The mmWave Market in 2026

Millimeter wave — mmWave for short — is the band of radio spectrum roughly between 24 and 100 GHz. It is the physics that makes multi-gigabit 5G possible in a stadium, that lets a car's radar spot a pedestrian in fog, and that powers the imaging portals at an airport security lane. For a founder, that breadth is both the opportunity and the trap: a plan that tries to serve every use case at once rarely raises money. A plan that picks one and proves it can.

The overall millimeter wave technology market was valued at roughly $4.1 billion in 2025 and is forecast to compound at about 17% a year, reaching close to $9.7 billion by 2035 (Market Research Future, 2025). The radar slice alone was worth $4.62 billion in 2025 and is growing faster — a stated 21.6% CAGR to $18.16 billion by 2032 (Coherent Market Insights, 2025). Those two figures overlap in places because analysts scope the category differently; the point for your plan is that both the component and the sensing sides of mmWave are growing at double-digit rates.

The demand drivers are concrete rather than hype. Automotive radar in the 76–81 GHz band is now standard in advanced driver-assistance systems for collision avoidance and parking. 5G operators are lighting up 24–40 GHz spectrum for capacity in dense urban zones. And a newer wave of "human sensing" — presence detection, vital-signs monitoring, gesture control — is putting 60 GHz radar into smart homes, hospitals and vehicles, precisely because mmWave can sense breathing and heartbeat without a camera.

Looking further out, the standards work on 6G already assumes heavier use of mmWave and the sub-terahertz bands above 100 GHz. That is a tailwind for anyone building capability now, but it is not a business on its own: 6G volume is years away, and a plan that leans on it as the revenue case will not fund the runway needed to get there. Treat 6G as a reason your capability compounds, not as your first customer.

Where do most guides on this subject stop? At the headline market number. The figure that actually determines whether an mmWave company survives is not the total addressable market — it is the cost of a single design win and how long that win takes to convert to shipped volume. A 5G module maker can win a small-cell socket in months; an automotive radar supplier can wait two to three years for the same. Your plan has to model that gap, because it is what your runway has to cover.

Questions Founders Ask First

Before the financials, most first-time mmWave founders — and their investors — want plain answers to a handful of questions. Here they are, in the order they usually come up.

What is mmWave technology actually used for?

Four commercial pillars: 5G and emerging 6G radio (small cells and fixed wireless access delivering fibre-like speeds without trenching); automotive and industrial radar; imaging and security screening; and satellite and backhaul links. A fifth, faster-moving pillar is short-range human sensing at 60 GHz. Each pillar has different customers, certification routes and margins, which is why the business plan, not just the technology, decides who wins.

Why is the range so short?

Higher frequency means more bandwidth but worse propagation. mmWave energy is absorbed by atmospheric oxygen and water vapour, reflected by buildings, blocked by walls and foliage, and cut by rain fade. A 28 GHz small cell might reach a few hundred metres with clear line of sight, versus kilometres for a sub-6 GHz macro cell. Beamforming — steering a focused beam electronically with a phased array — is how the industry claws back link budget, and it is why mmWave products carry more antennas and more signal-processing than lower-band radios.

Do I need a spectrum licence to build an mmWave product?

Sometimes. Licensed bands (28 GHz, 24 GHz, 37/39 GHz in the US; 26 and 40 GHz in the UK) need spectrum rights, which are usually held by carriers or won on a local Shared Access basis. Unlicensed bands (57–71 GHz, including 60 GHz) need no spectrum licence at all. But every transmitter — licensed or not — needs equipment authorisation from the regulator before it can be sold. The licensing section below breaks this down by jurisdiction.

Is this a hardware business or a software business?

Both models exist and they raise money very differently. A fabless chip or module company is capital-intensive up front and margin-rich later. A design-services or IP-licensing company is lighter on capital but caps out on scale. A test-and-measurement or "characterisation-as-a-service" company sells access to equipment most startups cannot afford. Your plan should name which of these you are — investors price them on different multiples.

Who Buys mmWave, and How You Reach Them

The fastest way to lose a deep-tech investor is to answer "who is your customer?" with "the 5G market." mmWave buyers are specific, technical and few, and they buy on evidence rather than marketing. A strong plan names the segment, the person who signs the purchase order, and the proof that person needs before they commit. Below are the buyer types most mmWave founders sell into, and what each one actually cares about.

• Network-equipment makers (OEMs): Tier-1 vendors like Ericsson and Nokia are hard to reach early; Tier-2 and private-network small-cell makers are the realistic beachhead. They value a component that shortens their own design cycle and comes with characterisation data they can trust.
• Automotive and Tier-1 suppliers: radar module buyers with long qualification programmes and demanding reliability standards. High lifetime volume, but a 24-36 month path to production that your runway has to survive.
• Defence and aerospace primes: buyers of phased arrays, satcom terminals and sensing, often reachable through SBIR/STTR or Innovate UK-funded programmes before any commercial sale.
• Industrial and smart-building integrators: the fast-moving 60 GHz sensing market — presence detection, people-counting, vital-signs monitoring — where sales cycles are measured in months, not years.

The go-to-market motion follows from the segment. mmWave is not a business you win with paid advertising; you win it with reference designs, evaluation boards, conference demonstrations and technical papers that let a customer's own engineers de-risk your part. Companies such as Texas Instruments built their radar franchise on exactly this — an evaluation module in every engineer's hands. A credible plan budgets for eval hardware and applications-engineering support as a sales cost, because in this market that support is the sales team.

Geography matters more than founders expect. The UK's RF strength clusters around Cambridge, Bristol and Glasgow; the US around San Diego, Boston and North Carolina's Research Triangle. Locating near one of these hubs buys access to scarce mmWave engineers, shared test facilities and the specialist investors who understand why a semiconductor company burns cash for three years before it scales. Your plan should say where you are and why, because for this category the answer is a genuine competitive factor.

What It Costs to Start

A fabless mmWave design or module startup typically needs $180,000 to $950,000 (about £145,000 to £760,000) to reach a working prototype and first customer over the first 12–18 months. The spread is wide because it depends entirely on whether you build your own test capability or rent it, and whether you tape out silicon or integrate off-the-shelf front-ends. Full semiconductor fabrication is an order of magnitude more and is almost always outsourced to a foundry rather than built in-house.

The single line that surprises founders is test and measurement. mmWave cannot be debugged on a hobby bench. A vector network analyser that reaches 67 GHz runs well into six figures, and one that reaches 110 GHz can exceed $250,000 on its own — before probe stations, spectrum analysers and over-the-air chambers. This is why so many mmWave startups cluster near a university or a shared lab: they are effectively borrowing capex.

Cost breakdown

• Test & measurement (VNA to 67/110 GHz, spectrum analyser, probe station): $60K–$350K (£48K–£280K)
• EDA / RF design software (Keysight ADS, Cadence AWR, Ansys HFSS): $25K–$120K per year (£20K–£95K)
• Prototype fabrication / multi-project-wafer foundry runs (GaAs, GaN, SiGe): $30K–$180K (£24K–£145K)
• Anechoic / over-the-air test chamber access or build: $20K–$150K (£16K–£120K)
• FCC / Ofcom equipment authorisation & compliance testing: $8K–$45K (£6K–£35K)
• Founding RF / DSP engineering team (first 12 months): $120K–$400K (£95K–£320K)

Notice how little of that is premises or inventory — the classic startup costs a generic template would lead with. An mmWave venture spends its money on measurement, talent and silicon access. A plan that lists "office rent" and "marketing" as its top costs signals to a deep-tech investor that the founder has not priced the real work. The template below reorders the cost model so the expensive, technical lines sit where they belong: at the top.

UK founders often model a lower headline number than US peers, and not only because salaries differ. Two things pull the effective cost down. First, non-dilutive grants — Innovate UK, and the shared facilities at catapults such as the Compound Semiconductor Applications Catapult in Newport — can offset a meaningful slice of the test and prototyping bill. Second, SEIS and EIS reliefs make early UK equity cheaper to raise against the same technical milestones. The mistake is to treat those offsets as certainties before they are secured: a plan should show the fully-loaded cost first, then the funded position after grants, so a reader can see the gap you are actually asking them to cover.

Suppliers, Foundries & Test Partners

No mmWave company builds everything itself. Your plan's operations section should name the supply chain you intend to rely on, because it tells an investor you understand what you are outsourcing and what you are keeping. These are the players most UK and US mmWave startups actually work with.

Component and chipset incumbents

• Analog Devices — mmWave transceivers, beamformers and the widely used automotive radar chipsets.
• Qorvo — gallium-nitride (GaN) power amplifiers and front-end modules for 5G and defence.
• Texas Instruments — the IWR/AWR single-chip 60 GHz and 77 GHz radar sensors many prototypes start on.
• Broadcom — RF filters, switches and front-end modules across 5G, Wi-Fi and mmWave.
• Qualcomm — 5G mmWave modem and antenna-module reference designs that set the pace for handsets.
• Sivers Semiconductors — mmWave transceiver modules; it acquired the startup MixComm in 2022.

Foundries and materials

• GlobalFoundries — SiGe BiCMOS and, increasingly, GaN processes suited to mmWave front-ends.
• WIN Semiconductors and OMMIC — GaAs / GaN pure-play RF foundries used for multi-project-wafer runs.
• IVWorks — GaN epitaxial wafers for RF and power; it has raised around $33M cumulatively.
• Finwave Semiconductor — GaN-on-silicon RF switches and amplifiers for 5G/6G; raised an $8.2M round.

Design and test tooling

• Keysight and Rohde & Schwarz — the mmWave VNAs, signal generators and OTA chambers you will rent or buy.
• Cadence AWR and Ansys HFSS — the circuit and electromagnetic simulators for phased arrays and antennas.
• Gapwaves — waveguide antenna technology, a reference point for anyone doing high-gain mmWave apertures.

Naming this chain does two things in a plan. It shows the venture is fundable without a fab of its own, and it exposes concentration risk — if your whole product depends on a single foundry slot, a reader will want to see a second source. That is exactly the sort of operational realism that separates a plan that gets a second meeting from one that does not.

How mmWave Companies Make Money

There are five common revenue models in mmWave, and most successful companies blend two of them. Component and module sales generate volume revenue at fabless gross margins of roughly 55–70% once yields mature. IP and design-service licensing — selling a proven beamformer or antenna block to a larger OEM — carries 35–50% margins but far lower capital needs. Non-recurring engineering (NRE) plus royalty deals front-load cash and tail into per-unit income. Systems integration — building the finished radar or radio for a customer — earns 25–40% and scales with headcount. And test-as-a-service monetises the expensive equipment other startups lack.

A worked example

Take a fabless startup selling a 28 GHz beamforming front-end module into private-network small cells. At an all-in price that nets $34 per module after bill-of-materials and channel margin, shipping 40,000 units a year to two Tier-2 5G equipment makers, the business books about $1.36 million in revenue. At a 58% gross margin that is roughly $790,000 of gross profit — enough to carry a lean 8-person engineering team, but only if the design win was secured before the runway ran out. Double the volume through a third customer and the same fixed cost base pushes net margin from thin to healthy. That jump in operating margin as volume grows is the whole investment thesis, and your model needs to show it explicitly.

The recurring-revenue question matters here too. Pure hardware is lumpy; investors pay more for anything that recurs. mmWave companies build recurrence through multi-year supply agreements, firmware and calibration subscriptions on radar sensors, and design-in royalties that pay every time the customer ships. A plan that shows how one-off sales convert into a repeatable, contracted book of business will always raise on better terms than one selling widgets one purchase order at a time.

Pricing discipline is its own subject. Because mmWave parts are hard to design and harder to second-source, the temptation is to price on cost-plus and leave money on the table. The better anchor is the value the part creates for the customer: a beamforming module that removes six months from an OEM's development schedule, or a radar sensor that lets a car maker delete a separate camera, is worth far more than its bill of materials. The plan should state the pricing logic explicitly and defend the gross margin it implies, because that margin is what every later financial line depends on. Vague pricing is the single most common reason a technically strong mmWave model still fails to convince a finance-minded reader.

US Funding & SBA Reality

mmWave hardware sits in NAICS codes such as 334220 (radio and wireless communications equipment manufacturing) and 334413 (semiconductor and related device manufacturing), with design-only firms often classified under 541715 (R&D in the physical, engineering and life sciences). That classification matters because it shapes which funding doors are open.

SBA 7(a) loans, the workhorse of US small-business finance, technically reach up to $5 million with terms to 25 years. In practice, pre-revenue deep-tech hardware is a poor fit for a cash-flow lender: 7(a) underwriting wants trading history and collateral that a fabless startup does not have. The realistic US capital stack for mmWave is different, and your plan should target it directly:

• SBIR / STTR grants: non-dilutive awards from the NSF, DoD, DARPA and NASA — a natural fit for mmWave's defence, radar and 5G-infrastructure angles. Phase I is typically $50K–$275K; Phase II can reach $1M–$2M.
• Seed venture capital: deep-tech and semiconductor-focused funds. Comparable rounds are named below — MixComm raised $8.6M; Finwave raised $8.2M; GaN peers have closed $2.6M–$7.5M seeds.
• Strategic and corporate investment: foundries and incumbents (as GlobalFoundries did with Finwave) invest in the ecosystem they will later supply.
• Equipment financing: where SBA loans do help — financing the test bench against the asset itself, rather than the business.

The competitor funding data is the useful part for a founder. It tells you the going rate for a seed round in this niche is single-digit millions, not a $50K friends-and-family cheque, and it gives you comparables to anchor your own ask. Our Research + Content and Bespoke packages build the SBIR-ready and VC-ready financials — five-year model, use-of-funds, and the milestone map investors expect — around exactly these benchmarks.

One line every mmWave investor reads closely is the intellectual-property position, because it is what stops a better-funded incumbent from copying you the moment you prove the market. In this field the defensible assets are usually patents on antenna and beamforming architectures, calibration and signal-processing know-how, and the accumulated characterisation data that makes your part trustworthy to a customer's engineers. A plan should state what is filed, what is trade secret, and — just as important — where you rely on third-party IP or standard essential patents that carry licensing obligations. The point is not to claim an impregnable moat; it is to show you understand exactly how deep your moat is and how you intend to make it deeper with each funding round.

Spectrum, Licensing & Compliance

Two separate approvals get confused constantly: the right to use a slice of spectrum, and the right to sell a device that transmits. You may need one, both, or (for unlicensed bands) only the second. Get this wrong in a plan and a technical investor will notice immediately.

United States (FCC)

• Licensed mmWave bands under the Upper Microwave Flexible Use Service (UMFUS): 28 GHz (27.5–28.35), 24 GHz (24.25–24.45 and 24.75–25.25), the 37/39 GHz bands, and 47.2–48.2 GHz — held via geographic-area or Partial Economic Area licences, generally won at auction.
• Unlicensed operation under Part 15 in the 57–71 GHz band (including 60 GHz) — no spectrum licence, but strict power limits.
• Automotive radar in the 76–81 GHz band — rule-compliant, no individual licence.
• Equipment authorisation (Part 2 certification via a Telecommunication Certification Body) for any intentional radiator before sale — budget $8K–$45K and roughly 6–12 weeks.

United Kingdom (Ofcom)

• 26 GHz (25.1–27.5) and 40 GHz (40.5–43.5) mmWave licences were auctioned in September 2025 across 68 High Density Areas. EE, O2 and VodafoneThree each won 800 MHz at 26 GHz plus 1 GHz at 40 GHz, paying about £13 million apiece (Ofcom, 2025).
• Local 26 GHz access is available through Ofcom's Shared Access licensing framework for campus, venue and private-network use — a realistic route for a startup that does not need national spectrum.
• UKCA conformity for radio equipment under the Radio Equipment Regulations 2017 before placing a product on the UK market.

European Union / CEPT

• The 26 GHz band (24.25–27.5) is harmonised as the EU 5G mmWave "pioneer band" under CEPT/ECC decisions, giving a single technical framework across member states.
• The 66–71 GHz range supports unlicensed wireless access, and CE marking under the Radio Equipment Directive is the gate to selling anywhere in the single market.

For most early-stage mmWave companies, the practical path is an unlicensed band or a local Shared Access licence plus equipment certification — not a national spectrum auction. Saying so in the plan, and putting a realistic figure and timeline against certification, is a credibility signal in itself.

There is also an export and controls dimension that generic plans miss entirely. High-performance mmWave components — particularly GaN power devices and phased-array beamformers with defence or satcom applications — can fall under export-control regimes such as the US ITAR and EAR, or the UK's strategic export controls. If any part of your addressable market is defence, aerospace or a restricted jurisdiction, the plan should acknowledge the licensing and end-user checks that come with it. Investors in this space expect to see the question raised; its absence reads as inexperience rather than as a clean bill of health.

mmWave Terms, Explained

An mmWave plan is read by both technical and non-technical people. Defining your terms once, clearly, makes the whole document easier to underwrite. These are the words that carry the most weight.

• Beamforming: steering a focused radio beam electronically using an array of antennas, so energy points where it is needed instead of radiating in all directions — the core trick that makes short-range mmWave usable.
• Phased array: the antenna hardware that does the beamforming, made of many small elements whose signals are combined with controlled phase shifts.
• Front-end module (FEM): the block of power amplifier, low-noise amplifier, switch and filter that sits between the transceiver and the antenna. Often the product an mmWave startup actually sells.
• Link budget: the accounting of every gain and loss between transmitter and receiver. mmWave link budgets are tight, which is why the number appears on the first page of any credible technical plan.
• Rain fade: signal loss caused by rain absorbing and scattering mmWave energy — a real coverage variable, not a footnote.
• GaN and SiGe: gallium nitride and silicon-germanium, two semiconductor materials that dominate mmWave. GaN handles power and heat; SiGe integrates cheaply at scale.
• OTA testing: over-the-air testing, where a device is measured radiating through an antenna in a chamber rather than through a cable — mandatory for phased-array products because you cannot probe every element directly.
• Design win: the moment a customer commits your part into their product. In mmWave it is the true unit of progress, and its cost and cycle time drive the whole financial model.

Five Mistakes That Sink mmWave Plans

Across deep-tech plans we have reviewed, the same avoidable errors keep appearing in mmWave ventures. Each one is a place where a sharper plan pulls ahead.

• Treating mmWave coverage like sub-6 GHz. Assuming macro-cell reach ignores rain fade, foliage and line-of-sight limits, and inflates every deployment and revenue number that follows.
• Under-budgeting test and measurement. Founders line-item a laptop and forget that a 110 GHz VNA can cost more than a year of salaries. Investors know this and check for it.
• No named beachhead. A plan that pitches 5G, radar, satcom and imaging simultaneously reads as unfocused. Pick one application, win it, then expand.
• Ignoring qualification cycle time. Automotive design wins can take 24–36 months to convert to volume. A runway sized for a 6-month sales cycle runs out before the first big order ships.
• Leaving certification to the end. Equipment authorisation and conformity are not afterthoughts. Plans that budget FCC or UKCA work up front avoid the classic post-tape-out cash crunch.

For a related build on the radio side of this market, see our 5G chipset business plan template and the 5G fixed wireless access business plan template, both of which share supply-chain and spectrum dynamics with mmWave.

Sample Business Plan Preview

Here is an extract from an mmWave business plan written by our team, so you can see the level of specificity a deep-tech investor expects:

What's in the Template

Every Avvale business plan template is pre-structured for its industry. The mmWave version is tuned for RF and deep-tech ventures, with prompts written for hardware, spectrum and IP rather than generic retail:

• Executive Summary — Your technology, beachhead application and ask, framed for a technical investor in 60 seconds.
• Company & Technology Overview — Legal structure, founding team, IP position, and where you sit in the mmWave value chain.
• Market Analysis — Market size, CAGR by segment, and the specific application you are attacking first, with room for cited sources.
• Customer & Design-Win Strategy — Who buys, how design wins are secured, and the cycle time from sample to volume.
• Competitor Analysis — Incumbents and funded startups, and where your unfair advantage lies.
• Operations & Supply Chain — Foundry, test partners, second-source risk, and the certification pathway.
• Go-to-Market Plan — Channels, reference customers, and how one-off sales become recurring revenue.
• Management Team — Founder credentials, technical advisers, and the key hires the funding pays for.

The optional Financial Forecast add-on (included in our $300/£250 and $1,000/£800 packages) provides a five-year Excel model with income statement, cash flow, balance sheet, break-even analysis, and a use-of-funds table sized to the SBIR and seed-VC benchmarks in this niche.

Because mmWave spans very different sub-niches, the template is built to be adapted rather than filled in blindly. A 60 GHz sensing startup selling into smart buildings will foreground short sales cycles and unit volume; a defence phased-array company will foreground programme funding, security clearance and export controls; a fabless chip company will foreground foundry relationships and tape-out milestones. The section prompts flag where each of these paths diverges, so the plan you produce reads as if it were written for your specific corner of the market — which, to the investor or grant reviewer on the other side of the table, is exactly the point.

Frequently Asked Questions