A hardware product takes roughly 12 to 18 months from a defined concept to a sellable product if it is simple and non-connected, 18 to 30 months if it is connected or battery powered, and two years or more if it enters a regulated market. Those ranges assume an outsourced engineering engagement with a partner who has done it before. They are not promises, because the largest single variable is not how fast anyone works.
Thank you for reading this post, don't forget to subscribe!Most of a hardware calendar is spent waiting. Waiting for a mold to be cut, waiting for a test chamber to come free, waiting for a component with a 40 week lead time, waiting for a container to leave a port. The engineering is the part you are paying for. The queue is the part that sets your launch date.
The short answer: realistic ranges from idea to sellable product
12 to 18 months for a simple electronic product
A mains or battery powered product with no radio, a straightforward enclosure and no regulated claim. Certification is a self-declaration path rather than a full grant. Tooling is one or two molds. Most of the schedule is engineering plus one tooling cycle plus one certification pass.
18 to 30 months for a connected or complex product
Add a radio and the calendar changes shape. Certification moves from a two to six week self-declaration to a multi-week grant through an accredited body, test labs have to be booked in advance, and a failed first pass adds a month or more. Add a lithium battery and you inherit transport testing that has to finish before you can even air freight your own validation units.
Two years or more for a regulated product
Medical, children’s products and anything making a health claim carry a review clock you do not control. In the EU, the European Commission’s notified body survey reports that most notified bodies put conformity assessment under the Medical Device Regulation at 13 to 18 months, with about a third indicating 19 to 24 months for combined quality system and product certificates. As at February 2026 that survey recorded 31,902 applications against 18,010 certificates issued. That backlog is a queue you join, not a process you manage.
What the timeline is actually made of
Engineering time versus waiting time
This is the distinction that makes a schedule readable. Two examples, both from the same project:
- EMC testing. Booking a lab in peak season takes six to eight weeks. The test itself takes two to three days for a self-declaration path and four to seven days for a single-band radio. Roughly 85% of “EMC testing” is queue.
- Ocean freight from Asia. Port-to-port sailing to the US West Coast runs about 15 to 20 days. Flexport’s Ocean Timeliness Indicator, which measures from cargo-ready at the factory to departure from the destination port, read 37.6 days for that lane in the week to 10 August 2026. The gap between those two numbers is waiting for a booking, waiting for a vessel and waiting at anchor.
Neither of those weeks appears on an engineering plan. Both appear on your calendar.
The three clocks
A hardware program runs three clocks at once, and they do not stop for each other.
| Clock | What it measures | Who controls it |
|---|---|---|
| Design | Engineering hours across industrial design, mechanical, electronics and firmware | You and your partner, mostly |
| Manufacturing lead time | Tooling, components, PCB fabrication, assembly, freight | Suppliers and the market |
| Compliance | Lab availability, test duration, review and grant, notified body capacity | Almost nobody you can call |
Compressing the first clock is what most founders try to do. It is usually the smallest of the three, and past a point, pushing on it makes the other two longer through rework.
The stage map from idea to first shipment
Feasibility and proof of concept
Four to eight weeks for most products. You are buying an answer to whether the physics, the certification path and the target price can coexist. Skipping it does not save the time, it moves the discovery to a point where fixing it costs ten times more. Our guide on product feasibility analysis covers what a proper one includes.
Engineering, prototype and the validation builds
The engineering itself is the part that behaves most like a normal project plan. The validation gates that follow are the part that surprises people.
| Gate | What it proves | Units | Duration |
|---|---|---|---|
| EVT | The design works | Tens | 6 to 10 weeks |
| DVT | It survives the real world and passes certification | Dozens to hundreds | 8 to 14 weeks |
| PVT | The line can build it repeatably at yield | A pilot run | 4 to 8 weeks |
What a founder is doing during those windows is mostly deciding. Approving samples, signing off test results, choosing between a fix that costs money and a fix that costs time. Our prototype testing checklist sets out what actually gets tested at each gate, and the difference between a prototype and an MVP explains why these builds are not the same thing as a product you can sell.
Tooling, certification, pilot and mass production
Tooling is where the calendar stops being about engineering at all. Production tooling runs 8 to 16 weeks or more. First sample shots typically land 45 to 65 days after order, and one contract manufacturer’s published examples put a small four-cavity ABS part at five weeks and a large single-cavity part at eight to nine.
Then the part that nobody plans for. Molders commonly report that most first mold trials need a modification cycle. The tool work itself is three to five days. The calendar cost, once you count sample production, shipping, review and approval, is one to three weeks per round. Assume one round.
Bridge tooling is the real expedite option, at three to six weeks instead of eight to sixteen, bought at the cost of tool life. Paying an expedite fee on a production tool adds 20% to 35% and buys back a fortnight.
Where the calendar actually goes: the queues
This is the section that changes how you plan. Every number below is a wait, not a work item.
Components
The single biggest schedule risk in 2026. Average peak semiconductor lead times reached 40 weeks in March 2026, a jump of about two thirds in a single month, against a 20 to 25 week range through most of 2025. Passive components stayed comparatively stable in a 10 to 20 week band, and the pressure was concentrated in active and optoelectronic parts: circuit protection, converters, diodes, logic and memory ICs, programmable logic and transistors.
A 40 week part ordered at the end of EVT arrives after the date you wanted to ship. That is not a supply chain problem you fix later, it is a design decision you make early, and it is the strongest argument for choosing parts with sourcing in mind rather than performance alone.
Certification
| Path | Typical elapsed time |
|---|---|
| Self-declaration, simple product | 2 to 6 weeks |
| Connected product on a pre-certified module | 3 to 8 weeks |
| Single-band custom radio | 6 to 16 weeks |
| Multi-radio product | 8 to 20+ weeks |
| Cellular, including carrier approval | 4 to 12+ months |
| Safety listing, typical with one or two rounds of findings | 12 to 20 weeks |
| UN 38.3 lithium battery transport testing | 3 to 6 weeks |
Two things worth knowing about this table. We could find no published FCC service-level target for equipment authorization, so every figure above is what labs and certification bodies report rather than an official standard. And UN 38.3 is a shipping prerequisite, not just a market one. Prototype and low-production-run cells can move under the exception at 49 CFR 173.185(e), on cargo aircraft only, with prior DOT approval and prescribed packaging. Anything you intend to ship commercially needs the test report, which puts it earlier on the critical path than most plans assume.
One genuinely counterintuitive finding. Bluetooth qualification is widely believed to take months. The Bluetooth SIG’s own process guide says listing can complete in as little as ten minutes and approval in as little as one business day. The months are test lab queue and test execution, and if you use an unmodified pre-qualified module, much of that disappears.
Fabrication, assembly and freight
- PCB fabrication. Standard runs about 20 working days. Quick turn is 1 to 5 days for simple boards, at a premium. Note that working days exclude weekends and factory holidays, which is how a “10 day” quote becomes a fortnight.
- Assembly. 10 to 15 days standard, 3 to 7 days expedited, once components are in hand. Components, not assembly, are the constraint.
- Customs. One to three business days for ocean freight after arrival. An intensive examination adds 10 to 20 days and $1,500 to $3,000.
Where founders lose months without noticing
- Failing EMC on the first attempt. Test labs have reported first-pass failure rates around 50% for years. A failure costs 4 to 12 weeks and $2,000 to $15,000. Pre-compliance scanning during design drops that risk to under 10%.
- A component going end-of-life mid-design. One documented case ran three weeks of engineering plus four weeks of requalification, pushing production start two months later. In a March 2026 Accuris and Fuld & Company survey of 439 engineering, procurement, quality and supply-chain professionals, 46% estimated the average cost of a single post-freeze change at over $50,000.
- Slow approvals on your side. One molder states plainly that the biggest cause of missed tooling deadlines is the back and forth between buyer and supplier during design review and sample approval, not the mold making. Your response time is on the critical path.
- Chinese New Year. This is the one that catches every first-time founder. Chinese New Year 2027 falls on 6 February, eleven days earlier than in 2026. The official holiday is a week or so, and the real impact is four to six weeks: slowdown from mid-January, complete shutdown from late January to mid-February, then a return at roughly a third of workforce capacity until early March. Complex orders need placing five to six months ahead.
- Golden Week. A second, smaller shutdown, 1 to 7 October, preceded by Mid-Autumn Festival, with backlogs into mid-October.
- A design choice that changes your regulatory route. Under the EU Radio Equipment Directive’s cybersecurity requirements, allowing a device to run with no password set at all removes your presumption of conformity under the harmonized standards, which pushes you into notified body assessment. Months, decided at schematic review.
What a credible partner should be able to tell you about dates
A firm that gives you a single date for a product it has not scoped is guessing. A firm that will not give you any date is not much more useful. What a good answer looks like:
- A range, with the assumptions named. “14 to 20 months, assuming a pre-certified radio module, one tooling revision, and no cellular.”
- The critical path identified. Which single item, if it slips, moves the launch date. Usually tooling, components or certification, rarely engineering.
- Gates, not a finish line. Dates you can hold them to at EVT, DVT and PVT, because those are estimable in a way that “launch” is not.
- What they need from you, and by when. Approval turnaround is a schedule input.
- Explicit buffer. Government schedule guidance treats a plan with no reserve as a plan with no credibility. The same applies here.
Vetting a manufacturing partner is a related but separate exercise, covered in our guide on how to find the right manufacturer.
How to use a timeline estimate when you are still deciding
- Work backward from a date that matters, then check whether the queues fit. A holiday season launch means tooling ordered before Chinese New Year, not after.
- Buy schedule where it is cheapest. A pre-certified module removes weeks of certification. A second-sourced component removes a 40 week risk. Bridge tooling removes five to ten weeks. All three are decisions made in the first quarter of the project.
- Plan the retest. Half of products fail EMC first time. A plan that assumes first-pass success is a plan with a hidden two month hole.
- Separate what you control from what you queue for. Two thirds of your calendar is queue. Fund the buffer and stop trying to compress the engineering.
- Fix the specification before the clock starts. Every change after design freeze costs schedule at a multiple, and cost tracks it closely. Our guide on hardware product development cost breaks down what that multiple looks like in money.
Frequently asked questions
How long does it take to develop a hardware product?
Roughly 12 to 18 months for a simple non-connected product, 18 to 30 months for a connected or battery powered one, and two years or more for anything regulated. Those assume an outsourced engagement with a defined specification at the start.
What takes the longest?
Waiting, not working. Component lead times reached an average peak of 40 weeks in March 2026, production tooling runs 8 to 16 weeks, and test labs in peak season are booked six to eight weeks out. Engineering is usually the smallest of the three clocks.
Can I make it faster by paying more?
Partly. Bridge tooling, expedited PCB fabrication, expedited certification review and air freight all buy real weeks. What money cannot compress is a notified body queue, a 40 week component, or a second EMC attempt.
How much time does Chinese New Year cost?
Four to six weeks of degraded or zero output, not the week the official holiday suggests. In 2027 the date is 6 February, eleven days earlier than 2026, so a schedule built on last year’s calendar will be wrong.
Why do timelines slip after everything looked fine?
Because the risks cluster at the end. Tooling revisions, EMC retests and component substitutions all land after design freeze, when the schedule has no slack left. Molders report that most first mold trials need a modification cycle, and test labs report around half of products failing EMC first time.
Is it faster to build it myself?
Nobody has published a credible paired comparison, so treat any confident answer with suspicion. What is structurally true is that an outside firm runs disciplines in parallel, has existing lab and factory relationships, and does design for manufacture before the tool is cut rather than after.
Where Inventornest fits
Inventornest plans hardware programs around the queues rather than around the engineering, because the queues are what move launch dates. That means component sourcing decided at schematic stage, certification strategy set before layout, and tooling ordered against a holiday calendar. Our product design and prototyping page covers the early stages, and you can see how these decisions played out on a real enclosure program in our smart brush cabinet project.
If you need a schedule for your own product rather than a range for a category, get a quote and we will give you the gates and the critical path with it.