply raising investment funds will not be enough to cover the massive costs of data centers and AI semiconductor investments; they must also tap into the corporate bond market.
According to a report by the Financial Times (FT) on the 12th, Goldman Sachs and Morgan Stanley recently discussed with major credit rating agencies the possibility of securing investment-grade ratings on behalf of OpenAI and Anthropic, respectively. If both companies successfully complete their initial public offerings (IPOs) and obtain investment-grade ratings, they could access a corporate bond market worth approximately 11.7 trillion won (about 16 quadrillion won). The ratings have not yet been finalized.
Investment grade essentially means the market evaluates that "it is relatively safe to lend money to this company." Generally, the higher the credit rating, the lower the interest rate at which corporate bonds can be issued. It also becomes easier to attract funds from large institutional investors such as pension funds and insurance companies, which are restricted to purchasing bonds above a certain rating threshold.
Both companies are currently operating at a loss. Although OpenAI and Anthropic are rapidly increasing their revenues, the costs associated with training cutting-edge AI models and providing services remain enormous. They must continue to invest in securing GPUs and funding the data centers, power supply, and networks required to run them. Neither company has yet achieved positive free cash flow. It is reported that credit rating agencies will assess the ratings after reviewing the IPO results and actual financial conditions.
Until now, both companies have relied heavily on equity investments from venture capital (VC) firms and institutional investors. In data center projects, they have also been able to borrow money under relatively favorable terms with the help of high-credit partner companies such as NVIDIA, Oracle, and Google.
However, as AI investment amounts have grown into the hundreds of billions of dollars, there is an increasing need to secure long-term funding based on their own creditworthiness. Going forward, even among companies possessing similar levels of AI technology, investment speeds may vary depending on which company can borrow more money at lower interest rates.
The battleground of the AI competition is expanding beyond model performance and GPU acquisition into financial markets. It is no longer just about "who can build a smarter AI," but also about "who can borrow money more cheaply" that has become a key measure of competitiveness for AI companies.
in TECH
