OpenAI CEO Sam Altman's recent multibillion-dollar partnerships with Samsung Group and SK Group for the Stargate AI infrastructure project have intensified scrutiny over the company's financial sustainability. With operating losses of US$7.8 billion in the first half of 2025 alone, the AI giant's "invest now, profit later" strategy is testing investor confidence despite its aggressive global expansion.
South Korea emerges as key Stargate partner
On October 1, 2025, Altman met with key figures, including South Korean President Lee Jae-Myung, Samsung Group chairman Lee Jae-yong, and SK Group chairman Chey Tae-won, pledging cooperation in AI development. In addition, Altman signed letters of intent (LOI) with Samsung and SK to support the Stargate AI infrastructure project, which include long-term agreements with Samsung Electronics and SK Hynix to supply as many as 900,000 DRAM chips per month, potentially guaranteeing substantial export volumes for South Korean memory makers over the coming years. Altman also inked an MOU with South Korea's Ministry of Science and ICT (MSIT) to advance AI transformation and ecosystem development, supporting the government's goal to become one of the world's top three AI powers.
As to why OpenAI chose South Korea as a critical Stargate partner, South Korean media have attributed the decision to the country's robust AI ecosystem, strong industrial capabilities, high acceptance of AI services, and proactive government policies. According to recent reports from iNews24 and other outlets, South Korea ranks second worldwide after the US in ChatGPT paid subscriptions, with more than 10 million monthly active users.
Concerns rise over aggressive expansion
Despite these developments, OpenAI's heavy reliance on external funding and rapid global partnerships has raised skepticism. Although OpenAI's stance is to take proactive measures to secure its leading position in the global AI race, Chosunbiz has reported growing uneasiness within the industry about the sustainability of these actions, as OpenAI continues to burn through cash in its pursuit of aggressive expansion.
According to CNBC, investors and analysts are now questioning the astronomical scale of OpenAI's expenditures and its increasingly complex network of infrastructure partnerships. Currently valued at around US$500 billion, OpenAI has doubled its valuation within a year, but remains unprofitable. Tech-industry outlet The Information has revealed that in the first half of 2025, OpenAI generated US$4.3 billion in revenue, but suffered operating losses of US$7.8 billion.
Record losses strain investor confidence
The record losses stem primarily from hefty investments in AI development, ChatGPT operations, and infrastructure. In just the first half of 2025, OpenAI spent as much as US$6.7 billion on R&D, and another US$2.5 billion on employee stock incentives to retain top AI talent. Although the company did secure approximately US$50 billion in funding last year, as well as a landmark US$100 billion investment deal from Nvidia, these amounts may still fall short given OpenAI's projected decade-long expansion costs.
As expenses mount, bridging the loss gap will take even more time. Market expectations suggest OpenAI must grow its paying user base by the hundreds of millions to achieve profitability in the mid-to-long term. To diversify revenue streams, OpenAI recently introduced shopping features within ChatGPT, targeting the AI e-commerce market, and launched the social platform Sora aimed at creating and sharing AI-generated videos.
Although OpenAI is rapidly expanding its AI ecosystem footprint, its "invest now, profit later" approach is fueling investor anxiety. While large-scale R&D and infrastructure spending will help the company maintain its pole position in terms of technology, failure to establish stable profits or scale up its user base soon could strain both cash flow and investor confidence. All eyes will be on how OpenAI balances growth and profitability to survive intense competition during the AI gold rush.
Article translated by Kevin Wang and edited by Jerry Chen
