The State of the Big AI Players (mid-2025)

Jeff -

Hey friends - Ben Thompson recently published an extremely insightful article titled "Checking In on AI and the Big Five" breaking down the state of AI among Google, Meta, Apple, Amazon, and Microsoft.

I've distilled his 5,000+ word analysis into a 5-minute read but I highly recommend checking out the full article (as much as I wish I did, I have no relationship with Ben Thompson 🥲).

A bit of context: Disruption vs. Sustaining Innovation

Before diving into each company, you need to understand one key concept. Professor Clay Christensen identified two types of innovation:

Sustaining innovation = Making your existing products better. Think of how cameras on phones keep improving each year. The business model stays the same, just better.

Disruptive innovation = New technology that completely changes or threatens your business. Think of how Netflix killed Blockbuster, or how smartphones disrupted digital cameras.

This matters because AI can be either one, depending on the company. Let's see how each tech giant stacks up:

Apple 🍎

  • Infrastructure: Minimal (they don't own many AI servers)
  • Model: None (no proprietary ChatGPT competitor)
  • Partner: OpenAI? (there are rumors of potential collaboration)
  • Data: No public data differentiation, potential private data differentiation
  • Distribution: Apple devices
  • Core Business: Devices are complementary to current AI use cases
  • Scarcity Risk: Could lose differentiation in high-end hardware
  • Disruptive/Sustaining: Sustaining
  • New Business Potential: Robotics

What this means: Apple is playing it safe. They're not building their own AI but making sure iPhones work great with ChatGPT and other AI apps. But there’s a massive risk: If AI becomes THE reason to buy a phone and they don't have the best AI, they could lose their premium position. Thompson suggests they should either go all-in with OpenAI partnership or buy an AI company like Mistral.

Google 🔍

  • Infrastructure: Best (massive server farms, custom chips)
  • Model: Good
  • Partner: None
  • Data: Best (YouTube, search data, books, etc.)
  • Distribution: Android devices, Search, Google Cloud Platform
  • Core Business: Chatbots are disruptive to Search
  • Scarcity Risk: Data feedback loops diminished
  • Disruptive/Sustaining: Disruptive
  • New Business Potential: Cloud

What this means: Google has “everything” but their biggest challenge is that AI chatbots could kill their search ad business (their money printer). But they're adapting well with AI search overviews and their cloud business is growing. Their massive data advantage from YouTube and search gives them unique strengths.

Meta 👥

  • Infrastructure: Good
  • Model: OK (Llama is falling behind)
  • Partner: None
  • Data: Good
  • Distribution: Meta apps, Quest devices
  • Core Business: AI delivers more individualized content and ads
  • Scarcity Risk: Attention diverted to chatbots
  • Disruptive/Sustaining: Sustaining
  • New Business Potential: Virtual worlds and generative UIs

What this means: Zuckerberg is panicking - offering hundreds of millions to top AI talent. Why? If people spend hours chatting with AI instead of scrolling Instagram, Meta loses. But AI could also supercharge their ad business and finally make VR/AR useful. They need to catch up fast or risk becoming the next MySpace.

Microsoft 💻

  • Infrastructure: Very Good
  • Model: None (relies on OpenAI)
  • Partner: OpenAI (but relationship is rocky)
  • Data: Good
  • Distribution: Windows, Microsoft 365, Azure
  • Core Business: AI drives Azure usage
  • Scarcity Risk: Access to leading edge models
  • Disruptive/Sustaining: Sustaining for Azure, potentially disruptive for Microsoft 365
  • New Business Potential: Agents

What this means: Microsoft looked unbeatable in 2023 with their OpenAI deal. Now? OpenAI is threatening antitrust complaints and the relationship expires in 2030. Microsoft needs backup plans (like investing in xAI or Mistral). The good news: their cloud business is booming from AI demand. The risk: if AI replaces knowledge workers, who needs Microsoft 365?

Amazon 📦

  • Infrastructure: Good
  • Model: Poor (no standout AI)
  • Partner: Anthropic Claude
  • Data: Good
  • Distribution: AWS, Alexa
  • Core Business: AI drives AWS usage
  • Scarcity Risk: Access to leading edge models, chip competitiveness
  • Disruptive/Sustaining: Sustaining for AWS
  • New Business Potential: Agents, Affiliate revenue from AI recommendations

What this means: Amazon is the dark horse winner. They don't need their own AI because everyone building AI uses AWS (their cloud service). Their Anthropic partnership is more stable than Microsoft-OpenAI. Plus, AI actually helps their business - more AI = more cloud usage = more money.

The Real Takeaway

Bill Gates famously said,

"We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten."

ChatGPT didn't kill Google Search, and Apple is still selling iPhones just fine despite being "late" to AI.

But the ten-year underestimation is starting to show: Meta's desperate hiring spree suggests they see AI eventually replacing social media entirely, Amazon's "boring" infrastructure play might make them the biggest winner (just like with cloud computing), and Apple's focus on hardware could position them perfectly for the AI-powered robotics coming down the road.

We're only two years into this revolution - the real disruption that changes how we live and work is still coming, and today's leaders might not be tomorrow's winners. Stay curious, but don't panic. The future will be bigger than we imagine - it'll just take longer to arrive.

Whenever you're ready, here are some other ways I can help you:

✅ Build Your Command Center in Notion: Create a single-dashboard system and organize everything in one place. Save yourself 5 years of trial-and-error.

💻 The Workspace Academy: Learn the workflow I taught to over 10,000 Googlers. Made for busy professionals using Google Workspace.