Automated Testing for AI-Enabled Manufacturing: Why We Invested in PreScalar

Everyone is talking about physical AI these days but hardly anyone is talking about how that
physical AI is manufactured, which is why we are excited to announce our investment in
PreScalar as part of their $7M Seed round, led by Dynamo Ventures with participation from Blue
Wire Capital, Uplift Ventures and Antler. PreScalar builds automated firmware testing and
software deployment infrastructure for advanced manufacturing.
‍
The products coming off manufacturing lines today look nothing like they did a decade ago.
Electric vehicles, autonomous systems, industrial robots, defense platforms, and consumer
electronics all share one thing in common: they are as much software as they are hardware. A
modern vehicle contains upwards of 150 distinct software-defined components, each requiring
firmware to be provisioned, tested, validated, and traced before the vehicle leaves the line. And
as OEMs move away from the old model, where Tier-1 suppliers managed software rollouts on
a quarterly batch cycle, toward continuous, rolling software deployment directly on the assembly
line, the factory floor has become the critical bottleneck in getting software-defined products to
market.

We size the addressable market at over $40B globally across automotive, autonomous vehicles,
consumer electronics, aerospace and defense, and medical devices - and it remains largely
underserved. Most manufacturers today address this through internal engineering teams,
bespoke systems integrators, or point tools stitched together without a common data layer -
approaches that do not scale as software release cadences accelerate and product complexity
grows. When something goes wrong on the line, isolating whether the failure is the tester or the
device under test is slow and expensive. The cost of a line going down is not theoretical; for any
manufacturer running high-volume production, unplanned downtime translates directly into
"bricked" hardware products, lost output, rework costs, and delayed shipments.
Why PreScalar Wins
The founding team, Skyler, John, and Ben, spent their careers building bespoke versions of this
product at every company they worked at - Cruise, Waymo, Rivian, Google, Lucid, and Chrysler.
That is not a background that can be replicated quickly. The team's extensive experience across
legacy OEMs and Silicon Valley AVs means the team can walk into a legacy OEM procurement
conversation and a robotics startup factory floor on the same week and be credible in both
rooms.
‍
PreScalar's architecture reflects a product insight that sounds obvious in hindsight but that
nobody has built properly: manufacturing testing and software deployment are the same
workflow, and collapsing them into a single platform requires deploying hardware on the line.
Most manufacturers today run test systems, firmware deployment tools, and software release
infrastructure as entirely separate workflows managed by separate teams with no common data
layer. When something goes wrong, engineering teams end up parachuting into factories for
weeks at a time to debug issues that should have been caught earlier. PreScalar collapses this
into a single platform - and because it requires deploying hardware to do so, software-only
competitors cannot compete here. Once the platform is integrated with a customer's test scripts
and manufacturing execution systems, removing it becomes very costly. But the more durable
advantage is the dataset. Every firmware version on every unit, every test result, every
calibration event is captured in a single structured layer that compounds with production
volume. Over time this becomes a proprietary record of how software-defined hardware
behaves at scale that no competitor arriving later can replicate without first deploying hardware
on the line and waiting years for the data to accumulate.
‍
Why We Invested
‍
1. Firmware failures in production are expensive and surprisingly common, from OEM
assembly plants down through the Tier-1 supply chain. Most manufacturers know this
and still do not have a good solution.
‍
2. No generic platform has ever solved this well. The default today is a collection of point
tools, custom scripts, and bespoke integrators stitched together by engineering teams
who would rather be building products.

3. OEMs are pulling software development back in-house and deploying firmware directly
on the line, making the factory floor a critical bottleneck in software delivery in a way it
simply was not five years ago. The toolchain has not kept up.
We are excited to partner with PreScalar as they build the infrastructure layer that every
advanced electronics manufacturer will eventually need.

‍