The fiber bet that nearly bankrupted CenturyLink is now wiring the AI economy. Up 140% in a year — and the turnaround just crossed the line that matters.

A near-death telecom reborn as the fiber layer of AI. The consumer business is gone; what’s left is the backbone.

Lumen Technologies — formerly CenturyLink — spent years as a debt-laden legacy telecom that an aggressive fiber overbuild nearly sank. Under CEO Kate Johnson (ex-Microsoft), it has been stripped down to one focused idea: be the high-capacity, low-latency fiber network that the multi-cloud, AI-driven economy runs on. The February sale of its consumer fiber business to AT&T for $5.75B made it a pure-play enterprise infrastructure company, and this quarter the strategic side of the business finally grew larger than the shrinking legacy side. The stock has responded — up ~140% over the past year.

Strategic revenue finally passed legacy. For a turnaround, that’s the moment the math flips.

For years the bear case on Lumen was simple: the new, growing business was too small to outrun the old, melting one. As of Q1 2026 that crossed over — strategic revenue hit 51% of total business revenue (up from 49% in Q4), officially surpassing legacy for the first time, and that was despite legacy holding up better than expected. Once the growing half is the bigger half, every quarter the mix works for you instead of against you. Q1 revenue of $2.90B beat the $2.83B consensus, adjusted EBITDA came in line, and management reaffirmed its full-year framework.

The asset that can’t be rebuilt. You can re-light fiber; you can’t re-dig America.

What separates Lumen from every other company with “AI infrastructure” in its pitch is the physical asset underneath it. Lumen owns the most extensive intercity (long-haul) fiber backbone in the United States — the routes between the metros, where the data centers are going. To replicate it, a competitor would need rights-of-way across thousands of municipalities, tens of thousands of miles of new trench, and permitting that takes 5–10+ years. The moat is the path in the ground, not the glass lit over it.

~$13 billion of fiber the hyperscalers already bought. And a digital platform to sell it on-demand.

The proof that the moat has value is the order book. Lumen has signed nearly $13 billion in Private Connectivity Fabric (PCF) deals — custom, dedicated fiber routes — with hyperscalers, neoclouds, social platforms, and AI companies who need guaranteed capacity between their data centers. These are long-dated builds: the cash comes in partly up front, but the revenue recognizes over future years, which is the key nuance for how it shows up in the model.

From Pipes to Platform

Lumen’s second leg is making that network programmable — sold on-demand through software (NaaS) rather than provisioned by hand. The data-center-interconnect (DCI) roadmap scales from 400G toward 1.6T to feed GPUs cost-effectively, and the company frames its pitch around “Cloud 2.0” metrics that matter to AI builders: time-to-first-token, GPU idle time, interconnect latency.

Still leveraged — but markedly less scary than it was. The AT&T sale changed the math.

Debt has always been the heart of the LUMN bear case, and it’s still the biggest risk — but the trajectory has genuinely turned. The $5.75B AT&T consumer-fiber sale closed in February, pulling net leverage below 4.0x, and management has been chipping at the structure: a new $825M revolver, a Level 3 refinancing, and roughly $300M of annual interest expense taken out. Free-cash-flow guidance was raised to $1.9–2.1B. None of this makes the ~$13.3B debt load small — but it moves Lumen from “solvency question” toward “manageable turnaround.”

Why the re-rating may not be over.

• Irreplaceable asset — the most extensive US intercity fiber backbone, the exact layer AI data centers need · 17M → 47M fiber miles by 2028.
• The mix inflected — strategic revenue passed legacy (51%) · from here the revenue mix works for the company.
• ~$13B PCF order book — hyperscalers and AI firms have already committed to the fiber · real demand, real contracts.
• Balance sheet turned — $5.75B AT&T sale, leverage below 4x, ~$300M less interest, FCF guidance raised.
• Alkira completes the platform — cloud-native NaaS bolt-on, capex avoided, TAM expands to ~$70B.
• Operating momentum — Q1 revenue beat, NaaS ports/services up 25–30% Q/Q, FCF guide up · execution is landing.
• Scarcity narrative — as AI moves from compute to connectivity, the “trusted network for AI” framing keeps Lumen in the AI-infrastructure conversation.

Why a +140% turnaround is a dangerous chase.

• It already ran — up ~140% in a year, near the 52-week high, and trading above the average analyst target (~$8.29) · you are not early.
• Still loses money on a GAAP basis — −$1.74B net loss, negative P/E · the FCF story is real but the income statement isn’t fixed.
• Total revenue is still shrinking — the mix flipped, but legacy decline still drags the top line until strategic growth outpaces it (guided ~2028).
• ~$13.3B debt — lower and cheaper than before, but still a heavy load that caps the equity’s margin for error.
• PCF is long-dated — the $13B is contracted over future years and front-loaded on cash, not near-term P&L · the revenue ramp is slower than the headline implies.
• Capital intensity — building toward 47M fiber miles costs money · capex and future financing remain part of the story.
• Execution + integration — the thesis needs flawless transformation and a clean Alkira close/integration · a stumble re-rates a momentum name fast.
• High beta — ~1.5 beta on a debt-heavy turnaround · it moves hard both ways on AI sentiment.

Own the thesis, respect the run. Disciplined, not euphoric.

The setup is genuinely better than it was in the spring — the mix inflected, the balance sheet improved, Alkira landed, FCF guidance rose. But the stock has done a lot of that work already (+140% in a year, +17% in a month, above the average target). That argues for treating LUMN as a thesis you accumulate on weakness and pullbacks rather than chase at the highs — and for sizing it as the leveraged, high-beta turnaround it still is, not a stable infrastructure bond-proxy.

What to Watch

The turnaround is working. The stock knows it. Bullish on the asset, disciplined on the entry.

Lumen has done the hard part. It survived the debt scare, sold the consumer business for $5.75B, pulled leverage below 4x, signed ~$13B of AI-fiber deals, watched strategic revenue finally pass legacy, and bought Alkira to complete the platform — all while raising free-cash-flow guidance. The asset is genuinely irreplaceable and sits in the right lane as AI shifts from compute to connectivity. But the tape has noticed: +140% in a year, near the 52-week high, trading above the average analyst target, still GAAP-unprofitable with ~$13.3B of debt and revenue that’s contracted for future years. That’s not a short — it’s a quality turnaround you want to own through the inflection, bought on weakness rather than chased at the high. The easy money was made on the way from $4 to $10. The next leg has to be earned one quarter at a time — strategic-share, PCF conversion, EBITDA. Own the backbone; respect that it’s run.