Marques Brownlee
Marques Brownlee5d ago
Tech

Apple Lost the AI Race

7 min video4 key momentsWatch original
⚡
TL;DR

Apple lost the current AI software race decisively — paying Google $1 billion annually for Gemini integration proves it — but might still win the long-term hardware play if on-device AI becomes the standard.

Key Insights

1

Apple paid Google $1 billion a year for Gemini integration into iOS — that's an admission Apple already lost the software race before it even tried.

2

Just has to be good enough — Apple Intelligence doesn't need to be good, just good enough to convince investors Apple is doing something in AI while users stay locked into iPhones for other reasons.

3

Hardware that runs local models — The real AI race isn't about ChatGPT competitors — it's about who makes the hardware that runs local on-device models when cloud computing becomes optional.

Deep Dive

Apple's visible failure in the AI rush

Marques opens by calling it obvious: Apple lost the AI race. When ChatGPT launched, Google, Meta, and Bing all moved fast. Apple did nothing — they wouldn't even say the word AI at their events for months. His counterpart pushes back, arguing Apple's traditional playbook is to wait for early adopters to work out bugs before shipping a polished version. But Marques counters that a multi-trillion dollar company should've at least tried something after watching Claude, Gemini, and GPT proliferate. The stakes felt urgent in 2023, not optional.

Apple Intelligence as corporate theater

The conversation shifts when the guest makes a provocative claim: Apple Intelligence doesn't have to be good. It just has to be good enough for investors to believe Apple is doing AI and for users to not feel alienated. Marques pushes hard here — he lists all the ways Android flagships have better AI features: smarter writing tools, photo editing, generative features across the board. The guest concedes these are mostly gimmicks nobody buys phones for, which suggests Apple's strategy is less about product and more about optics and ecosystem lock-in.

The real race: local on-device AI hardware

The turning point comes when they reframe the entire competition. As on-device AI models improve, reliance on cloud computing shrinks. Eventually, users won't need the cloud at all — everything runs locally on their device. The question becomes: who makes that device? Apple has the iPhone. OpenAI theoretically could make a phone. Nvidia wants RTX Spark hardware. In this future landscape, Apple's advantage isn't ChatGPT-level software; it's the hardware ecosystem and deep integration. Apple doesn't make search engines either, but it owns the device people search on. Same playbook could apply to AI.

Two races, two different winners

They land on a synthesis: there are two separate competitions. In the immediate race for AI software supremacy, Apple took an L and is paying Google $1 billion annually to not look embarrassing — that's a loss. But in the long-term hardware race for local AI devices, Apple still has the iPhone's distribution and integration advantage. The risk is real though — OpenAI or another AI company could enter hardware and outmaneuver Apple the way Apple outmaneuvered Nokia in phones. Marques notes Siri still doesn't work in the background or retain personal context like ChatGPT does, so Apple's integration benefits haven't materialized yet.

Takeaways

  • ✓Apple's $1 billion annual Gemini payment signals they conceded the AI software race before competing — expect integration partnerships, not innovation leadership.
  • ✓Watch whether local on-device AI adoption actually happens; if it does, Apple's iPhone dominance becomes the real defensive moat, not Siri's capabilities.
  • ✓If OpenAI or another AI company releases hardware in the next 2-3 years, iPhone's lock-in advantage disappears — monitor startup announcements in this space closely.

Key moments

0:08Marques declares Apple the loser

“I just feel like... it's so obvious that they lost.”

3:00Apple Intelligence only needs to look good

“I don't think Apple Intelligence has to be that good... it just has to be good enough for investors to believe Apple's competitive.”

4:08The real race is hardware, not software

“As these on-device models get better and better, you go to the cloud less and less. So in the future, someday you will basically never have to go to the cloud... Who's gonna make that device? Apple maybe?”

5:05Two races, two different outcomes

“In race number one, the competition is so far ahead, Apple's kind of already took the L. But in race number two, the finish line is way out on the horizon, and Apple does have a bit of a head start.”

Get AI-powered video digests

Follow your favorite creators and get concise summaries delivered to your dashboard. Save hours every week.

Start for free