An analysis said Meta Platforms and Anthropic, major corporations in artificial intelligence (AI), are competitors and partners at the same time.
The New York Times (NYT) on the 27th (local time) analyzed corporate relationships in the AI industry, saying Meta and Anthropic are keeping each other in check while forming a major client relationship.
Mark Zuckerberg, Meta chief executive officer (CEO), criticized Anthropic and OpenAI, leaders in closed AI models, in a 6,500-word essay titled "The future belongs to everyone" on the 10th.
Zuckerberg did not directly name specific corporations or founders, but said, "Instead of centralizing superintelligence, we should distribute it widely so that everyone can control it," and added, "If only one corporation holds superintelligence, the market will be less dynamic and less prosperous than it is today."
He added, "The idea that AI is so dangerous that the only way to make it safe is to concentrate power to the extreme is fundamentally flawed."
But while Meta is criticizing Anthropic's approach to AI development, it has also become Anthropic's largest customer, the NYT reported, citing anonymous sources.
According to the sources, Meta internally expected at one point this year to spend up to $10 billion annually on Anthropic's AI models. Given that Anthropic estimated its annual revenue at $65 billion in July, that is not a small amount.
Examples of competitors cooperating in the AI industry are not limited to Meta and Anthropic. Google and Amazon are developing their own AI models while pledging $73 billion in investment in Anthropic. Microsoft (MS) was OpenAI's largest early investor, but they now emphasize each other's independence.
Nvidia likewise makes AI-specific chips while forging partnerships worth billions of dollars with Meta and Google.
Sources said Meta is also developing and testing a next-generation AI product called "Hatch," based on Anthropic's AI models.
Zuckerberg introduced Hatch as Meta's next-generation innovation product, describing it as "a personal agent that acts on behalf of users 24/7 to help them achieve their goals and improve their lives, health, relationships and finances."
Meta reduced its expenditure on Anthropic somewhat this summer, but it is still said to be paying hundreds of millions of dollars each month. Anthropic likewise proposed that over two years it would purchase up to $10 billion worth of computing power from Meta's data centers.