Mid-range cloud is dead

The two-pronged squeeze on the middle market

Focusing on live broadcast and scale, my job is to see around corners; to assess platform risk before it becomes a problem. Right now, I see a fatal strategic manoeuvre being executed by the Big Three cloud providers: Amazon Web Services, Google Cloud, and Microsoft Azure. They are not just out-competing the mid-tier players; they are systematically eradicating them.

This is not a slow market shift: it is a controlled collapse of the middle ground, driven by two simultaneous pressures. Mid-tier providers: the DigitalOceans, Linodes, and their ilk: are being squeezed out of existence because they cannot compete on cost at one end, nor on capability at the other. It is a strategic chokehold designed to force every enterprise into one of two camps: a hyper-commodity general workload or a proprietary, high-performance AI future.

The commodity race to zero

The first strategic pressure is a brutal undercut on general-purpose compute. The Big Three have effectively declared general virtual machines and basic storage to be loss leaders. They possess an operational scale and capital expenditure budget that no mid-tier provider can realistically match. For Amazon, Google, and Microsoft, the goal of offering a cheap VM is simply customer acquisition; it is a foot-in-the-door strategy to onboard an enterprise and then up-sell them to high-margin managed services, databases, and network features.

I have always championed efficiency and constraint in our platform decisions. For years, the mid-tier cloud offered a competitive balance: predictable pricing, a simplified interface, and just enough horsepower for scale-out workloads. That advantage is gone. Today, attempting to run a commodity IaaS business in the shadow of the hyperscalers is a financially unviable position. Mid-tier providers are being systematically gutted of their general workload customers who are migrating toward the marginal cost leader.

The silicon chokehold

The real strategic weapon, however, is not price: it is the silicon. For any product leader building truly high-growth, modern applications in 2024, the differentiating factor is no longer a general-purpose CPU: it is access to specialised AI hardware, primarily advanced Graphics Processing Units (GPUs) and Tensor Processing Units (TPUs).

The Big Three are not just buying up this silicon; they are securing proprietary supply chain access. They are either co-developing chips with manufacturers or designing their own custom architecture (like Google’s TPUs and AWS’s Trainium/Inferentia chips). This is the key strategic firewall. It is virtually impossible for a mid-tier competitor to offer a competitive deep learning or high-intensity data processing infrastructure because the supply of the necessary cutting-edge chips is simply not available to them at scale or at a viable cost. This hardware scarcity is an existential threat. It forces every product team with a serious AI ambition to go directly to one of the three hyperscalers, regardless of their commodity pricing.

Prepare for the 2025 vendor migration

For product leaders, the time for vendor optionality in the mid-market is over. This is not a choice you can defer until the next budget cycle.

2025 will be the year of forced migration and accelerated vendor lock-in. You will be forced to bifurcate your strategy: non-critical, scale-out workloads will inevitably move to the most aggressively priced commodity offering, and critical, differentiating AI workloads will be forced into a Big Three provider where the specialised silicon is the only game in town. The mid-range market, unable to compete on either cost or capability, will vanish; absorbed, acquired, or simply out of business.

Start reviewing your mid-tier contracts today. If your product roadmap relies on scale-out IaaS or future AI acceleration, you must prepare an emergency migration plan. The old strategic benefit of balancing vendor risk across a wide array of providers is being replaced by a new reality: a necessary choice between hyper-competitive low cost and proprietary, specialised AI lock-in. The market is choosing for you; your only remaining choice is when and how you execute your platform migration.