If you're just getting caught up on the most powerful computing models that exist today in America, you may have heard the term Colossus, but not the spanish-language moniker 'El Capitan.'

While the former is something created by Elon Musk's XAI company, which I reported on a couple of weeks ago, the latter is a corporate project developed by Hewlett-Packard that outstrips Colossus in terms of pure computational capability.

So how do these two systems stack up?

According to reported data, Colossus has a greater number of GPUs. The system has an estimated 100,000 Nvidia Hopper GPU units, all linked together to provide quite a bit of number-crunching or other computation.

By comparison, El Capitan is reported to have 30,000 AMD APUs, and each of these elements has a CPU and GPU in one package.

But what about performance?

Although Colossus's chip count may be higher, experts give El Capitan the blue ribbon for power. This system provides an estimated 1.742 exaflops or quintillion calculations per second in terms of sustained real-world performance. Colossus is estimated to be able to deliver 6 exaflops in theoretical peak performance. That might look like a bigger number (and it is) but it's the context that's important.

For reference, here's what ChatGPT has to say about it:

"Experts generally agree that El Capitan is more powerful for tasks requiring high-precision computations, such as scientific simulations and national security applications. In contrast, Colossus excels in AI and machine learning tasks that can leverage its high throughput in single-precision calculations. Therefore, the determination of which is more powerful depends on the specific application and computational requirements."

In the case of the Colossus data center, Elon Musk suggested it was built to empower AI capability for the Grok chatbot that's part of premium services on the X platform.

By contrast, here's what stakeholders say about El Capitan - that the system "will support national security efforts, and will accelerate both classified and unclassified studies on energy, security, climate change, power grid modernization, drug discovery, and other areas."

As for the similarities between the two projects, both are big. Both rely on liquid cooling, but one is aimed at adding to the United States's national security position, where the other one is a private sector initiative.

Both of these systems also show us what the data center of the future will look like. These are massive engines that can support the kinds of reasoning and high-performing AI models that we've just engineered into existence. In other words, we need those big data centers to feed these models. We also need a lot of energy. That's why the U.S. is now pursuing more small nuclear capability to catch up with China, where smaller, safer nuclear power plants are the norm.

Here is some of my reporting energy from just weeks ago.

Market articles talk about how AMD is trying to catch up with Nvidia in terms of chip market power.

Specifically, though, AMD has yet to really close the lead that Nvidia has in the market. I reported recently about Nvidia's rapid rise to eclipse Apple and Microsoft as the dominant tech stock in the American economy.

You also have to look at the context, with TSMC supplying the raw materials for Nvidia's chips.

In any case, it's worth keeping an eye on the manufacturers, as well as the specific use case projects like Colossus and El Capitan. Each of these deserves scrutiny as it works behind the scenes to deliver robust artificial intelligence to the world.