Meta Announces Major Investment to Advance its AI Efforts

Following on from xAI launching its Grok 4 model, which it claims is now the best-performing AI system on all fronts, Meta has announced the next steps in its AI development journey.

And it’ll be tough for any other AI developer to match or better this.

Today, Meta CEO Mark Zuckerberg has announced that Meta’s looking to establish “several” multi-gigawatt datacenters in the U.S., beginning with its “Prometheus” project in Ohio, which will be followed by its new “Hyperion” cluster, which is being built in Louisiana.

Both of these new mega-clusters will establish major new compute power for Meta’s AI projects, and will cost, according to Zuckerberg, “hundreds of billions of dollars” to complete.

Which is a mind-boggling amount of money, more than the GDP of some nations.

And this, according to Zuck, is just the beginning.

Meta’s Prometheus data center will be its first major AI data project, aimed at powering Meta’s new “superintelligence” push. The project will incorporate the capacity of some 600k NVIDIA H100 units, which will give Meta a major advantage in compute.

These NVIDIA chips are key for AI development, enabling massive compute power to drive the decision-making of AI systems, which scan through trillions of data sources to come up with the best possible answers to queries.  

For comparison, both OpenAI and xAI reportedly hold around 200k H100s each.

That alone will give Meta a clear lead, but Meta’s also building more data processing capacity on top of that, as it looks to go beyond the limitations of the latest generative AI models.

Which will ideally, for Meta at least, lead us closer to the development of artificial general intelligence (AGI), and machines that actually “think” like humans, and replicate the human brain in circuitry.

The current wave of AI models, while impressive, are not close to this, as they can only replicate what they can find in data patterns based on human inputs. True AGI goes a step beyond that, and could eventually expand human development, by exploring possibilities beyond human thinking.

Which will either take us to the next stage of civilization’s development, or mark the end times for humanity. Nobody knows which path they’ll take us on as yet.

But Meta’s determined to find out, which is why Zuckerberg has put together a team of leading AI experts to explore the potential of AI in all new ways.

These new investments in AI infrastructure will support that effort, and will cost Meta a massive amount of capital in the quest to determine what’s possible on this front.

Which feels slightly concerning, but also inevitable, to some degree.

To be clear, the current generative AI tools are nowhere close to AGI, and there’s skepticism as to whether it’ll even be possible to simulate human-like thinking in machine systems.

But Meta’s seemingly going to find out, with Zuckerberg noting that:

“Meta Superintelligence Labs will have industry-leading levels of compute and by far the greatest compute per researcher.”

It has the money, the data, and now the staff to explore this next-level push.

Which, in the short term, could leave every other AI project in the dust. But in the longer term, you can expect to see a lot of spooky stories about AI working out the inefficiency of humans stemming from Meta’s next-level research.

Meta’s first mega datacenter, “Prometheus,” is expected to be fully activated early next year, so every other AI project will likely be looking to gather up investment and opportunity before that point.

At which stage Meta looks set to become the leading AI project, by an unmatchable margin.