The AI Race: Imagination vs. Infrastructure

For decades, massive firms like Meta, Google, and Apple have dominated the tech industry. But when we talk about artificial intelligence, these aren’t the companies that most people think of. Instead, it’s tools like ChatGPT, Midjourney, and Runway that are shaping public perceptions of AI.

Over just a few years, AI-native companies like OpenAI, Anthropic, and Stability AI (and tools developed by them) have gone from virtual unknowns to household names. That’s a very big problem for established tech giants, and it clearly has them on the back foot right now.

Although the big players have their own AI tools that people are actively using – Meta AI, Apple Intelligence, and Grok – many are doing so only because those services are embedded within the tools and ecosystems that they’re already familiar with: Instagram, iOS, X, etc. When it comes to innovation, however, most early adopters are looking elsewhere for their fix.

Missionaries and mercenaries

Tech giants are now competing against the scrappy, agile startups that they used to be. 

And, as the incumbents play catch up, they’re finding themselves in an unfamiliar position: chasing innovation, rather than setting the pace of it as they once did. But, instead of innovating their way out of the hole they’ve landed in, major players are increasingly looking to bring in talent from elsewhere or lean on existing services to help them up their game. 

Apple, for example, is reportedly considering outsourcing its core LLMs to OpenAI or Anthropic. It’s a move that feels unthinkable for a company that has historically been so eager to build everything in-house. Meta, meanwhile, has recently announced a major hiring spree to staff up a new “Superintelligence” team headed up by former Scale AI CEO Alexandr Wang. 

OpenAI’s Sam Altman has since slammed that move, which has poached talent away from Altman’s own company, arguing that attempts to import culture from elsewhere are always destined to fail and that “missionaries will beat mercenaries.” Altman’s subtext is clear: culture stems from the courage and conviction needed to take a risk and build something from nothing.

Of course, in the AI race, doing anything with nothing is difficult because building and running AI services is eye-wateringly expensive. Without significant investment, sustained growth is impossible. That’s a reality that many smaller AI-native firms are grappling with right now.

Deep pockets vs. deep thinking

Various well-known tools like GitHub’s Copilot and ChatGPT have already lowered their usage limits, while tools like Midjourney and Runway offer tiered pricing models with free offerings that are very limited. Even OpenAI recently announced that they need to 40x their revenue to become profitable. (Altman better hope that his investors are in it for the long haul…)

To put it simply, companies looking to dominate the AI space need deep pockets. Established tech giants like Meta and Apple fit that description. They have the infrastructure, they have the existing user bases, and they have the money. Whether or not they have the big ideas, they may well be able to wait out the competition until they are the last ones standing. 

In the meantime, it appears that many of these legacy companies are seeking to buy their way back into contention by hiring and acquiring what smaller AI firms are building from the ground up. It’s another example of a longstanding approach in the tech scene – c.f. Microsoft and Blizzard, Salesforce and Slack, or Meta and Instagram – if you can’t beat ‘em, buy ‘em.

Still, AI-native companies are very much the ones in the driving seat here. It’s not an overstatement to say that they’re the ones currently defining what AI is and what it can do. In the same way that Hoover, Xerox, and Jacuzzi have all become genericised trademarks, the names ChatGPT and OpenAI have already become synonymous with AI. 

Though that’s not to say that Google or Meta can’t catch up – or even overtake them. 

Working hard or hardly working?

Outpacing smaller companies that have innovation and agility baked into their DNA is a big ask, but granting their shiny new teams free rein to work quickly, take risks, and potentially make some mistakes along the way could pay off when it comes to counteracting their late starts. 

In the tech space, it’s not unusual for competitors once seen as the underdog to finish strong and come out on top. Right now, it just so happens that the incumbents are the underdogs. These recent strategic pivots may represent the beginning of a comeback, or they may turn out to be last gasp attempts at remaining relevant. Time will tell which turns out to be true. 

One thing’s for sure: if anyone does manage to overtake the current leaders in the AI space, they need to be planning their next moves long before they even start to close that gap. 

Because it feels very much like it’s imagination, rather than infrastructure, that will win this race.