Talk Around Town: A Latvian Startup Claims Its AI Graph Platform Could Rewire Hardware Engineering

What Folks Are Hollerin' About

Well, butter my biscuit and call me impressed — there's a seed-stage startup out of Riga, Latvia, making some pretty ambitious noise. Trace.Space, founded in 2022 by Janis Vavere and Karlis Broders, says it has rolled out a product called Trinity aimed at robotics, aerospace, automotive, and defence engineering teams who are trying to claw their way from prototype to mass production, according to Tech.eu and Dealroom summarizing the company's own announcement.

The company describes Trinity as a graph-based system that, in Trace.Space's own telling, pulls together requirements, testing data, design parameters, and product variants into one interconnected structure — and then cracks the door open so AI agents can go rooting around in all that engineering data. Think of it like wiring every drawer in your workshop to a single smart pegboard that also talks back at ya.

What We Actually Know for Certain

Here's the part where the concrete pours pretty solid. Trace.Space was confirmed to have been founded in 2022 in Riga by Vavere and Broders, who reportedly spotted a market gap while previously building and selling software for complex hardware delivery projects, per Tech.eu. The company has offices in both Latvia and the United States, though investor databases disagree on exactly which U.S. city — Pitchbook and Tracxn say San Jose, Crunchbase says Miami, and Tech.eu describes Riga as the founding home with U.S. offices added later.

The money trail is about as verified as anything gets at this stage. Pitchbook and Tracxn both corroborate that Trace.Space has pulled in $5.5M in total funding across two seed rounds, including a $4M seed round that closed in early 2025, with Cherry Ventures and Fiedler Capital among the investors. That's the kind of financial data you can hang your hat on — unlike some of the flashier product claims, which ain't got the same independent legs to stand on.

What's Still as Murky as a Pond After a Rainstorm

Now here's where we pump the brakes harder than a mule digging in on a muddy slope. The company claims Trinity currently covers roughly 40% of the hardware product development lifecycle — up from about 10% when the platform first launched — with aspirations to hit 80%, according to Dealroom summarizing Trace.Space's own assertions. No external auditor, analyst, or independent researcher has weighed in on how that percentage was even calculated, and the 80% target is purely aspirational as the company describes it.

Trace.Space also says, per Dealroom and Tech.eu, that engineering teams at Lucid Motors, Serve Robotics, and Xiphos are among its customers. The company further claims that former SpaceX employees told it Trinity resembles how they handled engineering data internally at SpaceX. Those are mighty impressive names to drop at a barn dance, but neither claim has been independently confirmed — no customer on record has publicly vouched for the platform through a third-party channel.

Then there's the Space Agent. Trace.Space describes it on its own website as an orchestration layer for what the company calls a coordinated set of specialized hardware and systems engineering agents that operate inside the graph, working across traceability, variants, coverage modeling, and requirement structure. That's a mouthful of buzzwords that sounds real pretty, but it remains entirely the company's own characterization with zero external benchmarking to prop it up. The company's own marketing also claims AI agents let teams build physical products ten times faster — a figure with no independent study or benchmark cited anywhere in the packet.

Our Analysis: A Promising Piglet, But Let's See It Grow

This is analysis, not reporting: Trace.Space is operating in a genuinely underserved corner of the software world. Hardware engineering lifecycle tooling has historically been fragmented enough to make a junk drawer look organized, and the idea of threading requirements, test data, design parameters, and product variants through a single graph — then exposing all of it to AI agents — is conceptually coherent and interesting. The strategic direction the company describes is at least pointing toward a real problem.

That said, also as analysis: the evidentiary picture here is about as thin as a paper napkin in a thunderstorm. The coverage of this launch flows almost entirely from a single journalist interview with the CEO, mirrored by an investor-data aggregator. No independent analyst has poked at the product, no customer has gone on record outside of company-cited references, and every headline performance number — the lifecycle coverage percentages, the ten-times-faster marketing claim — originates from Trace.Space itself. Seed-stage startups have every right to tout their vision, but readers ought to hold the specifics loosely until someone outside the company kicks the tires.

The Bottom Line From the Porch

Trace.Space has launched what it describes as a graph-based hardware engineering platform called Trinity, the company says, and has $5.5M in verified funding to pursue the vision. The ambitious lifecycle coverage figures, the customer name-drops, and the AI orchestration layer capabilities are all things the company asserts — none have been corroborated by independent parties. Worth watching, but worth watching with one eyebrow cocked until more outside voices start hollering in agreement.