The "$12 Billion" Private 5G Story Is Backwards

A $12 billion market, and three of tech's biggest names placing "radically different bets" on it. That is the pitch running through the research headlines about private LTE/5G networks — the cellular networks a factory, warehouse, or port builds for itself instead of renting capacity from a carrier. On paper it is a growth story with a floor under it. So it is worth reading what the three companies actually did, because their behavior contradicts the headline.

AWS, Microsoft, and Nokia all moved the wrong way relative to the pitch. In May 2025, AWS retired AWS Private 5G, the managed cellular service it had sold for three years, and now routes customers to "Integrated Private Wireless," a program where carriers build the networks and AWS sells them the cloud. In 2025, reporting by industry analysts pinned the same retreat on Microsoft: it retired Azure Private 5G Core at the end of September 2025, pointing customers to migrate to Nokia and Ericsson. And Nokia, the one vendor that actually leads this market, used its November 2025 Capital Markets Day to push the mainstream business aside.

A market that refuses to become a product

The mechanism explains the retreat better than any single announcement. Private wireless looks like a software company's dream on a slide: take cellular out of the carrier, put it in the enterprise. In practice it refuses to become a product. Every deployment is a bespoke integration job — radio planning for a specific site, spectrum licensing (CBRS exists in the U.S. but has no direct equivalent almost anywhere else), small-cell siting, backhaul, and the physical work of mounting radios where they need to go. None of that cloudifies through a web portal. It is a high-touch, slow, one-off grind, which is exactly the economics a software business cannot love.

Nokia is the cleanest proof because it needs no convincing about the market. Outside China, Nokia holds more than half of private wireless. Yet when it reorganized last November, it filed its enterprise campus-edge unit — the factory-and-warehouse private 5G business — under "Portfolio Businesses," the holding pen for assets it will sell or spin off. The reason was in the numbers: that unit lost on the order of €100 million on roughly €900 million of revenue.

The market isn't falling short of $12 billion. It isn't in the same order of magnitude at all.

The tell is in the spread itself

The spread among the numbers circulating is the tell. A firm that actually counts deployments, Berg Insight, tallied about 4,700 private LTE/5G networks worldwide at the end of 2024, worth roughly $1.8 billion. Dell'Oro, which tracks radio equipment sales, put private wireless at only about 3–5% of total radio-access-network revenue. The report mills that sell forecasts call the same market $6 billion by 2027. Others push it to $12 billion, and the optimistic edge of the registry stretches to $312 billion by 2035. A range from $1.8 billion measured today to $312 billion projected in a decade isn't a forecast; it is a row of marketing pages with a price tag attached.

That is the durable lesson, and it is not specific to cellular. These totals are published by firms that sell the reports containing the totals, and the companies that write real checks are voting the opposite way with their capital. In the same stretch that the market was being advertised as racing toward $12 billion, both hyperscalers and the market's own leader quietly concluded the per-unit economics don't repay the integration grind. When the smartest operators retreat in the same year the market is described as accelerating, trust the retreat over the brochure.

The one real bet isn't private 5G

That is also why Nokia's divergence deserves a sharper reading. The stock trades near $10.60, up more than 60% in 2026 and more than double its 52-week low near $4.50 — a re-rating the market is awarding for the AI data-center pivot, not for private 5G. Nokia's enterprise revenue is growing about 18% year to date, but analysts attribute that to hyperscalers buying Ethernet and optical connectivity for data centers, a business Nokia bulked up last year with its Infinera acquisition, not to private 5G campus deals.

The three "radically different bets" the headline advertises are, at the strategic level, one bet: get out of the bespoke grind and follow the capital that actually scales. For an investor, that reframing is the useful takeaway. Nokia is now a wager on optical and IP data-center infrastructure — a real, addressable, engineering-scalable market, and also a claim about execution it still has to prove with Infinera integration. Private 5G, whatever the TAM slide says, is the niche it may be selling off.