When a company sells you a device below cost, it isn't being generous — it's making a bet that the ecosystem surrounding that device will pay back the difference many times over. That trade-off is platform strategy in its most visible form.

Why this matters now

Platform strategy sits at the center of nearly every major technology business decision: which AI model to open-source, which app store to lock down, which API to price at zero. Understanding how subsidies, openness, and revenue capture interact helps you read any platform's long-term intentions — not just gaming hardware, but cloud services, developer tools, and AI infrastructure. When a major hardware maker publicly refuses the subsidy model and calls that refusal a core belief, it forces a clean look at what subsidies actually buy.

How it works

A platform business has two levers: the entry point (hardware, software, or service that pulls users in) and the ecosystem (the monetizable activity that happens once users are inside). The classic approach is to discount the entry point aggressively — sell hardware below cost — and then recoup through revenue cuts on every transaction, mandatory subscription fees, and rules about what third parties can build or sell inside the platform.

@title Platform subsidy cycle
  Discounted entry point ·········
     │
     ▼
  Users locked into ecosystem ····
     │
     ▼
  Revenue capture ················
     │  higher cuts, fees, exclusivity
     ▼
  Subsidy repaid with margin ·····
@caption Entry-point discounts fund ecosystem lock-in, which recaptures margin through platform fees.

The mechanism is self-reinforcing: lock-in justifies the upfront loss, and the upfront loss deepens lock-in. The moment a platform accepts a subsidy, it has implicitly committed to recovering that investment through tighter control — there is no other place the money comes from.

An open platform breaks this cycle by refusing the subsidy. Hardware or software is priced at or near actual cost. Because there is no subsidy to recover, there is no structural pressure to raise platform fees, enforce exclusivity, or restrict what runs on the device. Openness isn't a marketing claim here — it's an economic consequence of how the entry point was priced.

Real-world applications

The same logic plays out far beyond gaming hardware.

Cloud AI APIs — When a cloud provider offers a frontier model at steeply discounted inference rates, the question worth asking is: what does lock-in look like here? Proprietary SDKs, data residency requirements, and bundled services are the equivalent of the walled storefront.

Developer tools — An IDE or framework offered free to developers can be a genuine open platform or a funnel into a paid deployment target. The pricing of the downstream runtime reveals which one it is.

Open-source models — Releasing model weights publicly is structurally similar to pricing hardware at cost: it removes the subsidy mechanism and with it the leverage to enforce ecosystem control. Companies that open-source base models while monetizing fine-tuning infrastructure or inference are making an explicit platform strategy choice about where they want capture to happen.

In every case, the right diagnostic question is the same: where does this company need to recover its investment, and what control does recovering it require? Pricing is almost always the clearest answer.

Where to go deeper

  • Platform competition literature — Search for foundational work on two-sided markets and network effects; the economics of subsidizing one side of a marketplace to monetize the other is well-documented in academic and practitioner writing.
  • Open-core business models — Study how developer-facing companies decide which layers to open and which to monetize; the pattern maps directly onto hardware platform decisions.
  • EducationPals courses on AI infrastructure and LLM deployment — Courses covering API design, model serving, and cloud strategy will apply this same platform logic to the AI stack you're likely building on or selling into.