In the artificial intelligence industry, the dominant narrative has been that success depends on having the best model, the most advanced intelligence, the most capable chatbot. However, a deeper analysis reveals that some of the smartest companies are not competing to build the best model, but are collecting rent from those who do. SpaceX, Nvidia, and Apple have adopted strategies that allow them to win regardless of which AI model triumphs or fails. It is the modern equivalent of McDonald's transformation: from selling hamburgers to owning the land where they are sold.

SpaceX: The Landlord of Artificial Intelligence

SpaceX has replicated the business model that Ray Kroc applied to McDonald's. The movie The Founder shows how Kroc discovered that the real business was not hamburgers, but real estate. Similarly, SpaceX has built a massive computing infrastructure, Colossus, initially to train its own model Grok, but now rents it out to other AI labs. The company has signed multi-billion dollar agreements with startups like Reflection AI, which will pay $150 million per month to use Nvidia GB300 chips in Colossus 2, in Memphis. If the contract holds, SpaceX could generate around $6.3 billion. Additionally, Google has agreed to pay SpaceX approximately $920 million per month for 32 months, totaling nearly $30 billion. And xAI reached an agreement with Anthropic that could bring up to $45 billion in revenue to SpaceX. Thus, regardless of whether xAI, Reflection, Google, or Anthropic succeed or fail, SpaceX always wins.

This strategy turns Elon Musk into the Ray Kroc of AI. Just as McDonald's developed its ultra-fast production system to make its original location profitable and then used it to succeed as a landlord, SpaceX is using the Colossus infrastructure to be the 'landlord' of artificial intelligence.

Nvidia: The Mayor McCheese of AI

Nvidia has positioned itself as the indispensable hardware provider for AI. It supplies the GPUs and network infrastructure that all advanced AI labs need to train their models. Colossus, for example, started with 100,000 Nvidia H100 GPUs and has expanded to over 220,000 GPUs, including H100, H200, and Blackwell. The entire infrastructure runs on Nvidia's Spectrum-X Ethernet platform. Additionally, Nvidia has invested in companies that rent capacity on Colossus, such as Anthropic and Reflection AI. In this way, Nvidia collects a 'tax' on every node of the AI economy, earning hardware revenue regardless of who wins the model race.

Apple: The One Taking the Best Part

Apple's AI strategy is even more brilliant. Its three-tier routing system allows most requests (about 85%) to be processed on-device using small, efficient models. Approximately 12% is sent to Private Cloud Compute, its own server infrastructure with Apple-designed chips. Only the 3% most complex queries are routed to external partner models, such as Google's Gemini, for which Apple pays around $1 billion per year. This design avoids the enormous cost of training a frontier model. Apple benefits in several ways: Apple Intelligence drives device upgrades and higher-tier iCloud subscriptions. While others invest trillions in infrastructure, Apple collects billions without the massive investments of model developers.

Lessons for Businesses and IT Professionals

This paradigm shift has profound implications. Companies investing in building frontier models may be accumulating debt that will eventually need to be repaid. Meanwhile, those who own the infrastructure or control the hardware benefit without taking that risk. For IT professionals, this suggests that value lies not only in artificial intelligence itself, but in the assets that support it: data centers, GPUs, networks, and orchestration platforms. Strategies like Apple's, which keep data close and use small models, may be more sustainable, as discussed in Workday and AI. Additionally, context debt and the risks of relying on external models are key topics, as analyzed in Vibe Slop is the Symptom.

In summary, the AI industry is undergoing a transformation similar to McDonald's. The hamburgers (the models) are not what matters; what matters is the land (the infrastructure). And the most valuable land in the world is now in a data center complex in Memphis, filled with hundreds of thousands of GPUs. Elon Musk, like Ray Kroc, has understood that he is not in the artificial intelligence business: he is in the real estate business.

Original source: ComputerWorld. Analysis and adaptation by ForgeNEX.