The CEO launched xAI last year to compete with OpenAI; the startup develops a line of large language models called Grok. In May, xAI raised $6 billion at a $24 billion valuation to finance its AI development efforts.
Colossal-AI is an integrated large-scale deep learning system with efficient parallelisation techniques. The system can accelerate model training on distributed systems with multiple GPUs by applying parallelisation techniques. The system can also run on systems with only one GPU.
Techies statement on Elon Musk's xAI named Colossus
Despite the massive scale and impressive specifications of Colossus, several industry leaders have raised concerns. Reid Hoffman, cofounder of LinkedIn and a significant player in the AI space, was one of the first to downplay the hype. In an interview with The Information, Hoffman described Colossus as mere "table stakes"—essentially, a tool that allows xAI to stay in the race but doesn’t necessarily guarantee leadership.
Chris Lattner, the CEO of Modular AI, said during a panel discussion at The Information's AI Summit last week that Musk's heavy reliance on Nvidia's expensive and finite chips is also inconsistent with the billionaire's effort to build his own GPU, called Dojo, The Information reported.
Meta, Microsoft, Alphabet, and Amazon are all developing their own AI chips even as they continue to stockpile Nvidia GPUs.
"The difference is that Elon has been working on Dojo for many years now," Lattner told Business Insider in an email.
This raises an important question: Can Musk’s xAI truly lead the charge in AI development if it's dependent on third-party chips?
The challenges faced by Colossus
One of the main challenges facing Colossus is scaling its power needs. Running 100,000 GPUs simultaneously requires a staggering amount of energy and sophisticated networking technology. While Musk has touted that the supercomputer was up and running within 122 days, concerns about its actual operational capacity have surfaced.
In fact, environmental advocacy groups have criticised xAI for installing unpermitted gas turbines to meet the data centre’s energy needs.
The Southern Environmental Law Centre claimed that xAI was operating at least 18 unpermitted turbines, raising both environmental and regulatory concerns. To power the full 100,000-chip setup, xAI may need an additional 100 megawatts of power—far more than what is currently available.