OpenAI, the developer of ChatGPT, recently failed to meet targets for new users and revenue, putting Chief Executive Officer Sam Altman's leadership to the test. OpenAI signed long-term contracts worth about $600 billion (about 880 trillion won) to secure computing resources, the core of artificial intelligence (AI) infrastructure, but there are concerns that if growth slows, it may not be able to cover the computing contract expense.
It is seen as the result of intensifying competition in the AI market as rivals Google and Anthropic have rapidly closed the gap with OpenAI since the second half of last year. To make matters worse, the damages lawsuit trial with Tesla Chief Executive Officer Elon Musk also began this month, and depending on the outcome, OpenAI's initial public offering (IPO) plan scheduled for the end of this year could face setbacks.
◇ "May be unable to pay the 900 trillion computing contract" raises internal concern
According to the tech industry on the 30th, OpenAI is facing a triple whammy of slowing growth, financial concerns related to large-scale AI data center commitments, and legal risks from the $134 billion (about 198 trillion won) dispute with Musk. The Wall Street Journal (WSJ) recently reported that OpenAI failed to meet its internally set targets for first-quarter revenue and new users. It had aimed to reach 1 billion weekly active users (WAU) for ChatGPT by the end of last year, but as of February this year it was reportedly in the 900 million range. Its market share in the consumer AI chatbot segment is still No. 1 at about 60%–65%, but growth is slowing.
Sarah Friar, OpenAI's Chief Financial Officer (CFO), was said to have expressed to management her concern that "if revenue does not increase fast enough as we planned, we may not be able to cover the computing contract expense." According to the WSJ, Friar clashed with CEO Altman in the process, as Altman insisted on additional AI data center investment.
Altman sees ample computing resources as the key competitive edge that will determine long-term growth, and he has pursued an aggressive AI infrastructure expansion strategy. As part of that, OpenAI signed multi-year computing contracts with Nvidia, Oracle, AMD, and Microsoft (MS), but as the company's growth has slowed compared with before, some executives have raised financial concerns, signaling growing internal friction.
OpenAI recently raised $122 billion (about 181 trillion won), the largest amount in Silicon Valley's history, but at its current pace of investment expenditure, it is expected to exhaust those funds within three years.
◇ Market share eroded by Google's "Gemini" and Anthropic's "Claude" offensive
Analysts say OpenAI's stronghold is shaking under fierce offensives from Google and Anthropic. In particular, Google appears to have eaten into ChatGPT's market share by sharply improving the performance of its AI model "Gemini" in the second half of last year. Google is rapidly onboarding new users as it broadens its base with the AI image generation model "Nanobanana," the AI video model "Veo," and the general-purpose world model "Genie."
Unlike OpenAI, which is posting losses, Google is seeing synergy by integrating AI into its existing businesses such as search and advertising, cloud, and video (YouTube), and its performance is on track. Alphabet, Google's parent company, said that as AI-driven demand fueled rapid growth in its cloud business, first-quarter revenue rose 22% from a year earlier to $109.9 billion (about 163 trillion won). Net income jumped 81% to $62.6 billion. AI investments are beginning to translate into tangible business results.
Google Chief Executive Officer Sundar Pichai said, "Enterprise AI solutions became the primary driver of cloud growth for the first time in the first quarter," emphasizing that the number of paid monthly active users (MAU) for the enterprise AI model "Gemini Enterprise" rose 40% from the previous quarter.
This year, rival Anthropic has especially stood out. Anthropic is expanding its influence in the enterprise AI market with its coding tool "Claude Code." Early in the year, it unveiled the AI agent "Claude Co-Work," which can independently perform multi-step tasks, stoking fears of a "software end-times" scenario, and more recently it introduced the AI model "Mythos," equipped with powerful security vulnerability detection, roiling the security market. OpenAI has reportedly missed its monthly revenue targets multiple times as it lost ground to Anthropic in coding and enterprise markets.
Some predict Anthropic's valuation will soon surpass OpenAI's. According to major foreign media, Anthropic is discussing new fundraising at a valuation of about $900 billion (about 1,336 trillion won). If the deal goes through, it would overtake OpenAI's $852 billion (about 1,264 trillion won) valuation. Both OpenAI and Anthropic are seeking funding with the goal of listing within the year, and at the current pace, concerns are mounting that Anthropic could seize the advantage.
On top of that, the start of the $134 billion (about 198 trillion won) damages trial filed by Musk against OpenAI is adding another risk factor. Taking the stand the previous day in federal court in Oakland, California, Musk said, "I was a fool to fund OpenAI believing it was a nonprofit corporations," and likened it to "a museum opening a gift shop to sell off Picasso's works," strongly criticizing the company. At this point, Musk appears more likely to lose, but if he defies expectations and wins, OpenAI would not only suffer massive financial losses but also have to revise its IPO plans.
With an urgent need to broaden its revenue sources, OpenAI is expected to focus on advertising in the short term. OpenAI is showing ads when users chat with the chatbot on the free ChatGPT model and the low-cost plan "ChatGPT Go." OpenAI recently told investors that annual advertising revenue would grow about 40-fold from $2.5 billion this year to $100 billion in 2030.
With the recent end of its exclusive deal with Microsoft (MS) as a starting point, it is also expected to step up efforts to secure corporate customers. Amazon Web Services (AWS) said the previous day that on its AI model integration platform "Amazon Bedrock," it would offer OpenAI's AI model GPT and the coding tool Codex.
OpenAI, meanwhile, dismissed reports of internal leadership rifts as "ridiculous claims." It added, "We work closely every day to secure as much computing resources as possible."