According to foreign media The Information, OpenAI CEO Sam Altman revealed to employees this week that OpenAI’s annual revenue has reached $1.3 billion. In other words, OpenAI's average monthly revenue is more than $100 million, a 30% increase from this summer, when it revealed annual revenue of $1 billion. Since launching the paid version of ChatGPT in February this year, OpenAI has experienced significant revenue growth, mainly from user subscriptions to its conversational chatbot.


Looking at the timeline a little further, OpenAI’s full-year revenue last year was US$28 million. If OpenAI’s revenue this year is US$1.3 billion, that would be a year-on-year increase of more than 4,500%, which is equivalent to an increase of more than 45 times.

Some are happy and some are sad. When OpenAI took off with large models, the performance of another star AI cutting-edge research institution, Google DeepMind, showed decline.

Just yesterday, it was revealed that Google DeepMind's revenue in 2022 will fall by 21% year-on-year to approximately US$1.3 billion; its profits will fall by nearly 40% year-on-year. DeepMind's previous profits mainly relied on the continuous infusion of Google's parent company Alphabet. OpenAI, which is riding the wave of generative AI, has single-handedly matched DeepMind's revenue by relying on the commercialization of large models and ChatGPT.

According to this trend, OpenAI's performance will surpass DeepMind's just around the corner.

Since the release of ChatGPT, OpenAI has become a closely watched barometer of AI demand. ChatGPT and large models have not only quickly entered various life and industry scenarios, but also demonstrated significant growth in commercialization performance, which is undoubtedly a "boost" for the entire AI industry.

01.

Revenue surges, OpenAI valuation may reach $80 billion

OpenAI's revenue growth could help boost the company's valuation.

OpenAI has organized a takeover offer in which outside investors will be able to buy shares from some employees, sources said, marking the second such sale the company has held this year. According to the Wall Street Journal, OpenAI employees may soon try to sell shares for a total market value of $80 billion or more.

The valuation of US$80 billion has nearly tripled in a short period of time compared to the US$27-29 billion valuation when financing was completed in April this year.

The rapid growth of OpenAI also highlights the prospects of ChatGPT-like large language models (LLM).

Although this technology is far from perfect, some LLM developers, including Andrej Karpathy, the AI ​​guru who returned to OpenAI from Tesla, have begun to call it the next software operating system, using it to write and run code, access the Internet, and retrieve and reference files.

In June of this year, The Information first reported that Altman hoped to build ChatGPT into a "super-intelligent work personal assistant."

According to sources, Altman has discussed developing AI hardware devices with former Apple design director Jony Ive and SoftBank CEO Masayoshi Son.

Whether in terms of storytelling or actual revenue, OpenAI has set a good example for the industry.

02.

OpenAI plans to raise US$100 billion, "not short of money"

It’s unclear what OpenAI’s development costs will be in relation to revenue this year, which also includes the consumption of computing resources to power its AI. The computing resources required by OpenAI are run on Microsoft's cloud servers, and it is unclear how much revenue OpenAI can get from these transactions.

According to sources, OpenAI recorded a loss of US$540 million last year due to the development of ChatGPT and GPT-4. Well-known industry leaders such as Zoom, IKEA, and Volvo all say they use GPT-4, but some companies obtain it through OpenAI and others through Microsoft. OpenAI still needs to share revenue with Microsoft.

According to previous reports by The Information, in order to achieve OpenAI’s ambitious goals, Altman himself has privately stated that the company may try to raise as much as $100 billion in funding in the next few years.

On top of that, Microsoft has committed more than $10 billion to OpenAI to fund its operations and AI development in exchange for using OpenAI's technology. Investors expect OpenAI to raise another large funding round as early as this year, in addition to its ongoing employee buyout offer.

OpenAI's recent growth, driven by a series of product launches, may ease competition from archrival Google and its upcoming Gemini large language model, as well as generative AI startups like Anthropic. In August, Anthropic told investors it had reached $100 million in annualized revenue.

03.

OpenAI eats up Google DeepMind falls

While ChatGPT has already generated revenue, OpenAI is rushing to strike while the iron is hot.

According to people familiar with the matter, OpenAI, which has been established for eight years, plans to hold its first developer conference in San Francisco on November 6. OpenAI will improve ChatGPT and other functions to speed up the performance of its LLM.

OpenAI pays special attention to cost optimization this time. For example, one of the major updates is that OpenAI will add services such as memory storage to development tools, which is said to reduce the cost of application developers to 1/20 of the original cost.

OpenAI relied on generative AI to turn around, but Google DeepMind's life was not easy.

According to CNBC, according to a document recently submitted to the British government agency, both revenue and profit of Google DeepMind plummeted last year: revenue was approximately 1.081 billion pounds (approximately 1.329 billion U.S. dollars), a year-on-year decrease of 20.8%; profit was approximately 60.9 million pounds (approximately 74.9 million U.S. dollars), a year-on-year decrease of nearly 40%.

It is worth noting that before the advent of the GPT-3 large model, DeepMind was even more popular. It has a series of star AI models that have defeated humans, such as AlphaGo, AlphaGoZero, and MuZero. Later, it was also blessed by AlphaFold, which solved major scientific research challenges. It has a halo of its own.

Cutting-edge models in the directions of AI+science and AI+games are following a "money-burning model." The reason why DeepMind has achieved profitability in recent years mainly relies on the business path of paying for the commercial projects of parent company Alphabet. Last year's poor performance is obviously a wake-up call for DeepMind's business model.

As the trend of generative AI becomes more and more popular, OpenAI, the leader of GPT large model, becomes more and more popular, but DeepMind has quietly disappeared. It was not until the news of the merger with Google Brain was announced that it returned to the focus of attention in the AI ​​field.

It is reported that before merging with Google Brain, DeepMind cut employee costs by 39% last year, but still reached 594.5 million pounds (about 731 million U.S. dollars). Following these cuts, DeepMind also announced in January that it would close its first international AI research office in Edmonton, Canada.

A Google DeepMind spokesperson told CNBC that the company will continue to invest in basic research and world-class interdisciplinary teams because it sees the promise of AI to unlock scientific discovery.

04.

Conclusion: The takeoff of generative AI brings a new dawn for AI commercialization

OpenAI is the benchmark for the current AI industry. Whether OpenAI can make money, how much it makes, and whether it is growing fast is a key reference for major manufacturers and entrepreneurs in the industry. The news that OpenAI’s annual revenue is US$1.3 billion has given people a glimpse of the prospects for AI companies to generate revenue from large-model algorithm software. Especially compared with DeepMind's weak revenue, this further confirms the commercialization prospects of generative AI.

At the same time, OpenAI’s investment in ChatGPT and GPT-4 is also huge. Combined with the news that DeepMind’s profits fell by nearly 40% last year, it can be seen that the industry still needs to pay attention to the high cost of cutting-edge AI models. Even companies like OpenAI and DeepMind that are backed by giants have to be careful about their budgets, not to mention more start-up AI startups.