OpenAI's Trillion-Dollar Gamble on a Closed-Loop Ecosystem

OpenAI is building an "AI closed-loop economy" centered around itself with unprecedented capital scale. Following collaborations with tech giants such as Oracle, NVIDIA, and AMD, this AI unicorn recently announced another significant development: a strategic partnership with Broadcom to jointly launch custom chips with a total capacity of 10 gigawatts (GW).

According to industry estimates, OpenAI's forward-looking investments have already leveraged trillions of dollars in infrastructure support, covering multiple critical areas such as computing power and cloud services. However, its ambitions may extend even further. Last week, OpenAI CEO Sam Altman stated in an interview that the massive investments in infrastructure represent a strategic gamble at the corporate level, hinting that more deals are in the pipeline.

Yet, the concept of "suppliers investing in customers, who then purchase products" raises the question: Is this closed-loop model leading a technological revolution or merely creating a bubble? The industry has both skepticism and approval. Angel investor and senior artificial intelligence expert Guo Tao believes that the key to ultimate success lies in whether OpenAI can consistently deliver disruptive innovations.

Looking domestically, should Chinese companies follow this path? On October 15, SenseTime announced a strategic partnership with Cambricon to advance the joint optimization of software and hardware, which may be a starting point. However, Chen Ran, founder of the large model ecosystem community OpenCSG, argues that China should not simply replicate OpenAI's closed-loop approach but instead focus on building an open ecosystem.