Metro Connect 2026: Andy Lipman’s top 10 things to watch in a broadband market on fire

“I will be playing once again the role of Conan O’Brien today,” he quipped, welcoming delegates to the 25th annual gathering of what he called the “reunion of the leaders of the broadband industry”.

A veteran of MFS, MCI WorldCom and the early days of Level 3, Lipman has helped shape the programme for more than two decades. This year, against a backdrop of AI acceleration, record data centre demand and unprecedented public funding, he framed the agenda as a David Letterman-style countdown.

Here are his Top 10 things to watch in 2026.

A friendlier macro climate and renewed optimism

The “aspirational soft landing” appears to be materialising. Inflation is easing. Investors are broadly comfortable with interest rates. Broadband, Lipman reminded the audience, is not a short-term asset class.

Tax policy remains supportive, with bonus depreciation and R&D credits. The antitrust environment has improved, and the Administration is seeking to reduce environmental and permitting obstacles.

Broadband investment, he argued, is “fundamentally built on optimism” and optimism is returning. Metro Connect, traditionally an incubator for deals, may yet amplify it.

The $65bn question: BEAD and beyond

2026 is the year when the bulk of US broadband funding hits the market.

The $65bn Infrastructure Bill is being complemented by roughly $40bn more from Treasury, the FCC and other programmes. The “White Whale of BEAD”, as Lipman described it, is finally moving.

NTIA has approved close to $23bn in state awards covering approximately four million locations – just over $5,000 per location on average with a further $20bn expected to follow.

Among large operators, Comcast and AT&T secured the highest number of locations and funding totals. Fibre is the preferred technology in 41 out of 54 states and territories. It accounts for 65% of locations and 86% of funding, with satellite at 21%, fixed wireless at 12% and cable at 2%.

CAF, RDOF, E-Rate and other programmes add further billions. The capital injection is historic. The execution risk is equally significant.

AI, hyperscalers and the “trees grow to heaven?” moment

“How humungous is the AI opportunity?” Lipman asked. “Do trees actually grow to heaven?”

Hyperscalers are spending at an extraordinary pace. “If you’ve got the money, spend it” appears to be the working motto. Data centres are the most obvious beneficiary, but towers, fibre and power infrastructure are all in demand.

Hyperscale operators are pre-leasing capacity 24–26 months in advance. Enterprise demand remains robust, reacting to hyperscale buying patterns and tight supply. Data centre investment has become a major contributor to US GDP growth.

Key questions loom:

Will AI-driven efficiencies improve operational performance and EBITDA?
How much incremental volume will AI generate?
Which existing business models face disruption?

The growth is undeniable. The shape of the spoils is still being negotiated.

The labour bottleneck

Rapid fibre deployment and data centre construction have collided with a simple constraint: people.

The industry faces shortages of skilled technicians, contractors and materials, compounded by Buy American requirements. One client recently told Lipman there isn’t a fibre splicer in the country under 50 years old.

There is Department of Labor funding available for workforce training, yet hiring struggles persist. Expanding participation, particularly among women and minorities – remains both an economic and moral imperative.

The return of big M&A

“M&A and Metro Connect go together like peanut butter and jelly,” Lipman observed.

Scale continues to drive strategy. A new DOJ and FCC are expected to be more receptive to deals than their predecessors, though still serious about antitrust enforcement.

The industry may see a shift towards a blended wireline/wireless product market definition, potentially unlocking further consolidation. Strategic combinations across cable, fibre and wireless are already underway, and Lipman suggested that deals once deemed “unthinkable” may enter the conversation.

The consolidators, particularly national wireless operators, remain active. In broadband, scale still matters.

Integration or disintegration?

Will digital infrastructure players expand laterally across fibre, towers and data centres? Or will we see further divestitures and specialisation?

Vertical integration and cross-pollination are logical responses to AI-driven demand. Yet recent asset sales demonstrate that disintegration remains equally viable.

The direction of travel will say much about capital discipline – and investor appetite for complexity.

Is FTTH still red hot?

Fibre-to-the-home represents a $125bn US opportunity, perhaps larger than any other broadband segment. Nearly 65% of US households still lack access to fibre.

“People will give up the keys to their house and car, but not their broadband,” Lipman noted, quoting a forthcoming speaker.

FTTH remains “red hot”. Yet greenfield assets require significant capital before turning EBITDA-positive. Metro fibre has consolidated, but 75% of US multi-storey office buildings still lack fibre connectivity.

Further consolidation in 2026 appears likely.

Capital everywhere, and getting creative

Infrastructure funds and private equity firms continue to raise capital. Broadband is increasingly recognised as core infrastructure, expanding the universe of financial and strategic buyers.

Activity is picking up. Unallocated capital sits on the sidelines. The valuation gap between buyers and sellers is narrowing.

Recapitalisations, minority stake sales and ABS securitisations are all being deployed to refresh capital structures and fund growth. In Lipman’s view, creativity in financing is matching creativity in technology.

Power: The data centre constraint

AI needs data centres. Data centres need power.

By 2028, data centres could account for 7–12% of total US electricity consumption. A Department of Energy report projects an additional 50 gigawatts of demand by 2030.

States and utilities are experimenting with new rate classes. The Administration is accelerating grid connection for the new generation. Operators are exploring self-help measures: modular reactors, co-location strategies, behind-the-meter solutions and on-site generation.

Electricity is the new bottleneck. Solving it will determine the pace of AI expansion.

The populist pushback and the AI tsunami

Not everyone welcomes data centre growth. Opposition has emerged across the political spectrum, driven by rising residential electricity costs, up more than 30% since 2020, privacy concerns and perceived job displacement.

More than 20 projects representing over $100bn in revenue have reportedly been blocked nationwide.

Yet the industry is responding. Some operators have pledged to shoulder their own power costs. “We don’t think taxpayers should be asked to pay the burden, especially at the local level,” said one senior executive cited by Lipman.

Still, he ended on a bigger theme: the sheer scale of AI’s impact.

Morgan Stanley predicts global capital expenditure on AI infrastructure could exceed $3tn through 2028. The federal government has compared the AI race to the Manhattan Project.

Will AI be bigger than social media? Bigger than smartphones? “Finally,” Lipman asked, “will AI be bigger than fire?”

The only certainty, he suggested, is that Metro Connect 2027 will be even larger and more dynamic.

As ever, the conference aims to separate “reality from myth and hype”, to ask, in Lipman’s words, whether something is truly “Fibre Optics or Fibre Optics”.

Metro Connect USA: Tracking trends, deals & industry leadership