From Beauty Contest to Price Tag
On July 15, 2026, Japan's Ministry of Internal Affairs and Communications (MIC) formally certified JTower — a neutral-host tower-sharing company — to operate 26.8-27.0 GHz millimeter-wave spectrum across 13 metropolitan areas, including Tokyo's 23 wards, Yokohama, Osaka, Nagoya, Fukuoka, and Chiba. The certification closed out results from an auction MIC ran June 11-25, 2026, and announced June 30 — the first time Japan has ever assigned spectrum by price rather than by regulatory judgment (MIC press release).
For the six decades since Japan began licensing mobile spectrum, MIC allocated it through "comparative review" (比較審査): carriers submitted rollout plans, coverage commitments, and network-openness pledges, and the ministry picked winners by administrative judgment rather than willingness to pay. That system worked reasonably well when Japan had three or four carriers competing for broadly similar sub-6 GHz allocations. But as CMS's regulatory survey notes, MIC spent 2023 through early 2025 studying a shift to "conditional auctions" for spectrum above 6 GHz, originally targeting implementation by the end of FY2025 — a timeline that slipped roughly a quarter before the actual auction ran (CMS Expert Guide to 5G Regulation).
What the Auction Actually Produced
Five companies bid: NTT Docomo, KDDI/Okinawa Cellular, SoftBank, JTower, and High Tech Inter. For the nationwide 25.8-26.2 GHz block, Docomo prevailed after four bidding rounds at ¥6.288 billion — about 60% above the ¥3.93 billion floor price, beating out KDDI and SoftBank (Telecompaper). For the regional 26.8-27.0 GHz blocks, JTower won the major metro package for ¥468.71 million, while High Tech Inter — a much smaller regional operator — picked up sparser allocations in Hokkaido and Yamanashi for just ¥8.18 million (Dempa Times).
That spread is the real story. A four-figure premium for prime Tokyo-Osaka-Nagoya spectrum next to a rounding-error bid for rural Hokkaido is exactly what a price mechanism is supposed to reveal and a comparative review never could: the market's own estimate of where mmWave capacity is actually worth deploying. Under the old system, MIC would have had to guess that relative value through paperwork; the auction produced it as a number in three weeks.
The JTower Wrinkle Is the More Interesting Precedent
Docomo winning a nationwide license was the expected outcome. JTower winning a spectrum license at all was not. JTower's business has always been to lease tower and in-building capacity to Docomo, KDDI, SoftBank, and Rakuten, not to hold spectrum of its own — its own announcement calls this "the first domestic case of an infrastructure-sharing operator receiving frequency allocation certification" (JTower press release). Regulators built the regional-block structure precisely to let a neutral host or a regional entrant compete without staring down Docomo's balance sheet in the national round, and MIC's own framing of the 26.8-27 GHz band as slots for "new and regional operators" reflects that intent.
The Case for Comparative Review — Fairly Stated
Before declaring the auction an unambiguous win, it's worth taking the strongest objection seriously. Comparative review let MIC bake coverage obligations directly into the license: winners in Japan's 2019 5G assignments had to commit to building out base stations nationwide within five years, trading upfront revenue for guaranteed rural build-out. A pure price auction has no such lever — a bidder can pay for spectrum and deploy only where it is commercially convenient, leaving MIC to layer coverage conditions on afterward rather than screen for them upfront. Critics of auction-based allocation elsewhere have also warned that licensing costs get passed to consumers, and that auctions can favor deep-pocketed incumbents who can outlast smaller rivals on price alone even when the smaller rival's technical plan is better for a given region.
Why the Trade-Off Still Favors the Reform
That critique lands harder in dense sub-6 GHz bands, where every megahertz carries broad public traffic, than in 26 GHz millimeter-wave, whose short range and high-capacity profile make it inherently a dense-urban and venue-specific technology — exactly the deployment pattern the auction's own bid spread just confirmed without MIC having to mandate it. And the incumbent-capture worry cuts the other way here: JTower, a company with no history of holding spectrum, outbid a formal comparative process would have made it far harder to reach as a first-time spectrum licensee, since MIC's old criteria rewarded demonstrated national rollout capacity — the thing new entrants by definition lack.
Japan was one of the last G7 economies to adopt price-based spectrum assignment, and the amended Radio Act underlying this auction applies only to bands above 6 GHz, leaving sub-6 GHz comparative review intact for now. That's the right scope: use price discovery where deployment economics are genuinely local and heterogeneous, and keep administrative allocation where universal-service obligations matter most. MIC should track whether JTower's regional licenses translate into faster indoor and venue mmWave deployment than Docomo's national block — if the neutral-host model outperforms, that's the strongest argument for extending price-based regional carve-outs to the next mmWave tranche.