The data: growth at the margin, concentration at the core
TRAI's September 2026 subscription data, published as Press Release No. 140 of 2026, puts India's broadband connections at 1,107.68 million, up 6.47 million (0.59%) from 1,101.21 million in August, according to MediaNama's report of 27 October. The headline is dominated by mobile broadband, at about 1.039 billion connections. The more interesting line is fixed wireless. It reached 19.70 million connections, up from 19.34 million in August, and was the fastest-growing segment at 1.84% month on month. Within that, 5G fixed wireless access (FWA) rose from 13.45 million to 13.61 million.
The operator split of that 13.61 million is the policy-relevant fact. MediaNama's reading of the TRAI tables gives Jio 9.52 million and Airtel 4.09 million. Those two figures sum to the whole segment. The wider broadband market shows the same pattern: Jio and Airtel hold 83.96% of all broadband connections, against 83.95% in August.
Two caveats apply. First, TRAI began counting 5G FWA in the fixed wireless segment only in November 2024, so earlier comparisons are not like for like. Second, MediaNama notes that Airtel's machine-to-machine SIMs have been counted since December 2025, so some growth in the overall total reflects reclassification rather than new subscribers. The FWA numbers themselves are not affected by that note, but readers should treat month-on-month broadband growth with care.
The strongest case for the status quo
The incumbents' position has a respectable defence. Jio and Airtel paid for the 5G spectrum, built the networks, and took the risk of offering a fixed-line substitute before anyone knew whether households would buy it. India has few fixed-line connections relative to its population: about 48.89 million wired connections in a country of well over a billion mobile broadband links. Wireless is the fastest way to reach homes that fibre will not serve for years. If concentration is the price of speed, a regulator could reasonably accept it. Competition authorities also have fair grounds to worry that forcing artificial entry, through mandated sharing or price controls, would reduce the incentive to invest.
That argument holds for conduct rules. It is weaker for spectrum, which is a scarce public resource that the state allocates. The question is not whether incumbents should be restrained but whether the state's allocation rules shut out everyone else.
Why spectrum rules are the real competition policy
5G FWA runs on the same mid-band and mmWave spectrum that powers mobile 5G. A new entrant, or a weaker incumbent such as Vodafone Idea, cannot offer a competitive home-broadband product without enough contiguous spectrum at a price that makes the economics work. Vodafone Idea has 131 million broadband connections, per the MediaNama figures, yet none of the reported 5G FWA base. That gap suggests the constraint is not customer demand.
TRAI's February 2026 recommendations on the auction of IMT spectrum, released on 24 February 2026, are the pending policy vehicle. According to The Week's summary, they cover lower entry barriers for new players, a longer runway for the 600 MHz band and revised reserve prices. Secondary coverage reports that TRAI recommended auctioning the full available spectrum in 600 MHz and the existing bands. It also advised that parts of the upper 6 GHz range (6425-6725 MHz and 7025-7125 MHz) stay out of the next auction until the outcome of WRC-27 is known. I could not extract the text of the PDF directly, so the exact reserve prices by circle and band should be checked against the primary document. I also could not confirm whether the Department of Telecommunications has acted on the recommendations.
The design questions are concrete:
- Reserve prices. Prices set above what bidders will pay leave spectrum unsold, as happened with 600 MHz and 26 GHz in earlier rounds, and unsold spectrum is no help to anyone.
- Entry thresholds. Net-worth and eligibility rules decide whether a newcomer can bid at all.
- Block sizes and timing. Small, flexible lots let a regional or specialist operator bid for what it needs rather than a national bundle.
- Unlicensed or shared access. Keeping 6 GHz out of the auction preserves the option of licence-exempt use, which has powered cheap broadband in other markets, but only if the decision is made deliberately and not left to drift.
What a proportionate policy looks like
The pro-innovation position is not to cap Jio and Airtel. It is to make sure the state does not hand them durable advantages through allocation design. Four steps follow.
- Hold the auction without delay. Every month that spectrum sits idle, the segment grows inside a two-firm market.
- Price for use, not for revenue. Reserve prices should reflect demand evidence from earlier rounds. Maximising auction receipts is a legitimate goal, but not at the cost of a decade of weaker rivalry.
- Decide the 6 GHz question openly. If the band is held back pending WRC-27, the government should say when it will revisit it and consider licence-exempt use, rather than leaving it undefined.
- Track FWA as its own market. TRAI already publishes the data. A regular, public breakdown by operator and by rural and urban use, 7.06 million rural against 6.54 million urban in September, would let policymakers see whether concentration is easing.
The bottom line
One month's data is a small sample, and 5G FWA is still under 2% of all broadband connections. But the segment is growing faster than anything else, and it is already a duopoly. India's regulator has put forward a framework that, on the reported evidence, leans toward lower barriers. Whether that becomes reality depends on the Department of Telecommunications and the final auction rules, not on TRAI's data releases.