BLS Says Fiber Jobs Are Shrinking. The Capex Says No [2026]

Signal vs. Noise

The Projection Says Declining: BLS Modeled the Last Build, Not This One

Two occupations. One agency. One projections cycle. Opposite answers to the same question.

By Scot Free

Every blueprint on this site runs on Bureau of Labor Statistics data. Median pay, annual openings, ten-year growth. It is the best public record of what the American labor market is doing, and nothing else comes close.

Which is why it's worth saying plainly when one of its numbers doesn't hold up.

The two projections

Telecommunications line installers and repairers (SOC 49-9052) — the people who pull and splice fiber:

  • 103,400 jobs in 2025, projected at 99,900 in 2035. A decline of 3%, or 3,500 jobs.
  • BLS reasoning: employment is expected to decrease as recent 5G infrastructure upgrades and government-funded broadband expansion projects are completed, reducing the number of workers needed for new installations.

Electrical power-line installers and repairers (SOC 49-9051) — the people who pull and splice power:

  • 131,900 jobs in 2025, projected at 145,500 in 2035. Growth of 10%, much faster than average.
  • BLS reasoning: the increasing use of artificial intelligence and the data centers needed to support it are expected to raise electricity demand; the national grid will need upgrades and new infrastructure, creating opportunities for these workers.

Same agency. Same 2025–35 cycle. Same page update date. The AI buildout is named explicitly as a growth driver for the people who string power lines to data centers, and it is absent from the outlook for the people who string fiber to the same buildings.

That is not a contradiction in the data. It's a modeling omission, and it has a signature: the telecom forecast is describing a build that is ending rather than the one that is starting.

What the capex says

  • AT&T committed $250 billion over five years to fiber, wireless, and satellite networks, and said it plans to hire thousands of new technicians in 2026. For scale, AT&T reported $20.8 billion in capex for all of 2025.
  • The footprint is still expanding fast. AT&T's fiber reached 32 million locations at the end of 2025 and is targeted at more than 40 million by the end of this year, on the way to 60 million by 2030.
  • The AI-specific build is separate from consumer broadband. Lumen had deployed 17 million intercity fiber miles at the close of 2025 and is on track for 47 million by the end of 2028, with an additional $2.5 billion in private connectivity contracts signed.
  • Consolidation is not contraction. AT&T acquired Lumen's consumer fiber business for $5.75 billion, adding roughly 1 million existing and 7 million planned locations. Verizon closed its acquisition of Frontier in a deal exceeding $20 billion. PwC's read: AI infrastructure is a second demand engine for fiber, with dense and long-haul networks critical to connecting the data centers powering AI, and operators building dark-fiber corridors between AI hubs.
  • The labor shortage is already being reported. Coverage this summer put the figure at roughly 66 million miles of new fiber-optic cable needed to connect data centers by 2029, with a shortage of the fiber splicers and cable technicians to lay it, while the U.S. remains short an estimated 58,000 skilled tradesworkers for rural broadband goals alone. The AI buildout and the rural program are competing for the same people.

A company does not commit $250 billion and announce technician hiring into an occupation it expects to shed 3,500 jobs.

What's actually true in both directions

The honest version is not "BLS is wrong." It's that the projection is answering a narrower question than the headline suggests:

  • BLS projections are built on industry employment models with long lead times. A capital commitment announced in 2026 does not show up in a projection published from earlier data.
  • The 5G and BEAD wind-down is real. Those programs did employ a lot of people, and many of those projects do finish. BEAD-approved work is expected to run into 2027.
  • A declining occupation still hires. Even with employment falling, people retire and leave, and those seats get backfilled.
  • Pay has not behaved like a declining occupation. The May 2025 median for telecom line installers is $74,330, well above the $50,980 median for all occupations.

So the seat isn't shrinking the way the headline number implies. It's changing employers. Fewer crews on subsidized rural broadband, more on long-haul and dense metro fiber between data centers.

The move

This is why a projection is a starting point and never the finish line. Run the same two-step on any occupation you're considering:

  • Read the reasoning, not the percentage. BLS publishes why in the Job Outlook tab. The reason tells you what the model is watching. If the stated reason is a program that is ending, ask what's starting.
  • Check the capex against the projection. Earnings calls, 10-Qs, and 8-Ks are public. When the people who sign the checks are describing a different decade than the forecast, believe the checks and watch the forecast.
  • Follow the employer, not the occupation. If the growth has moved from one customer to another, the job title stays the same and the job board changes.

The Scot Free Take

This site cites BLS constantly, and the way to earn that is to say so out loud when a number doesn't survive the check. Citing a source and auditing it are the same job.

The lesson isn't that forecasts are useless. It's that every forecast carries an assumption, and the assumption is usually printed right next to the number if anyone bothers to read it. Here, the assumption is that the last build was the build.

Putting a date on it: if the fiber-to-AI buildout holds, the next BLS cycle revises this one upward. That call is now on the record, with today's date on it.

Get the MoneyZoo LAB Report and instant access to The Side Door Playbook. One seat, one receipt, one move, every Monday. Subscribe here.

Sources

  • BLS, Occupational Outlook Handbook, Telecommunications Technicians (SOC 49-9052 projections and outlook reasoning), last modified Aug. 27, 2026.
  • BLS, Occupational Outlook Handbook, Electrical Power-Line Installers and Repairers (SOC 49-9051 projections and outlook reasoning), last modified Aug. 27, 2026.
  • BLS OEWS, May 2025 (telecom line installer median $74,330; all-occupations median $50,980).
  • Fierce Network, "AT&T unveils $250B investment to expand 5G, fiber, satellite network infrastructure," 2026.
  • Data Center Dynamics, "AT&T has appetite for more fiber, targets 40 million locations by end of 2026," 2026 (Stankey remarks; 2025 footprint).
  • Lumen Technologies Form 8-K and Q4/FY2025 results, Feb. 3, 2026 (17M intercity fiber miles, 47M by 2028, $2.5B PCF contracts, $5.75B AT&T transaction close).
  • PwC, US Deals 2026 Midyear Outlook, Telecommunications (fiber consolidation; AI as second demand engine; dark-fiber corridors; Verizon–Frontier).
  • RCR Wireless, "AI fiber build-out undermines rural BEAD skills-drive," Aug. 2026 (66 million miles by 2029 via NBC; 58,000-worker BEAD shortfall).

Corrections: ScotFree@TheMoneyZoo.com

Next
Next

The Door Just Opened: McKinsey Says 11 Million Need to Retrain. Pell Just Started Paying for It. [2026]