Recommended Articles
Fundraising

Lumaril Raises $66M Series A at $400M Valuation to Digitize Factory-Floor Knowledge

By Angeled · November 2, 2026 · Fundraising
Lumaril Raises $66M Series A at $400M Valuation to Digitize Factory-Floor Knowledge
Lumaril Raises $66M Series A at $400M Valuation to Digitize Factory-Floor Knowledge

Lumaril, a New York-based industrial AI company, has raised a $66 million Series A funding round at a $400 million valuation, bringing its total funding to $78 million. The round was led by returning investor Elad Gil, with participation from Felicis, Koch Disruptive Technologies, Pax Ventures, Lux Capital, and General Catalyst. Lux Capital and General Catalyst had co-led the company's earlier seed round.

The company was founded by Ben Duffield and Mark Elliot, who met as colleagues at Palantir Technologies and both worked on Palantir Foundry, the data-integration platform used by large industrial and government customers. Lumaril's premise is that one of the biggest constraints on American manufacturing is not a shortage of factories, but the quiet loss of knowledge as experienced plant operators retire.

The knowledge walking out the door

Running a production line well is, in the words of coverage of the company, closer to an art than a checklist. A veteran operator knows how a particular machine sounds when it is about to drift out of tolerance, how a change in humidity affects a batch, and which dial to nudge when a raw material shipment arrives slightly off specification. That knowledge is rarely written down. It lives in people — and when those people retire, it leaves with them.

This is known in manufacturing as the loss of "tacit knowledge": expertise that is learned by doing and is difficult to formalize. It is a growing problem as the skilled manufacturing workforce ages. Factories can buy new equipment, but they cannot easily buy back decades of judgment about how a specific plant behaves under real conditions.

Lumaril's answer is to capture that knowledge systematically. Its platform connects to a plant's existing control systems — the industrial computers and sensors that already run the facility — and builds what the company calls a living, continuous model of the entire operation. The model learns from how the plant's best operators run it, then recommends adjustments in real time as materials, weather, or equipment conditions change. Over time, the system is designed to become partially or fully autonomous, making routine adjustments itself while human operators supervise.

A different business model: selling production, not software

Lumaril's commercial model is as notable as its technology. The founders have said plainly that they do not sell software — they sell production, and they are paid on the throughput and yield they add. In other words, the company's revenue is tied to measurable improvements in what a plant produces, rather than to a flat subscription fee.

That alignment matters in an industry famous for skepticism toward new software. A vendor paid on results has a stronger incentive to make sure the system actually works on the factory floor, not just in a demo. It also changes the sales conversation: instead of asking a customer to bet on a tool, Lumaril asks to share in the gains it creates.

Why industrial AI is drawing this kind of capital

A $66 million Series A is exceptionally large by historical standards — Dealroom noted it sits in the 97th percentile of all-time Series A rounds for US enterprise software startups. The size reflects both the ambition of the bet and the cost of proving it: industrial software must be validated in live plants, integrated with legacy control systems, and trusted with production lines where a mistake costs real money.

It also reflects a broader shift. For years, venture capital favored software that could be deployed with a credit card and scaled without touching the physical world. Industrial AI is the opposite: slow to deploy, hard to build, and deeply embedded in customer operations. But the prize is correspondingly large. Manufacturing, chemicals, agriculture, energy, and other physical industries make up a huge share of the global economy, and even small percentage gains in throughput or yield translate into very large sums.

For founders, Lumaril's story carries a practical lesson about wedge and credibility. The company's founders did not arrive cold: their Palantir Foundry background gave them both the technical toolkit and the credibility to be trusted inside industrial operations. And rather than promising to replace existing systems — a "rip and replace" pitch that plant managers resist — Lumaril plugs into what a factory already runs.

For investors, the round is a signal that the industrial autonomy race is now a funded category, not a thesis. The question going forward is not whether AI will enter the factory, but which layer of the stack captures the value: the control systems, the models, or the companies that tie payment to performance.

Angeled.com connects founders with investors.

The Wall Street Journal: "AI Startup Lumaril Wants to Make Factories the Best Version of Themselves"
Pulse 2: "Lumaril Raises $66 Million Series A At $400 Million Valuation To Expand Industrial AI Platform"
Dealroom News: "Lumaril raises $66M Series A to run factories on autopilot"
Lumaril company announcement: "Nothing is built from nothing"

Read Next