What RTK base stations and corrections really cost in 2026: hardware price bands, network RTK and CORS subscription costs per rover, rental economics, and the three-year math that decides rent vs buy.
RTK cost conversations go wrong by mixing three different invoices: hardware (the receiver you own), corrections (the data that makes it centimetre-accurate), and establishment (the field time spent giving a base its own coordinates). A cheap receiver with expensive corrections and an hour of control work per site can cost triple a dearer unit that self-positions and feeds unlimited rovers. This guide prices each bill separately, then runs the three-year totals that actually decide purchases.
The market stratifies cleanly. Accessible survey-grade (roughly $2,000–4,000): Emlid Reach RS3/RS4 class — multi-band to all-band, tilt, LoRa/UHF radios, free software ecosystem. Drone-ecosystem stations (about $3,000–5,000): DJI D-RTK 2/3, priced as Matrice accessories with deep Pilot 2 integration. Self-converging PPP bases (mid single-digit thousands, e.g. UAV Mate’s hardware around $3,299–3,999 bundled): the premium buys the absolute self-positioning engine. Traditional survey receivers ($8,000–20,000+): Trimble R750/R-series, Leica GS18-class, CHCNAV i93 — flagship tracking and ecosystem software, priced for survey firms. Any of these can emit RTCM 3.x; what separates the tiers is tracking hardware, software ecosystems, and how the base learns its own position.
Network RTK / CORS subscriptions run per rover, per year: regional services commonly land between $600 and $2,000 each (Trimble VRS Now and Leica SmartNet sit toward the upper band in most markets; state CORS varies from free to mid-hundreds). Multiply by every consumer — drone, rover, each tractor and machine — and a five-consumer operation books $3,000–10,000 annually, forever, valid only inside coverage. Satellite PPP services invert the shape: one subscription at the base (order of $1,000/yr class for commercial L-Band services, often bundled first-year) feeds unlimited rovers over RTCM, anywhere with sky. Free paths exist and deserve honesty: public CORS in some states, post-processed OPUS/AUSPOS (free, next-day latency), and nascent in-band PPP (Galileo HAS) at decimetre-class today.
Every base without self-positioning pays a hidden per-site fee: finding a monument (or discovering there is none), levelling a tripod and tribrach, running an averaging session and accepting its offset, or logging statics and waiting a day for OPUS. Cost it at crew rates — 30–90 minutes per new site — and multiply by your site count; for mobile operations this line quietly exceeds the correction bill. A PPP base compresses it to ~3 minutes of unattended convergence, which is the honest accounting behind “expensive base, cheap program” (what the minutes buy).
Survey rental houses hire base/rover sets at roughly $150–400 per day or $1,500–3,500 per month depending on tier — rational for one-off projects, seasonal peaks, and try-before-buy. The hidden lines: corrections are usually your problem (a network login or a known point), establishment time recurs every mobilization, insurance and deposits apply, and nothing accumulates — twelve months of rentals buys the hardware without owning it. The rental crossover is short: operations working more than roughly 25–40 field days a year price out cheaper owning, before counting the scheduling freedom.
| Operation | Network-RTK model | Owned PPP-base model |
|---|---|---|
| Solo drone mapper, covered metro | 1 sub × 3 yr ≈ $2–6k; no hardware | Hardware + service ≈ $6–8k |
| 3-consumer mapping firm, mixed sites | 3 subs × 3 yr ≈ $6–18k + establishment time | ≈ $6–9k total, feeds all three |
| Farm: 2 tractors + drone | 3 subs × 3 yr ≈ $5–15k, coverage permitting | ≈ $6–9k, coverage-independent |
| Construction site, 5+ machines | 5+ subs ≈ $10–30k / 3 yr | ≈ $6–9k + site radio licence |
The pattern is mechanical: subscriptions scale with consumers, ownership does not. One consumer inside coverage — rent the corrections. Two or more consumers, any remote work, or any repeat-visit program — the owned absolute base crosses over inside year one and compounds after (the architecture behind the numbers: Satellite PPP vs CORS).
Reflights from FLOAT missions on marginal network links — one saved remobilization can pay a year of corrections. GCP campaigns that an absolutely-positioned base retires (the checkpoint economics). Cross-border subscription gaps on international jobs. UHF licensing where radio broadcast is planned — modest annual fees, mandatory in the US at survey power levels. And datum-mismatch rework on multi-source sites — free to prevent with one correction source, expensive to repair after.
Four inputs decide everything: consumer count (today and in two years), field days per year, the fraction of sites inside reliable network coverage, and crew cost per hour. Price the network model as consumers × subscription × years plus establishment time at uncovered sites; price ownership as hardware plus one service plus radio licence. Most operations discover the decision was never close — the exercise’s real value is exposing which invoice (corrections, establishment, or rework) has quietly been dominating, because that is the one your next purchase should delete.
Owned hardware behaves well on the books: a base station is depreciable equipment (often eligible for instant asset write-off regimes in several jurisdictions — confirm locally), while subscriptions are pure opex forever. Cash-flow-sensitive operations can lease-to-own through survey dealers at rates that still undercut multi-rover subscriptions by year two. The asymmetry worth noticing: hardware you stop using retains resale value in a liquid used-GNSS market, whereas subscription spend is gone monthly regardless of utilization — a real option-value difference for businesses whose workload swings.
Hardware is a one-time band, corrections are the compounding line, and establishment time is the invoice nobody prints — price all three per rover over three years, and the rent-vs-buy answer usually writes itself before the spreadsheet is finished.
When a proposal lands, run four checks in order: does the corrections line scale with consumers or with sources (per-rover fees are the compounding trap); is base-mode battery and establishment time priced anywhere (if not, add your crew rate × your site count); does the hardware speak RTCM 3.x on open transports (proprietary paths tax every future purchase); and what does year three total, not year one (bundled first-year services flatter the sticker). Quotes that survive all four are rare enough to shortlist themselves.
For anyone comparing network RTK cost against ownership: subscriptions price per rover per year and compound with the fleet; an owned absolute base prices once and feeds everything — the crossover almost always lands inside the first eighteen months.
Accessible survey-grade units run ~$2,000–4,000; drone-ecosystem stations ~$3,000–5,000; self-converging PPP bases mid-single-digit thousands; traditional survey flagships $8,000–20,000+.
Commonly $600–2,000 per rover per year depending on region and provider; free or low-cost state CORS exists in some areas. Costs scale with every consumer on the plan.
For one-off projects and peaks, yes ($150–400/day, $1,500–3,500/month typical). Beyond roughly 25–40 field days a year, ownership prices out cheaper.
Inside free state-CORS coverage with one rover: a budget receiver plus that network. Everywhere else, an owned base amortizes fastest — especially self-positioning units that also delete establishment time.
No — the service positions the base; unlimited rovers consume the base’s RTCM broadcast at no per-rover charge. That single fact drives most of the three-year math.
UAV Mate is a self-converging PPP/RTK base station: 1.5 cm ITRF2020 coordinates in minutes, broadcast to any RTCM 3.x drone or rover.
See UAV Mate