Manufacturing readiness levels, explained by someone who has watched a line stop

A founder once told me his product was basically done. He had a working unit. It had survived a pilot. He showed me a video of it running. Then I asked what it would take to build fifty of them, and the room went quiet. The unit had been assembled by the two engineers who designed it, over three weeks, with parts from four suppliers they knew personally. That is not a product. That is a very expensive prototype, and the distance between the two is measured in manufacturing readiness levels.

MRL is a scale from 1 to 10 that describes how ready your manufacturing is, separate from how ready your technology is. It came out of the US Department of Defense, which got tired of buying prototypes that could not be built twice. It is dry, bureaucratic, and one of the most useful honest mirrors a hardware startup can stand in front of.

Your technology level and your manufacturing level are different numbers

Technology readiness (TRL) asks: does the science work. Manufacturing readiness asks: can you build it, repeatedly, at a cost and quality someone will pay for. Deeptech companies routinely sit at TRL 8 and MRL 4 at the same time. The technology is proven in a relevant environment, and production is still two people with a torque wrench and good intentions.

Investors and customers increasingly know this. When a serious industrial buyer or a deep tech fund asks about your MRL, they are not being pedantic. They are asking whether the number on your quote is real.

The levels in plain language

The official definitions read like a procurement manual, so here is the working translation. Levels 1 to 3: you have research and a concept. Level 4: you can build it in a lab, by hand, by the people who invented it. Levels 5 to 6: you can build it in a production-relevant environment, meaning roughly the real processes and roughly the real materials, at prototype quantities. Level 7: you can build it in a production-representative environment, a pilot line, at low rates. Level 8: pilot line proven, ready for low-rate production. Level 9: low-rate production running, quality stable. Level 10: full-rate production, lean and boring.

The single most common self-deception is claiming level 7 while standing at level 5. The test is simple: if your two lead engineers got hit by a bus, could someone else build the next ten units. If the answer involves the phrase well, they would need to talk to, you are not at 7.

The money cliff sits between 6 and 8

Moving from lab builds to a pilot line is where the cash goes. Tooling, test equipment, process documentation, quality systems, a supply chain that can deliver the same part twice. This is the climb from MRL 6 to MRL 8, and it routinely costs more than everything before it combined.

This is the real valley of death for hardware, and it is commercial as much as technical. Nobody funds a pilot line on a promise. They fund it against purchase orders, or at least signed commitments. Which is why the companies that cross it are the ones that started selling before the product was finished: the LOIs, the design partners, the paid pilots. The manufacturing plan and the first customer are the same project wearing different clothes.

Your first customer is a manufacturing audit

Here is what surprises founders: the first serious industrial customer does not only buy your product. They audit your ability to make it. Before the purchase order, there is often a factory visit, a quality questionnaire, sometimes a full supplier qualification. They want to see process control, traceability, and what happens when a unit fails a test.

This is MRL as a sales asset. A startup at MRL 6 with honest documentation and a credible path to 8 beats a startup claiming 9 with nothing on paper. Industrial buyers have been burned by optimistic hardware companies before. Showing them you know exactly where you stand is itself a trust signal.

Climb the levels in the order that keeps you alive

The mistake is trying to buy your way to MRL 9 before the revenue exists. The smarter path is staged: use low-volume production methods that are expensive per unit but cheap to set up, price the early units accordingly, and let the value-based price of the problem you solve carry the cost of immature manufacturing. Customers pay for the outcome, and a €40k hand-built unit that saves €500k a year is still a bargain.

Then invest in the next level only when the order book justifies it. Each MRL climb should be funded by evidence of demand, not by hope. Manufacturing readiness is not a race to 10. It is a sequence of bets, each sized to what you have already sold.

What to do this week

Write down your honest MRL, using the bus test. Then write down what the next level costs, in money and months, and which customer commitment would pay for it. If those two numbers do not connect, that gap is your real roadmap, and closing it is partly a sales job. If you want help connecting the manufacturing plan to a pipeline that can fund it, that is the work we do at SLSCRW.