Orbit Without a Launch: India's Satellite Internet Market, Its Pending Framework, and the Economics That Decide It

India has spent four years licensing a service that nobody can buy. Eutelsat OneWeb received its satellite-communications licence in April 2022, Jio Satellite Communications in October 2022, and Starlink in June 2025. In October 2026, all three still cannot carry a paying customer’s traffic — and the reason is not satellites. It is a spectrum order that has not been signed, a security clearance that has not been granted, and a unit-economics problem that no licence solves.

The flashpoint came in the second week of October. Elon Musk accused unnamed “oligarchs” of blocking Starlink in India “to maintain their monopolistic chokehold”, and asked whether “Ambani is the real boss of India”. The Ministry of Communications replied the next day with a dry procedural fact: no applicant had been cleared, security assessment was under way for all three licensees, and “at present, all the three entities are at broadly the same regulatory stage.” Two days later the Communications Minister repeated that India permits no monopolies in any sector.

Both statements are true, and together they describe the market precisely. This is not a contest of technology or capital. It is a queue, with two gates left in it — and behind the queue, a market whose willingness to pay is the lowest of any large economy on earth.

Who is actually licensed

The clearest public inventory comes from a Lok Sabha reply of 5 August 2026. Unified Licence with Global Mobile Personal Communication by Satellite authorisation has been issued to exactly three companies: OneWeb India Communications, Jio Satellite Communications, and Starlink Satellite Communications. A second, quieter layer exists beneath them: Unified Licence with Commercial VSAT CUG authorisation, held by eleven companies — a list headed by Nelco and Hughes Communications India, and including BSNL, Reliance Jio Infocomm, OneWeb India, Jio Satellite, Starlink, Avantel, Cloudcast Digital, ONEOTT Intertainment and Inmarsat India.

That two-tier structure is the business map. Three entities can operate constellations. Eleven can operate ground and enterprise networks. Distribution, installation and the last-mile relationship sit with the second group — which is exactly why the first group has been signing with them rather than competing against them.

Licences, though, are not launches. Two further authorisations stack on top. The first is IN-SPACe’s: Starlink’s Gen1 constellation was authorised on 8 July 2025, valid to 7 July 2030. The second is the security clearance that sits above both, and which, as of October 2026, no licensee has.

The regime itself is being rebuilt while the applicants stand in it. The Telecommunications Act 2023 places satellite-based services — GMPCS among them — in its First Schedule, which is what allows spectrum to be assigned administratively rather than auctioned. The Telecommunications (Authorisation for Provision of Principal Telecommunication Services) Rules, 2026, notified on 23 June 2026, then collapsed the old licence categories into five broad service groups, folding satellite communications into general telecom authorisation rather than a separate class. The immediate casualty was Amazon Leo, formerly Project Kuiper: its GMPCS application was cancelled mid-process, it must reapply, and it is now reported to be sounding out partners — Nelco among them — for last-mile delivery. Leo has more than 390 satellites in orbit, plans up to 5,105, and targets direct-to-device from 2028. Existing GMPCS holders migrate to the new regime only at the end of their twenty-year terms.

What the Indian market actually is

Two numbers have to be held together.

The first is the space economy, which is where the growth story lives. Parliament was told on 2 April 2026 that India’s space economy stood at about $8.4 billion in 2023 — roughly 2% of the global total — and is projected to reach approximately $44 billion by 2033 (about 8%) and $100 billion by 2040 (about 10%). That is an 18% compound annual growth rate between 2023 and 2033, the kind of figure that gets a sector financed.

The second number is what satellite is actually competing against. In the year to March 2026, India had 1,092.79 million internet subscribers — 1,065.88 million of them on broadband. Average revenue per wireless user was ₹180.73 a month. Indians consumed 2,85,376 petabytes of wireless data, generated ₹2,39,874 crore of data revenue, and paid an average of ₹8.02 per gigabyte — a realisation that fell 10.59% over the year. Average use per subscriber reached 25.51 GB a month.

Put those together and the market defines itself. India is not a country where satellite internet will win on price, because no technology wins on price against ₹8.02 per gigabyte and falling. Satellite’s Indian market is exactly the set of places where terrestrial economics fail: the sites fibre cannot reach, the vessels and rigs and aircraft that move, the disaster zones, the defence deployments, and the towers that have no backhaul.

That set is smaller than the headline “700 million unconnected” framing suggests. BharatNet has achieved service-ready status in 2.19 lakh of a revised 2.26 lakh Gram Panchayats — 96.9% — at a cost of ₹39,888 crore, with the Comptroller and Auditor General finding significant utilisation deficiencies against that spend. The coverage gap that remains is real but narrow; the affordability gap is what actually keeps people offline. The honest size of the satellite opportunity is therefore a low single-digit percentage of the ₹2,39,874 crore India already spends a year on wireless data. One per cent of that is ₹2,398.74 crore.

There is also a hard physical ceiling on the urban case, and the regulator has stated it. A single satellite constellation, TRAI’s recommendations note, can offer just 10,000 to 20,000 connections in an area like Delhi against a requirement of 50 lakh broadband connections. That is 0.3% of one city’s demand. Satellite is not an alternative to Indian urban broadband. It is a rounding error in it — and it only makes sense where the alternative is nothing at all.

Per-head acquisition: the arithmetic that decides everything

This is where a business case either survives or does not.

Start with what the state will charge you before a customer pays you anything. TRAI recommended that satellite spectrum be assigned for five years with a two-year extension, at a levy of 4% of adjusted gross revenue, subject to a minimum of ₹3,500 per MHz per year. The Department of Telecommunications preferred 5%. The Digital Communications Commission settled it on 3 September 2026: 5% of AGR, with a one-percentage-point discount — 4% — for services delivered in government-notified hard-to-connect and remote border areas. A Cabinet note was ready by 11 October and approval was expected within days. Note the shape of that instrument: a per-head charge on a market whose per-head revenue is among the lowest in the world. TRAI also proposed an additional ₹500 per subscriber per year for urban subscribers, with rural exempt, so the cost of acquiring a subscriber lands precisely where the ability to pay for the service is highest and the capacity is scarcest. In rural and unserved areas — where satellite is genuinely needed — TRAI recommended a subsidy for user terminals, with one-time hardware typically running ₹20,000 to ₹50,000.

Now the other side of the ledger. India’s blended wireless ARPU is ₹180.73 a month. Starlink’s residential plans across its 26 live African markets run roughly $30 to $55 a month; at an exchange rate near ₹96.5 to the dollar, $30 is about ₹2,895 — 16 times India’s monthly ARPU. The ₹8,600-a-month plan with a ₹34,000 kit that briefly appeared on Starlink’s India site in December 2025 was, the company said, dummy test data exposed by a configuration error. Taken at face value it would have been around 48 times Indian ARPU. No announced Indian price exists, and anyone quoting one is guessing.

Model it anyway, because the shape of the answer does not depend on the exact number. Assume a ₹2,000 monthly retail plan — already generous, and already 2.5 times the market’s effective price of data. Take the midpoint of TRAI’s own hardware range, ₹35,000. Recovering that terminal alone takes 17.5 months of revenue at 100% gross margin before spectrum charges, gateway operating costs, security-compliance overhead or any return on capital. Push it to a per-gigabyte comparison and the problem sharpens: ₹2,000 for 100 GB is ₹20 per gigabyte, 2.49 times the ₹8.02 blended realisation Indian operators actually achieved in FY26 — against a benchmark that is deflating by more than 10% a year.

A service priced at two and a half times a falling benchmark, with a five-figure upfront kit, has no mass-market path in India. That is not a marketing problem, and better satellites will not fix it. It is the arithmetic of a low-ARPU market served by a cost structure set globally.

Where the money actually is

The escape from that arithmetic is to stop selling to households.

Carrier backhaul is the clearest contract market in the country. In October 2026 BSNL issued a global tender for baseband and VSAT remote equipment to provide backhaul for 2,800 mobile towers and 750 Gram Panchayats under its 4G saturation programme, with an earnest deposit of ₹8 crore. That tender is a template: public connectivity money in India already flows to satellite ground equipment, it is awarded on capability rather than on consumer price, and the buyer is a balance sheet rather than a household.

Enterprise verticals pay for service, not for gigabytes. Offshore energy, mining, plantations, maritime fleets and aviation all need connectivity in places where no tower or cable will ever go, and all of them compare satellite against the cost of not operating. SES is positioning its O3b mPOWER medium-earth-orbit fleet for Jio’s 5G backhaul and enterprise business, with a shared gateway at Kadapa — and publicly frames itself as carrier-grade with service-level agreements, distinct from what it calls Starlink’s more consumer-grade proposition.

Community and institutional access spreads one terminal across many users. A school, a primary health centre or a panchayat office can run one connection across dozens of people. That is where per-head economics invert, and it is precisely where TRAI’s recommended terminal subsidy is aimed.

Own the channel, not the constellation. Within twenty-four hours in March 2025, both Bharti Airtel and Jio Platforms announced agreements with SpaceX to bring Starlink to Indian customers — Airtel first, on 11 March, and Jio the next day. Both deals put Starlink equipment in retail stores, both cover enterprise customers, and Jio additionally committed to a mechanism supporting installation and activation. Airtel separately holds an alliance with Eutelsat OneWeb, in which the wider Bharti group is a major shareholder; Jio holds a satellite licence of its own and plans roughly 1,600 satellites. The telcos own the customer relationship, the billing rails and the service organisation. In India that is the scarce asset — Jio counts 524.4 million customers and Airtel about 650 million across India and Africa. Starlink’s global subscriber base, by comparison, is around 10.3 million.

The pending framework, item by item

As of mid-October 2026, the queue looks like this.

• Spectrum basis: settled in principle — administrative assignment under the First Schedule of the Telecom Act 2023, no auction.
• Spectrum price: agreed at DCC level on 3 September 2026; Cabinet note drafted, approval pending.
• Authorisation regime: notified on 23 June 2026, with new applications required.
• Consultation: TRAI’s Consultation Paper 06/2026, issued 8 April 2026 on satellite network authorisation and spectrum assignment, with comments closing 6 May 2026.
• IN-SPACe authorisation: granted per constellation, five-year terms.
• Security clearance: the second stage, with the Ministry of Home Affairs, still outstanding for every licensee.
• FDI: up to 74% under the automatic route for satellite manufacturing and operation, satellite data products, and ground and user segments, with government approval beyond that, and 100% automatic for components.
• Ground segment: Starlink says twenty gateway sites have been built in India.

The hurdles, ranked

1. Security clearance is the binding gate. Everyone is at the same stage, so no commercial strategy can outrun it.
2. The price ceiling. ₹8.02 per gigabyte, falling 10.6% a year, is the number every satellite tariff must be argued against.
3. Per-head regulatory charges in a per-head-cheap market. ₹500 per urban subscriber per year is immaterial to an enterprise contract and fatal to a household plan.
4. Terminal cost plus duty. Hardware of ₹20,000 to ₹50,000, with personal courier imports facing roughly 30.98% in customs and integrated tax and commercial imports taxed at their tariff line, is a five-figure barrier in a market where entry-level devices are the growth engine.
5. Urban capacity physics. Ten to twenty thousand connections against fifty lakh of demand in one city is not a market-share question; it is a category question.
6. Incumbency and the public-interest frame. Government sources have said domestic companies would be encouraged so that India does not depend solely on foreign players, citing security, accountability and equipment sourcing. For a foreign constellation, the fastest route into India runs through an Indian balance sheet.
7. And structurally, India is a low-ARPU market with a global cost base. Satellite capacity is priced internationally. Indian willingness to pay is not.

What the evidence supports — and what to watch

The consumer ISP thesis for satellite in India is wrong, and the market data says so. The capacity-wholesale and institutional thesis is right, and it is where the only signed contracts currently point. The licence is not the moat; the security clearance, the gateway, the channel and the government contract are. And the first real satellite revenue in India will arrive on a corporate or government invoice, not a household bill.

Two triggers resolve the fog. The Cabinet order on spectrum usage charges sets the floor under every tariff. The first commercial tariff filing tells you what the operators believe Indian customers will actually pay. Until both land, treat any figure quoted for the Indian satellite internet market — including the ones in this article — as a framework, not a forecast.

References

Lok Sabha Unstarred Question No. 2902, Ministry of Communications, 5 August 2026 — GMPCS and Commercial VSAT CUG licences; Telecommunications (Authorisation for Provision of Principal Telecommunication Services) Rules, 2026. · PIB / Rajya Sabha, 2 April 2026 — India’s space economy and the Decadal Vision projections. · TRAI, “The Indian Telecom Services — Yearly Performance Indicators 2025-26” (PR No. 127/2026), 5 October 2026. · TRAI Consultation Paper No. 06/2026, 8 April 2026 — satellite communication network authorisation and spectrum assignment. · TRAI satcom recommendations (May 2025) and Digital Communications Commission decision, 3 September 2026; DoT Cabinet note on satcom spectrum charges, reported 11 October 2026. · Ministry of Communications statement, 8 October 2026 — status of security clearances. · Bharti Airtel press release, 11 March 2025; Reliance Industries / Jio Platforms media release, 12 March 2025. · Comptroller and Auditor General of India, Report No. 19 of 2026 — Performance Audit on BharatNet. · BSNL tender MM/RADIO/VSAT/T-892/2026, October 2026. · Union Cabinet amendment to the FDI policy on the space sector.