Is OpenRelay profitable?

No revenue, ARR, or customer figures have been disclosed. OpenRelay is a 2-person, pre-seed Y Combinator Summer 2026 startup founded in 2026, with no independent funding announced as of July 22, 2026, making profitability highly unlikely at this stage.

Who are OpenRelay's biggest competitors?

Baseten (raised $1.5B Series F at a ~$13B valuation, June 22, 2026, spanning 18 clouds and 87 clusters), GPU marketplaces Vast.ai and RunPod, and ZML, which open-sourced a free hardware-agnostic inference engine on July 8, 2026, matching OpenRelay's core pitch.

Could open-source tools replace OpenRelay?

Yes, potentially. On July 8, 2026, Paris-based ZML released ZML/LLMD v2, a free, Yann LeCun-endorsed, hardware-agnostic inference engine offering the same 'run any model anywhere' proposition OpenRelay sells, directly undercutting its differentiation with a zero-cost alternative.

How much does OpenRelay charge for GPU compute?

OpenRelay advertises $0.98/hr for RTX 5090 and A100 32GB tiers and $2.30/hr for H100 80GB, claiming savings of up to 90% versus AWS, Azure, and GCP; these are the company's own vendor-claimed figures as of July 22, 2026.

OpenRelay is privately held, based in Seattle, WA, founded in 2026, and part of Y Combinator's Summer 2026 (S26) batch with 'Active' status. No parent company, acquisition, or independent funding round has been publicly disclosed as of July 22, 2026.

Has OpenRelay raised a seed round?

As of July 22, 2026, no independent, non-YC funding announcement has been found for OpenRelay; only its participation in Y Combinator's Summer 2026 batch is confirmed, with no disclosed valuation or round size.

YC Summer 2026 · United States

OpenRelay

2-person YC startup routing AI inference across cheap, hardware-agnostic GPUs.

Little protection left

Assessed 27 July 2026 · Evidence as of 27 July 2026

Not much stands between this business and its pressures.

On the recordSince 27 July 2026 · Checkable from Q4 2026-Q1 2027 · Every call on the record Checkpoint: OpenRelay's post-Demo Day seed fundraising outcome

The call

OpenRelay sells cheap, compatible-API GPU routing into a room where Baseten already scaled and ZML gives the same idea away free.

OpenRelay helps companies run AI models on cheaper computers by directing the work to the best available hardware, charging 90% less than big cloud companies. The problem: a free competitor called ZML launched the same idea in July 2026 with backing from a major AI researcher, and bigger companies like Baseten already handle this for customers across multiple cloud services. What might save OpenRelay is if it discovers a specific type of customer or use case that the bigger players ignore, but right now it is competing purely on price in a race where larger companies with more money usually win.

Why

AI pressure: ZML (open-source hardware-agnostic inference engine). ZML gives away the same hardware-agnostic routing OpenRelay charges for, backed by Scaleway and VSORA hardware partners. Could AI do the core job? Yes, likely.

Bigger players: Baseten. Baseten already runs multi-cloud inference across 18 clouds and 87 clusters, fresh off a $1.5B raise at a $13B valuation. Could a giant copy it? Yes, easily.

What protects it. Nothing proven yet; the only unclaimed opening is regulated inference like HIPAA or FedRAMP where hyperscalers can't reach.

What would change this call

OpenRelay announces a named seed round over $5M by mid-2027. Would show investors see a defensible niche beyond commodity GPU arbitrage.
ZML's free engine gains adoption from named enterprise customers beyond Scaleway/VSORA. Would show the free open-source alternative is closing OpenRelay's cost-arbitrage wedge.
OpenRelay publishes a named customer or vertical focus (e.g. HIPAA, FedRAMP, EU data residency). Would show a pivot to a regulated niche the incumbents can't easily enter.

Key numbers

Baseten Series F: $1.5B at ~$13B valuation (2026-06-22)
Baseten spans 18 clouds, 87 clusters (2026-06-22)
OpenRelay team: 2 people (2026-07-22)

Named risks

No shipped moat; OpenAI/Anthropic-compatible API lowers switching costs for customers to leave. Severity: high
Free open-source ZML engine (July 2026) directly replicates core hardware-agnostic pitch at zero cost. Severity: high
Only 2-person team with no disclosed independent funding or named customers as of July 2026. Severity: high
Commodity margin-tax race against well-capitalized incumbents like Baseten and Vast.ai. Severity: high
Industry-wide B200/Blackwell GPU supply constraints could limit hardware sourcing. Severity: medium

Rivals, as read in July 2026

Rival	Direction	Evidence
Baseten	losing to	Raised $1.5B Series F at ~$13B valuation with 18 clouds/87 clusters as of June 22, 2026
ZML	losing to	Released free open-source hardware-agnostic inference engine July 8, 2026, matching OpenRelay's pitch
Vast.ai	losing to	Established GPU marketplace with five-year head start, offering AMD support since May 2024
RunPod	contesting	Operates competing GPU rental/serverless marketplace as of July 15, 2026
Modal Labs	losing to	Reportedly in talks near $2.5B valuation as of July 8, 2026

Questions people ask

Is OpenRelay profitable?: No revenue, ARR, or customer figures have been disclosed. OpenRelay is a 2-person, pre-seed Y Combinator Summer 2026 startup founded in 2026, with no independent funding announced as of July 22, 2026, making profitability highly unlikely at this stage.
Who are OpenRelay's biggest competitors?: Baseten (raised $1.5B Series F at a ~$13B valuation, June 22, 2026, spanning 18 clouds and 87 clusters), GPU marketplaces Vast.ai and RunPod, and ZML, which open-sourced a free hardware-agnostic inference engine on July 8, 2026, matching OpenRelay's core pitch.
Could open-source tools replace OpenRelay?: Yes, potentially. On July 8, 2026, Paris-based ZML released ZML/LLMD v2, a free, Yann LeCun-endorsed, hardware-agnostic inference engine offering the same 'run any model anywhere' proposition OpenRelay sells, directly undercutting its differentiation with a zero-cost alternative.
How much does OpenRelay charge for GPU compute?: OpenRelay advertises $0.98/hr for RTX 5090 and A100 32GB tiers and $2.30/hr for H100 80GB, claiming savings of up to 90% versus AWS, Azure, and GCP; these are the company's own vendor-claimed figures as of July 22, 2026.
Who owns OpenRelay?: OpenRelay is privately held, based in Seattle, WA, founded in 2026, and part of Y Combinator's Summer 2026 (S26) batch with 'Active' status. No parent company, acquisition, or independent funding round has been publicly disclosed as of July 22, 2026.
Has OpenRelay raised a seed round?: As of July 22, 2026, no independent, non-YC funding announcement has been found for OpenRelay; only its participation in Y Combinator's Summer 2026 batch is confirmed, with no disclosed valuation or round size.

Timeline

2026-07-22YC lists OpenRelay as 2-person, Active S26 company with no open job postings
2026-07-08ZML releases free open-source hardware-agnostic inference engine ZML/LLMD v2
2026-06-22Baseten announces $1.5B Series F at roughly $13B valuation
2026-06-18TechCrunch reports Baseten raising $1.5B with ~20x YoY revenue growth
2026-04-12Industry-wide B200/Blackwell GPU supply constraints noted
2026OpenRelay founded in Seattle, WA, joins YC Summer 2026 batch
2024-05-01Vast.ai becomes first GPU rental marketplace to offer AMD support

Revision history

27 July 2026

First assessed. Little protection left. The text above is fixed as of this date and is not edited afterwards.

In the same band

Cerenovus Belvedir Hoplite