Memory, Not Compute, Will Decide the AI Race

Everyone is fixated on raw compute power—but that’s yesterday’s game. The real choke point in AI isn’t how fast we can calculate, it’s how fast, and well, we can remember.

Memory has become the hidden kingmaker. The ability to move, store, and retrieve vast amounts of data at speed is now what separates breakthroughs from bottlenecks. Margins follow memory, not processors. Contracts are secured not by hype, but by reliability, efficiency, and the ability to scale without collapse.

And here’s the twist: regulation, geopolitics, and infrastructure will shape this frontier even more than clever chip design. Restrict memory access and you choke innovation. Miscalculate the demand for bandwidth and your “cutting-edge” AI stalls in deployment. Meanwhile, new architectures—stacked layers, hybrid bonding, or even memory alternatives—are reshaping what “advantage” looks like.

As I explore in Now What? How to Ride the Tsunami of Change, the biggest risks come from chasing the wrong signals. Compute may dominate headlines, but memory defines outcomes.

So here’s the challenge: are you still measuring success in teraflops, or are you rethinking what really drives intelligence at scale?

Read the full article on Financial Times.

----

💡 We're entering a world where intelligence is synthetic, reality is augmented, and the rules are being rewritten in front of our eyes.

Staying up-to-date in a fast-changing world is vital. That is why I have launched Futurwise; a personalized AI platform that transforms information chaos into strategic clarity. With one click, users can bookmark and summarize any article, report, or video in seconds, tailored to their tone, interests, and language. Visit Futurwise.com to get started for free!