OpenAI CEO Sam Altman has projected that while the costs of building AI systems will continue to rise, user access to these technologies will become significantly cheaper. In a blog post on February 9, Altman explained that gains in AI performance still depend on substantial investments in training and infrastructure. However, he expects the cost for users to access AI services to drop by a factor of ten every year, which could drive widespread adoption. He cited a significant price reduction for OpenAI’s GPT-4 model between early 2023 and mid-2024 as a recent example.
Altman’s post comes just ahead of the Paris AI Action Summit, where he is expected to discuss AI’s future further. His comments follow increased competition from Chinese AI startup DeepSeek, whose low-cost, open-source AI model has shaken the tech industry. DeepSeek’s rapid rise has intensified global concerns over AI development costs and access.
Rising development costs, falling usage costs
Altman highlighted that building advanced AI systems requires ever-increasing computation, data, and infrastructure investments. These costs are likely to keep escalating as companies race to develop more powerful models. However, for users, accessing AI services is becoming far more affordable. Altman predicts a tenfold decrease in costs every 12 months, a trend he believes will accelerate AI’s integration into industries like education, healthcare, and finance.
OpenAI is working to secure funding of up to $40 billion to support future AI developments. Major corporate partners, including Microsoft, NVIDIA, Oracle, and SoftBank, have committed $500 billion over four years to expand AI infrastructure in the United States.
Altman envisions AI systems becoming ubiquitous in various fields of knowledge work, with millions of AI agents enhancing productivity. However, he cautioned that while these systems will excel in certain areas, they will still have notable limitations.
Debate over open source vs. safety
The rise of DeepSeek’s open-source AI model has prompted renewed debate over whether AI companies should adopt a more open or restricted approach. Altman acknowledged that offering greater user control may be necessary, including open-sourcing parts of AI technology. However, he warned that such openness could lead to safety risks, such as misinformation and exploitation by malicious actors.
OpenAI is reportedly reassessing its open-source strategy in response to DeepSeek’s popularity. Altman stressed that striking a balance between user empowerment and security will require trade-offs.
The path to AGI and its economic impact
Altman also addressed the concept of artificial general intelligence (AGI)—a system capable of solving complex problems across multiple domains with human-like intelligence. Although some experts believe AGI is still years away, Altman remains optimistic about its development. He envisions AGI being deployed on a large scale to boost efficiency in various industries.
Economically, Altman predicts that AGI will drive down the cost of many goods, although prices for luxury items and limited resources like land may rise. He also warned of the risks of AGI being used for mass surveillance by authoritarian regimes.
Altman reaffirmed OpenAI’s strong partnership with Microsoft, clarifying that his views on AGI do not affect their collaboration. Microsoft, which has invested over $13 billion in OpenAI, continues to play a crucial role in the company’s long-term growth.
As AI technology advances, Altman’s vision highlights the opportunities and challenges ahead. Balancing development costs, accessibility, and safety will be essential as the global AI race intensifies.
