OpenAI raises $122 billion to expand its AI

OpenAI has closed a $122 billion funding round, giving the company a post-money valuation of $852 billion. It says it will use the capital to expand its models, products and computing infrastructure so it can offer AI to more people and businesses.

The figure places OpenAI among the world’s most valuable technology companies. It also shows the cost of competing in AI: training advanced models and responding to millions of users requires data centers, chips and computing capacity that keeps growing.

A platform that aims to cover everything

OpenAI no longer presents itself as just the company behind ChatGPT. Its strategy combines four pieces:

• Consumers, who use ChatGPT to search for information, write, learn or solve tasks.
• Businesses, which integrate its models into internal processes and work systems.
• Developers, who build applications through its APIs, the services that let other software connect to OpenAI’s models.
• Infrastructure, made up of data centers, chips and cloud partnerships used to train and run those models.

The company describes this system as a cycle: more computing capacity makes it possible to create more useful models; those models attract more users and customers; and the revenue makes it possible to buy even more capacity.

OpenAI says it currently generates $2 billion in revenue per month. According to the company, it reached $1 billion in annual revenue less than a year after ChatGPT launched and was generating $1 billion per quarter by the end of 2024.

It also says it is growing four times faster than companies that defined the internet and mobile phone eras, such as Alphabet and Meta. These figures were reported by OpenAI itself, not verified through an independent audit included in the announcement.

ChatGPT surpasses 900 million weekly users

The company says ChatGPT has more than 900 million weekly active users and more than 50 million subscribers. It also claims to record six times more web visits and mobile sessions than the next-largest AI application.

Its use as a search engine has nearly tripled in one year, according to OpenAI. In addition, ChatGPT’s advertising program reportedly surpassed $100 million in annual recurring revenue less than six weeks after its pilot launch.

In the business sector, OpenAI says this operation now accounts for more than 40% of its revenue and could match its consumer business by the end of 2026. Its APIs process more than 15 billion tokens per minute. A token is a unit of text that the model processes, such as a word or part of one.

Codex, its programming-focused agent, has more than 2 million weekly users, five times as many as three months ago, according to the company. In practice, these tools aim to move beyond answering questions to completing tasks, such as modifying code, preparing documents or coordinating several steps in a process.

More chips, cloud services and data centers

The funding round will also strengthen infrastructure. Nvidia remains the main provider of chips used to train the models and generate their responses, but OpenAI is expanding its partner network so it does not depend on a single technology.

Its strategy includes cloud agreements with Microsoft, Oracle, AWS, CoreWeave and Google Cloud. On the chip side, it works with Nvidia, AMD, AWS Trainium, Cerebras and an in-house design developed with Broadcom. It also has data center partnerships with Oracle, SBE and SoftBank.

The funding is backed by partners including Amazon, Nvidia, SoftBank and Microsoft. SoftBank co-led the round with a16z, D. E. Shaw Ventures, MGX, TPG and other institutional investors.

OpenAI also expanded a revolving credit facility to approximately $4.7 billion. The company clarifies that it had not used the facility when the transaction closed.

The plan: an AI super app

The next announced step is a unified application that brings together ChatGPT, Codex, web browsing and other functions capable of taking action. The idea is that you will not have to switch between separate tools to ask questions, search, code or complete a task.

The difference matters. A chatbot responds; an agent can interpret a goal, use tools and move through multiple steps with more limited supervision. For example, it could research options, compare data and prepare a result within the same workflow.

For OpenAI, the super app is also a commercial strategy. ChatGPT’s familiarity in daily use could make it easier for businesses to adopt the same functions at work. And a single platform would allow the company to add new models without forcing users to learn different products.

For you, the most visible change will be that AI is no longer limited to a question-and-answer window. OpenAI’s bet is to turn it into a layer that searches for information, writes, codes and takes action inside other applications. What remains to be seen is whether that integration genuinely improves productivity and whether the company can sustain the enormous computing cost required to offer it at global scale.