Liz’s view
The AI boom has minted a new crop of fresh-faced celebrities like Sam Altman and Dario Amodei. When I got the tip in January 2023 that Microsoft was investing $10 billion in OpenAI, I had to Google it. Now Altman is a household name.
But look across the AI landscape, and the faces are notably wizened. Oracle was born in the floppy-disk era; Larry Ellison made his first billion in 1992. Microsoft invented the operating system before memorably blowing mobile. AMD is three months older than the moon landing, and Intel is only a year younger. Even Nvidia was founded a year before Perplexity’s Aravind Srinivas was born. Altman’s tech origin story is founding a social-media app with a fashionably missing vowel; AMD CEO Lisa Su’s involves loose wires and, later, the world’s top prize in microelectronics.
This Space Cowboys vibe makes sense because this tech boom looks less like the last one — social apps powered by subway ads and nap pods — than the 1980s PC revolution, powered by tinkerers with motherboards and soldering irons. Unlike asset-light software unicorns, AI is not infinitely scalable. It scales brutally, and expensively, in a bare-knuckled fight for scarce resources: chips, electrons, export licenses, land, zoning permits, President Donald Trump’s approval. There may be more proprietary value in knowing how to cram GPUs into racks so dense they are cracking data centers’ floors than in writing algorithms, my colleague Reed Albergotti noted recently.
And AI’s financial success lies more in selling its products to businesses than to consumers, as so many of last decade’s “Uber-but-for-fill-in-the-blank” startups aimed to do. The real money in AI will ultimately come from corporate users, and the current buildout requires inking deals with suppliers and hyperscalers, CEO to CEO. So perhaps it isn’t surprising to see a generation of enterprise B2B giants on center stage.
Whether these experienced hands can keep up with their mad scientists is less clear. But this may be the first tech revolution that doesn’t vaporize the dinosaurs, the way mobile killed Nokia and threatened Microsoft, but grafts their DNA onto something new. And with tech as powerful as AI, the more adults in the room, the better.
In this article:
The View From Semafor’s Reed Albergotti
The fact that our AI infrastructure is being built by some of the same companies that were around at the dawn of the web is really an indictment on American innovation. The country took its foot off the gas after the fall of the Berlin Wall. We pulled back on government R&D funding that laid the groundwork for the internet era. We stopped pursuing the big ideas that got us here. At a fundamental level, we have been improving and iterating on an old invention and networking computers together. Even the algorithms — neural networks — were conceived decades ago. The difference is that we have a lot more compute power than we used to. It should be a worrying sign for those who want to see the US retain its technological edge.
You can get more of Reed’s views in Semafor’s Technology newsletter here.
Notable
- A newly relevant line from Bill Gates’ 1976 open letter: “Most of you steal your software. Hardware must be paid for.” He was complaining, and software dominated the tech value chain for decades, but that’s changing.