hyper: the first-party provider

Proposed first-party provider research with a strict neutrality gate. No deployment, date, model, capacity, price, or audit-reference commitment.

Owner

Supply

Version

proposal-2026.07

Verified

2026-07-13

Hyper is a Proposed first-party provider concept, not a deployment and not part of the 17 July mainnet Beta commitment. No Omnious-owned GPUs are serving traffic, and no launch date, model portfolio, capacity, price, or service level has been approved.

The hypothesis under review

A first-party open-model provider could help with cold-start liquidity, create a controlled reference for supervised model-integrity work, and test whether well-utilized self-hosting can compete with hosted APIs. Those are research objectives, not customer benefits available today. The opening Beta uses allowlisted external supply.

Neutrality requirements

If the proposal advances, Hyper must behave like an ordinary provider. It must sign the same short-lived quotes, enter the same second-score/3 auction, receive no ranking preference, expose the same measured service data, accept the same review controls, and reconcile under the same 7% platform fee. Its fills and price setters must appear on the same receipt and analytics surfaces as every other provider.

This is a publication gate, not an aspiration: if a first-party provider needs a scoring preference, audit exemption, capacity reservation, or undisclosed subsidy to win, the proposal is not ready to launch. Rewards, a vault, a token, or a related operator identity may not change auction ranking or manufacture an independent rival.

Research inputs are not terms

The design files explore rented and owned hardware, open-weight models, throughput assumptions, cache economics, and possible unit costs. Those figures are sensitivity inputs. They are not quotes, availability promises, capacity reservations, or service-level agreements. Every input must be refreshed and evidenced before a pilot decision.

What would count as success

The objective would be to bootstrap a contestable market and then lose fills whenever independent providers offer better effective cost or measured service. A first-party provider becoming the only liquid supplier would be a failed marketplace outcome, not a moat.