Artificial General Intelligence (AGI), the ultimate goal of artificial intelligence, has long been regarded as a milestone in technology and philosophy to reach the human level or even surpass human intelligence. Recently, however, OpenAI and Microsoft have taken a different approach and made a unique definition of AGI based on revenue capacity: when an AI system can generate at least US$100 billion in profits, it can be considered to achieve AGI.
According to media reports, OpenAI signed an agreement with Microsoft last year that Microsoft will lose the right to use its technology when OpenAI’s AI system reaches the $100 billion profit threshold. The agreement has sparked a broad discussion about the definition of AGI, which has traditionally been thought of as a system with intelligence equal to or superior to that of humans rather than a measure centered on financial performance. The terms add complexity to the partnership between OpenAI and Microsoft. It is speculated that OpenAI may announce an AGI in advance to shorten the time frame for Microsoft to acquire the technology. However, OpenAI is still far from reaching its goal of $100 billion in profits.
In fact, OpenAI expects to face huge losses this year and has informed investors that its earnings may not be realized until 2029. On a technical level, while OpenAI’s o3 inference model has reached a standard close to human inference ability in some tests, the high computational cost limits its commercial potential. It makes the financial definition of AGI more challenging to achieve. It is reported that OpenAI and Microsoft have also recently negotiated relevant terms, and there are even rumors that the two parties may modify the agreement and remove the condition linked to AGI income, with the goal of attracting more investment for OpenAI.
While OpenAI’s technological breakthrough is seen as an important step towards AGI, there is currently no unified standard to measure when AGI will be achieved. Whether it’s the philosophical perspective of human equivalence intelligence or the tens of billions of dollars of profits from a commercial perspective, it reflects the see-saw between technology and profit of this concept. Technological development requires not only breakthrough innovation but also marketization and commercialization capabilities.
