The Five Companies That Control Your AI Future (And You’ve Never Even Heard of One of Them)
You’ve been hearing about AI nonstop for two years. AI underwriting. AI lead generation. AI marketing tools. AI this, AI that. And you’ve probably started using some of it — maybe even depending on it.
But here’s something almost nobody in the mortgage world is talking about: a tiny group of five tech companies now controls the majority of all the AI computing power on Earth. Not most of it. Not a big chunk. We’re talking about two out of every three computers running AI — globally. And that number is growing fast.
Before we get into what that means for you as a loan officer or real estate agent, let’s back up and answer the question you might be asking right now.
So, What Exactly Is a Hyperscaler?
A hyperscaler is a company that runs massive, planet-sized computing infrastructure — data centers so enormous they can power entire cities’ worth of computing demand on demand. Think of them like the electric utility companies of the internet age, except instead of running power lines, they run the servers, chips, and networks that everything digital depends on.
The five hyperscalers we’re talking about are Google, Microsoft, Meta (Facebook’s parent company), Amazon, and Oracle. You know four of those names well. Oracle might surprise you — but in the data center world, they’ve quietly become one of the biggest players on the planet.
Together, these five companies now own 67% of all global AI compute capacity, according to new data from Epoch AI, a research institute that tracks chip ownership worldwide. That’s up from roughly 60% just one year ago. The trend is moving in one direction only — toward more concentration, not less.
Why Does This Matter to Anyone Who Uses AI Tools?
Here’s the real-world connection. Every AI tool you use — whether it’s a chatbot on your loan origination system, an automated marketing platform, or a large language model you use for writing client emails — almost certainly runs on computing power rented from one of these five companies.
Even the biggest, most famous AI companies in the world — OpenAI (makers of ChatGPT) and Anthropic (makers of Claude) — don’t actually own the computers that run their AI. They rent it. From these five hyperscalers.
That means the entire AI industry, from the smallest startup to the most powerful chatbot on the planet, is almost entirely dependent on a handful of corporations to keep the lights on.
This Has Happened Before — And It Didn’t End Well for Small Players
Think about what happened when social media consolidated. In the early days, there were dozens of platforms. Then there were a few. Then effectively two or three controlled almost all the eyeballs. Suddenly, reaching your customers meant paying for ads — on their terms, at their prices, with their rules.
The same pattern appears to be forming in AI infrastructure. When a few companies control the pipes that all AI runs through, they have extraordinary leverage over who gets access, at what price, and under what conditions.
For large banks and national lenders with deep pockets, this might not matter much. They can afford whatever the hyperscalers charge. But for independent mortgage brokers, community lenders, and small real estate firms, the concentration of AI compute could eventually mean higher costs and fewer choices for the AI tools you rely on.
The Numbers Are Staggering
To put the scale of this in perspective: Google alone holds the equivalent of about 5 million Nvidia H100 GPUs in computing capacity. That’s roughly 25% of all the AI compute power on Earth — sitting inside one company’s data centers. Microsoft, Amazon, Meta, and Oracle own much of the rest.
The top three hyperscalers — Microsoft, Google, and Amazon — plan to spend more than $500 billion on new AI infrastructure in 2026 alone. That’s not a typo. Half a trillion dollars. In one year. On data centers, chips, and the power grids to run them.
This level of investment creates enormous moats. It gets harder and harder for anyone else to compete at that scale. And when competition shrinks, prices tend to rise and innovation tends to slow — at least for the people at the bottom of the food chain.
What Should You Do With This Information Right Now?
For now, keep using the AI tools that are making your business more efficient — there’s no reason to pump the brakes on that. But start paying attention to who controls the AI platforms you depend on. Read the terms. Understand the pricing structure. And watch for consolidation happening above you in the tech stack.
The AI revolution is real, and it’s going to reshape mortgage and real estate in ways we’re still figuring out. But like any powerful technology, it matters enormously who owns the infrastructure. Right now, five companies do.
And that’s a story worth watching very closely.
If you want to stay ahead of the trends reshaping mortgage and real estate, don’t miss a single post — subscribe to Well That Makes Sense today. Fair warning: we make this stuff way more interesting than it sounds. (That’s a low bar, and we clear it every time.)