MNT Research's modular laptops show a useful way to describe hardware risk before taking a customer's money: name who makes each subsystem, where final assembly happens, what can force a redesign, and which component controls the calendar. For the Reform Next, circuit boards come from PCBWay, aluminum body parts from JPR, hinges from a Shenzhen supplier, cells from EREMIT, and final assembly and testing happen at MNT's Berlin workshop. That map does not eliminate delays. It tells a buyer which delay is plausible and tells a founder where working capital and engineering attention can become trapped.
The product is modular because the supply chain is not stable
Lukas F. Hartmann founded MNT around open hardware computers whose schematics, mechanical files, and software can be inspected and modified. The laptops use replaceable processor modules rather than soldering every major computing component into one permanent board. Users gain repair and upgrade options, but the company gains something just as valuable: a path around component obsolescence.
That does not mean modules are easy to source. In a 2023 interview about Pocket Reform, Hartmann said a processor module from Laird had a minimum 24-week lead time. He described using that waiting period to finalize aluminum parts and test another beta build. The dependency stayed on the critical path, but parallel work kept the calendar from becoming pure waiting.
This is the practical value of a modular architecture for a small manufacturer. When a processor becomes unavailable or poorly supported, the company can design an adapter or a new module instead of discarding the display, keyboard, enclosure, battery system, and customer documentation. Modularity does not remove redesign. It shrinks the part of the product that must change.
A public bill of responsibilities
The Reform Next campaign publishes an unusually concrete production plan. Custom keycaps were already made. PCBWay handles keyboard-switch soldering. JPR machines, bead-blasts, and anodizes the aluminum body. Smooth Technology supplies hinges. MNT sources display panels and cables from several suppliers, LiFePO4 cells from EREMIT, and packing material locally. The Berlin team performs final assembly, quality assurance, and packaging, then sends completed units to Mouser for fulfillment.
That list does two jobs. For customers, it distinguishes an engineered manufacturing plan from a render and a delivery date. For the company, it defines the handoffs that need incoming inspection. A hinge can be dimensionally wrong while the electronics are perfect. A finished circuit board can pass electrical testing but fail to align with the case. The more suppliers involved, the more important the interface drawings and acceptance checks become.
MNT also states the central risk: the Reform Next contains hundreds of electronic components. A shortage or price increase can force a PCB change, and the replacement then needs engineering and validation. That is more credible than claiming experience makes the schedule safe. Experience helps a team respond to surprises; it does not stop suppliers from creating them.
The calendar is a network, not a line
Use a simple critical-path exercise. Suppose the processor module requires 24 weeks. Enclosures require 10 weeks, boards eight, packaging four, and final assembly three. If all orders can begin together and the product is already validated, the nominal duration is not 49 weeks. It is roughly the longest supply path plus final assembly:
max(24, 10, 8, 4) + 3 = 27 weeks.
Now add a board redesign that can start only after the processor module arrives, taking four engineering weeks and another eight-week board order. The path becomes:
24 + 4 + 8 + 3 = 39 weeks.
This illustrative arithmetic explains why a one-part substitution can move delivery by months. Founders should draw dependencies with arrows, not add every supplier quote into a single total. The question is which work must wait for which evidence.
MNT's April 2026 update makes the consequences visible. The company said the Reform Next was production-ready and that it was ordering remaining parts for an initial 90-unit Crowd Supply shipment. The same update listed other Reform orders waiting six to eight months for RK3588 modules. The table separates items in assembly, items in quality assurance, and items blocked by a component. That vocabulary is useful because "in production" can otherwise mean almost anything.
Build a supplier map before accepting preorders
A small hardware team can reproduce the useful part of MNT's disclosure on one sheet. Give every critical item six fields:
- named supplier and manufacturing location;
- quoted and worst observed lead time;
- minimum order and payment date;
- substitute part and redesign time;
- incoming inspection method;
- the product milestone this item blocks.
Color only the items with no qualified substitute. Those are the risks worth discussing in a campaign, not a generic paragraph about global supply chains.
Then add a cash column. If 90 computers require a hypothetical $600 of paid components each, the batch ties up $54,000 before assembly labor and fulfillment. A 24-week part purchased first and a case purchased later have different cash exposure even if both end up in the same unit. Preorders can finance that gap, but the money is not revenue available for unrelated hiring. It is a liability attached to a future machine.
Where openness helps, and where it does not
Published designs let customers repair a board, let contributors inspect choices, and let MNT move modules without disguising proprietary connectors. Openness can also widen the pool of people who catch errors. It does not qualify a substitute component, negotiate a refund, or pay for a repeated CNC batch.
That boundary matters for every founder working with emerging hardware. A public repository is a design asset, while manufacturing competence lives in tolerances, supplier relationships, fixtures, test procedures, and recorded failure modes. Those operational assets resemble the durable processes in a small, owner-operated company, but they cannot be copied with a download. A supplier dependency audit makes the missing substitutes visible.
Within an hour, turn the current bill of materials into the six-field supplier map and mark the longest no-substitute item. If the promised ship date does not include its worst observed lead time, change the promise before taking more orders. MNT's example is not that transparent updates prevent frustration. It is that a buyer can understand whether a delay comes from assembly capacity, quality work, or a missing processor, and a founder can manage those as three different problems.
Updates should preserve that separation. Report units waiting for parts, units assembled, units in test, units rejected, and units handed to fulfillment. A percentage such as "90 percent complete" combines unlike work and often moves backward when defects appear. Counts let a backer see flow and let the team locate the bottleneck. They also prevent a founder from treating inventory sitting beside an unfinished fixture as nearly delivered revenue. In hardware, a unit becomes finished only after the defined acceptance test passes and responsibility moves to the next named stage.
A manufacturing schedule becomes trustworthy when every promised unit has a named next stage and a test it must pass.
Discussion
Sign in with Google or just a name. No email link, no password to remember.