OpenAI announced the completion of a new round of financing, with a total committed capital of US$122 billion and a post-investment valuation of US$852 billion. For the AI industry, this is not just a huge fundraising, but also a positive increase in computing power, products and distribution entrances. OpenAI is trying to further bundle ChatGPT, Codex and enterprise services into a unified platform.
What is the US$122 billion financing buying?
The most noteworthy thing about this round of financing is not just the amount itself, but the very clear direction of the use of the funds. OpenAI defines "the next stage of AI" as an infrastructure race: whoever can get more stable computing power is more likely to continue to train stronger models, lower reasoning costs, and then turn capabilities into product growth and business revenue.
According to the disclosed information, this round of financing was participated by strategic partners such as Amazon, NVIDIA, and SoftBank, and Microsoft continued to follow suit. OpenAI has also simultaneously expanded a revolving credit line of approximately $4.7 billion. Capital markets are no longer betting on a single model release, but on a complete AI infrastructure system covering training, reasoning, deployment and commercialization.
ChatGPT and Codex are becoming growth engines
OpenAI's operating data in the article shows that it wants to emphasize that it is no longer just a "hot model company." As of the latest disclosure, ChatGPT has more than 900 million weekly active users, more than 50 million subscription users, and current monthly revenue has reached US$2 billion. The corporate business has accounted for more than 40% of total revenue, and plans to be close to the level of consumer business by the end of 2026.
More importantly, AI programming and agent products have been pushed to a higher position. OpenAI defines Codex as a flagship coding agent and claims that its weekly active users have exceeded 2 million, a five-fold increase in three months. The API side processes more than 15 billion tokens per minute, which means that the developer ecosystem is no longer a supporting role, but part of OpenAI's commercial closed loop.
Computing power and super applications are the focus of the next stage
OpenAI has written computing power as the core logic of this article. It made it clear that the infrastructure has expanded from a few core vendors to multi-cloud, multi-chip and deeper collaborative design, covering Microsoft, Oracle, AWS, CoreWeave, Google Cloud, as well as NVIDIA, AMD, Trainium, Cerebras and self-developed chip directions in partnership with Broadcom.
Two signals are released behind this. First, the competition focus of AI companies is shifting from "who releases new models first" to "who can supply intelligence stably in the long term." Second, OpenAI wants to consolidate this capability to the unified portal. For the first time, the article clearly proposes the direction of "AI superapp", which is to integrate ChatGPT, Codex, browsing and broader agent capabilities into a product surface, allowing users to complete understanding, execution and cross-application operations in a single system.
When OpenAI put financing, computing power, ChatGPT distribution and Codex agent in the same article, the industry signal is already clear: the next round of AI competition is no longer just model parameters and list scores, but around infrastructure, unified product entry and enterprise-level deployment efficiency. Whoever can make intelligent capabilities an entrance to daily work will have a better chance of taking away the business dividends of the next stage of AI.
