Sam Altman Wants 100 Million AI Chips – And That’s Just the Beginning


 

The AI Arms Race Just Got Wilder: Sam Altman Eyes 100 Million Chips

Anyone thinking the AI hype was slowing down better think again. OpenAI’s CEO, Sam Altman, is back with yet another jaw-dropping ambition-securing up to 100 million AI chips to power what could become one of the largest computing clusters in history. This bold move signals that the race for artificial general intelligence (AGI) isn’t just heating up-it’s going nuclear (possibly literally).

Altman’s track record of throwing out massive figures isn’t new. Just months ago, he made headlines for trying to raise trillions to fund a global network of chip fabs. While those plans are still in development limbo, he’s already prepping for the next phase: acquiring the raw GPU power to train next-gen models that could make GPT-4 look like a toy.

Let’s break this down. A cluster of 100 million GPUs-assuming each runs at 750 watts-would demand around 75 gigawatts of power. That’s roughly three-fourths of the entire UK’s electrical grid output. Unless OpenAI has a secret handshake deal with the nuclear energy industry, the energy logistics alone are staggering.

Still, the tech industry isn’t shying away. Companies like Microsoft, Google, and Meta are also scaling up, hoarding AI chips like they’re gold in a digital gold rush. Despite all that, computing resources are still stretched thin. OpenAI appears to be following the logic that more silicon equals more intelligence-and possibly a pathway to AGI.

And while some scoff at Altman’s “unrealistic” dreams, it’s hard to dismiss them entirely. The CapEx in AI infrastructure is ballooning. NVIDIA, the crown jewel of this movement, is practically printing money thanks to the demand. Whether this move is visionary or delusional, one thing is certain: Altman is betting big, again.

Critics point to the obvious hurdles-funding, supply chains, and a physical infrastructure that doesn’t exist yet. Others argue that such a massive investment doesn’t guarantee better AI. But love it or hate it, Altman is dragging the industry into uncharted waters-and the splash could be massive.

Related posts

Windows Boot Issue With Ryzen 4000G/5000G APUs and 2x 32GB RAM: How to Fix It

Samsung Sharpens Its 2nm GAA Process to Compete with TSMC

Intel CEO Lip-Bu Tan Faces a Catch-22 in His Turnaround Strategy