Prototype vs MVP: Which One Your Product Actually Needs
A prototype answers whether it can work. An MVP answers whether anyone will buy it. Building the wrong one first is one of the most expensive mistakes a founder can make.
June 10, 20245 min read
Written by Konstantin Dolgan, Ph.D., NPDP
Founder & CEO, Product Development Engineer
Published June 10, 2024Updated September 2, 2026
A prototype is a build made to answer a technical or experiential question — can it work, does it fit, does it feel right — and is usually shown only to your team, testers and investors. An MVP is the smallest sellable version of the product, released to real customers to test whether they will actually buy and keep using it. One de-risks feasibility. The other de-risks demand.
The core differences
Prototype | MVP | |
|---|---|---|
Question it answers | Can it work? Does it fit? | Will people pay for it? |
Audience | Team, testers, investors | Real paying customers |
Build quality | Throwaway, often hand-built | Production-intent, supportable |
Feature scope | One question at a time | One complete job, end to end |
Regulatory burden | Minimal | Full compliance for the market |
Typical timeline | Days to weeks | Weeks to months |
Typical cost (hardware) | $500 to $50,000 | $50,000 to $500,000+ |
Failure means | Change the approach | Change the business |
Types of prototypes, and what each one proves
- Proof of concept — isolates the one physics or technology question everything else depends on. Ugly by design.
- Looks-like model — appearance, size, weight and finish, for buyer feedback and photography.
- Works-like model — the functional mechanism, usually in a housing nobody would ship.
- Looks-like-works-like — both together, for user testing, trade shows and pre-orders.
- Engineering prototype — production materials and processes, used for regulatory pre-testing and DFM.
Our guide to getting a prototype made covers the files, costs and timelines for each level.
What a real hardware MVP looks like
In software an MVP can be a stripped-down release shipped in a week. In hardware the minimum viable product still has to be safe, certifiable and repairable, which sets a floor on cost and time.
- A small production run — 50 to 500 units, often bridge-tooled or 3D-printed at production quality rather than injection molded.
- Real compliance — FCC, CE, UL or FDA pathways as applicable. This is not optional because you sold only a few.
- Packaging, instructions and support — returns and questions are part of the test.
- A price — a free unit measures curiosity, not demand.
- Instrumentation — you must be able to see usage, failures and churn, or the MVP teaches you nothing.
The order that saves money
Stage | What you build | Spend before proceeding |
|---|---|---|
1. Concept validation | Sketches, renders, landing page | $0 to $5,000 |
2. Proof of concept | Bench prototype of the hard part | $2,000 to $25,000 |
3. Looks-like-works-like | Testable, demonstrable unit | $10,000 to $60,000 |
4. Pilot / MVP run | 50 to 500 sellable units | $50,000 to $300,000 |
5. Scale | Hard tooling and volume manufacturing | $100,000+ |
Each stage should either produce a decision to continue or kill the program cheaply. Skipping stage two to reach stage four faster is the classic hardware failure, because tooling cannot be un-cut.
Where patents fit
Prototypes and patents interact, and the timing matters more than most founders realize.
- You do not need a working prototype to file. US patent law requires an enabling written description, not a physical unit.
- Public disclosure starts a clock. In the US you have a 12-month grace period after public disclosure or sale; most other countries have none, so file before you show it publicly.
- A provisional application establishes a priority date for around $130 to $320 in USPTO fees plus attorney time, and gives you 12 months to test the market.
- Prototype first, patent the version that works. Filing on your first idea often protects the concept you later abandon.
- Use NDAs for private testing and keep dated records of development.
If you are budgeting the IP side, our breakdown of the trademark application fee covers the brand-protection half of the same question.
Common mistakes
- Calling a prototype an MVP. If nobody can buy it, it is not testing demand.
- Polishing a prototype nobody asked for. Fidelity should match the question.
- Shipping an MVP with no way to measure it. No instrumentation, no learning.
- Filing patents before the design settles. Expensive protection for the wrong embodiment.
- Treating certification as a later problem. Compliance shapes design and cannot be retrofitted cheaply.
What the first hardware MVP actually costs
Cost line | Typical range | Notes |
|---|---|---|
Design and engineering to release | $40,000-$120,000 | Industrial design, mechanical, electrical, firmware |
Prototype rounds (3-5) | $8,000-$35,000 | Increasing fidelity toward production intent |
Tooling (soft or bridge) | $10,000-$60,000 | Depends on part count and cavitation |
Certification (FCC, UL/ETL, CE) | $12,000-$45,000 | Emissions, safety, and any wireless module testing |
Packaging design and transit testing | $5,000-$15,000 | Artwork, structure, ISTA testing |
First production run (500-2,000 units) | $25,000-$150,000 | Unit cost plus freight and duty |
Decide with a validation ladder
Rung | What you build | Spend | Question answered |
|---|---|---|---|
1 | Sketches and a spec | Under $2,000 | Is the problem worth solving? |
2 | Looks-like model | $2,000-$8,000 | Do people want this object? |
3 | Works-like rig | $5,000-$20,000 | Can the mechanism or circuit perform? |
4 | Integrated prototype | $15,000-$50,000 | Does it work as one product? |
5 | Production-intent MVP | $50,000+ | Can we make and sell it repeatedly? |
Key takeaways
- A prototype answers a question; an MVP earns revenue and must therefore be safe, compliant and supportable.
- Budget the whole path — certification and packaging routinely equal the engineering spend.
- Climb the validation ladder one rung at a time and stop when an answer is negative.
- File a provisional before any public demonstration if the mechanism is novel.
Where patents fit between the prototype and the MVP
Patent timing is the decision most founders get wrong in this sequence, usually by filing too early on a design that changes, or by disclosing at a trade show before anything is on file. In the United States, a public disclosure, offer for sale or public use starts a twelve-month clock to file. Most other jurisdictions have no grace period at all, so a public MVP launch can permanently forfeit foreign rights.
Stage | What exists | Sensible IP action | Typical cost |
|---|---|---|---|
Concept and sketches | An idea and a spec | Prior-art search; keep an inventor notebook | $500-$3,000 |
Proof of concept | The core mechanism works | File a provisional if the mechanism is settled | $2,000-$5,000 |
Works-like prototype | Function proven, form still moving | Provisional covering the working principle | $2,000-$5,000 |
Production-intent build | Design frozen for tooling | Non-provisional; design patent on the housing | $8,000-$20,000 |
MVP on sale | Customers can buy it | Foreign filings via PCT within 12 months | $4,000-$12,000+ |
The pattern that works: file a provisional once the working principle stops changing, keep iterating the embodiment, then convert to a non-provisional before the twelve months expire with everything you learned folded in. A provisional filed on a mechanism that then gets redesigned protects nothing useful, which is why the proof-of-concept prototype and the first filing usually belong in the same month.
A sequence that avoids paying twice
Prototype and MVP are not competing options; they are consecutive answers to different questions. Programs waste money when they skip a rung and buy a production-intent build to answer a question a $400 printed model could have settled, or when they ship an MVP whose feasibility was never proven.
- Prove the physics with a proof of concept before anything else gets designed around it.
- Prove the experience with a looks-like model in real users' hands; this is where feature scope gets cut cheaply.
- Prove the integration with a works-like build that combines mechanism, electronics and firmware in the intended envelope.
- Prove the process with a pilot run of 25 to 200 units off production-intent tooling or soft tooling.
- Prove the market with the MVP: real customers, real prices, support and warranty included.
Skipping the pilot run is the most expensive shortcut. Assembly problems that are annoying at 25 units are unrecoverable at 5,000, and a pilot build is also where the assembly instructions, fixtures and inspection criteria get written. Our guide to prototype cost breaks down what each rung buys.
Frequently asked questions
What is the difference between a prototype and an MVP?
A prototype is built to answer a technical or experiential question and is shown mainly to the team, testers and investors; it is usually thrown away. An MVP is the smallest complete, sellable version of the product released to real paying customers to test whether demand exists.
Should I build a prototype or an MVP first?
Build the prototype first. It costs far less and answers whether the product can work at all. Only once feasibility, form and usability are settled does an MVP make sense, because an MVP requires production-quality units, compliance and support.
How much does a prototype cost compared to an MVP?
Hardware prototypes typically range from $500 for a simple appearance model to $50,000 for a full engineering prototype. An MVP pilot run of 50 to 500 sellable units generally runs $50,000 to $300,000 once compliance, packaging and support are included.
Do I need a prototype to get a patent?
No. US patent applications require a written description that enables someone skilled in the field to make and use the invention, not a physical prototype. Building first is still often wiser, because prototyping reveals the version actually worth protecting.
Can an MVP be a prototype?
Only if customers can buy it and it is safe and compliant for the market. A demonstration unit sold to a friendly early adopter under a pilot agreement can serve as both, but a giveaway or demo unit tests curiosity rather than willingness to pay.
How many prototypes will I need?
Most hardware programs go through three to five prototype rounds: a proof of concept, one or two functional iterations, a looks-like-works-like unit for testing and demos, and an engineering prototype in production materials before tooling. Work with LA NPDT: if you are moving from here to execution, start with our rapid prototyping services or talk to us about prototype design .
Filed under:Uncategorized
Services related to this guide
- Best prototyping companiesDevelopment firms, instant-quote bureaus and local shops — who fits which job.
- Free prototype cost calculatorFive to eight questions, an instant cost range for a first prototype.
- Prototype development companiesWhich firms take a sketch to a working prototype — and which only fabricate parts.
- Rapid prototypingWorking prototypes in days, from 3D printing to vacuum casting.
