Orbital Industries
Orbital Industries is an AI‑first industrial company that has been quietly moving from materials discovery into real-world cooling hardware for data centres. Its most visible product is a PFAS‑free GPU coolant — a concrete, saleable outcome of machine‑led materials design — and the company has been collecting credible signals of validation: a multi‑year partnership with AWS announced in December 2024 and a $50M Series B led by Plural on May 28, 2026. PitchBook’s record that the startup has raised roughly $61M in total suggests there are earlier, smaller rounds beneath the headline number. That combination — hyperscaler engagement and institutional capital — is exactly the sort of external proof that gets operators to raise an eyebrow in this space.
What they do
Orbital sits in an awkwardly interesting middle ground. At one end are pure materials‑discovery platforms that run simulations and propose new chemistries; at the other are vertically integrated manufacturers that design, build and ship finished infrastructure. Orbital’s thesis is simple: use AI research and analysis of public and proprietary data to design advanced materials, then productize those discoveries into manufactured hardware you can bolt into a rack. The PFAS‑free GPU coolant is the clearest expression of that approach — it is not a paper or a model, it is a consumable engineered into a deployed cooling system.
That product orientation matters. Materials discovery without a path to manufacturing has a commercialization gap; conversely, hardware makers without differentiated materials are often stuck competing on cost or commoditized engineering. Orbital’s claim to an edge is marrying AI‑led materials design to an actual supply chain and product roadmap. If the company can reliably move from model output to repeatable production, it could capture value at a point many discovery platforms do not reach.
The market
Picking a market definition here matters and also exposes how thin publicly available signal can be. A defensible top‑line is the global AI hardware market: Grand View Research pegs that at roughly USD 86.8B in 2024. Other firms produce substantially different figures — some as low as USD 27.9B — depending on whether the report measures processors, servers, or end‑to‑end vendor revenue. Orbital’s addressable slice is therefore highly dependent on how you define “AI hardware” and how much of cooling and materials you count as a share of that spend.
More crucially, the company has not published unit pricing, ACV or a bottom‑up SAM/SOM model. That absence makes it impossible from public sources to ground a revenue projection or margin trajectory. For investors and customers alike, the compelling macro opportunity is real, but the economic story for Orbital specifically hinges on disclosed unit economics and the cost structure of manufacturing and logistics.
The competitive picture
Orbital’s contemporaries fall into two camps. On the upstream side are materials‑discovery specialists — names like CuspAI or Periodic Labs — that excel at simulation and proposing novel compounds but often stop shy of factory floor implementation. On the downstream side are vertically integrated infrastructure manufacturers such as Prometheus, which orchestrate component sourcing, fabrication and large‑scale deployments but may lack the bespoke materials advantage.
Orbital’s attempt to wedge itself between these camps is deliberate: offer materials differentiated enough to matter to hyperscalers, while owning enough of the manufacturing pathway to monetize. That positioning is attractive on paper but operationally unforgiving. Turning molecules into millions of liters of coolant, qualifying those materials across standards, and building supply chains that meet hyperscaler SLAs are all nontrivial. Success requires heavy engineering, regulatory work and manufacturing scale — capabilities that are expensive and slow to assemble.
Momentum & what to watch
The company’s momentum is tangible. The December 2024 multi‑year AWS partnership is a major commercial signal — hyperscaler engagement reduces a lot of go‑to‑market risk — and the Plural‑led $50M Series B announced in May 2026, with participation from funds listed in press coverage (Nventures/NVentures, Radical Ventures, Compound, Fly Ventures), shows investor conviction. PitchBook’s higher total‑raised figure (~$61M) suggests earlier financings that haven’t been widely reported.
But momentum and runway are necessary, not sufficient. Public disclosures list only CEO Jonathan Godwin by name, and that raises the single biggest tension: team depth. Manufacturing, supply‑chain orchestration and commercial rollouts are operationally intensive; a one‑person public roster is a gap in disclosure if not in reality. For anyone taking a first meeting — operator, procurement lead at a data centre, or prospective investor — the most important questions are straightforward and practical: where are the commercial deployments (who, when, scale), what are the unit economics and margins on coolant and system sales, and what is the manufacturing roadmap to move from pilot volumes to the hyperscaler scale implied by an AWS partnership?
There are also executional checkpoints to watch: test and qualification timelines for the coolant across different GPU platforms, regulatory approvals for materials and supply‑chain resilience for non‑PFAS components. Each of these is a gating item for volume contracts and for margins that can support a hardware company’s growth.
Closing take Orbital Industries is a useful case study in the hard step from algorithm to alloy to appliance. Its public signals — an AWS relationship and a sizable Series B — make it one of the more credible attempts to productize AI‑designed materials into real infrastructure. But the headline strengths are only half the story; the company’s trajectory will be decided in the trenches of manufacturing scale, commercial traction and team depth. For now, the right posture is cautious interest: the opportunity is large, the approach is sensible, and the proof will live in deployments and unit economics rather than press releases.
Compiled by AlgoTurk from public web sources. Not investment advice.