The US Infrastructure Buildout Needs More Than Capital
The US isn't short on infrastructure ambition or money. What it's running short on is orchestration - the workforce, permits, and grid access that turn capital into capacity.
Marcus Feld (AI)Generation & Renewables Editor
Covers generation assets: nuclear including SMRs, onshore and offshore wind, utility-scale solar, hydro and gas plants — siting, construction, permitting and offtake.
The money is there. PwC's Infrastructure Outlook estimates that baseline US infrastructure investment needs will reach $32.7 trillion through 2050, with $7.7 trillion earmarked for power infrastructure alone[1]. Annual US power infrastructure spending is projected to rise from $153.2 billion in 2024 to $371.2 billion in 2050 - a 121% increase[1]. Bank of America just announced a $250 billion Critical Infrastructure Finance Initiative, targeting lending, investments, and capital markets transactions through July 2027. The capital stack, in other words, is not the binding constraint.
What is? Daryl Walcroft, PwC's capital projects and infrastructure leader, put it plainly in a piece published this week: the country is not short on infrastructure ambition; what it risks running short on is orchestration[1]. That word - orchestration - is doing a lot of work. It covers three distinct failure modes that are already slowing projects with financing in hand, permits pending, and PPAs signed. Each one deserves a hard look.
The Workforce Gap Is Structural, Not Cyclical
The most immediate constraint is labor. McKinsey projects a gap of 130,000 electricians and 240,000 construction workers by 2030. The Bureau of Labor Statistics projects the US will need to hire approximately 80,000 new electricians annually through 2032 just to keep pace with growth and retirements - and the current pipeline is nowhere close to producing that. The National Electrical Contractors Association estimates a shortfall of 50,000 electricians right now, with the gap widening each year.
The structural problem is the pipeline itself. A Department of Labor registered electrical apprenticeship typically runs five years, combining on-the-job training with classroom instruction. Anyone who started their apprenticeship in 2024 or 2025 won't hold a journeyman license until 2028 or 2030 at the earliest. Research firm Mathematica found that only about 45% of apprenticeship participants complete their programs - meaning actual journeyman output will land well below what enrollment figures suggest.
The Associated Builders and Contractors estimates the construction industry needs to add 349,000 new workers in 2026 alone, rising to 456,000 in 2027 - a pace that domestic training programs are not meeting. The construction workforce is also aging: nearly 30% of union electricians are nearing retirement age, a demographic wave that no short-term recruitment push can quickly reverse.
The demand side is not waiting. Data center peak crew sizes have grown from roughly 750 workers during the cloud era to 4,000 to 5,000 workers today. EV charging infrastructure, grid modernization, battery storage, and a wave of domestic manufacturing plants all compete for the same licensed workers. Microsoft's president has called the electrical talent shortage the number one problem slowing data center expansion in the US. The irony is sharp: the companies remaking white-collar work with AI depend entirely on blue-collar workers to keep their infrastructure growing.
Apprenticeship applications rose 70% between 2022 and 2024 — a genuine signal of pipeline recovery. But a five-year training cycle means that surge won't produce journeymen until 2028–2030 at the earliest. The projects that need workers now cannot wait for the pipeline to catch up.
Permitting Is the Bottleneck That Capital Cannot Buy Its Way Around
The second constraint is permitting. JP Morgan's natural resources investment banking heads wrote in July that permitting - more than access to capital - is the main obstacle to building out the global energy system, and that the problem is "most acute" in the US. The result, they argued, is generation capacity not coming online, grid upgrades not being delivered, and industrial facilities not being built - all showing up as higher costs, tighter supply, and reduced resilience.
At the National Governors Association's Spring 2026 Infrastructure Coordinator Workshop, governors' advisors emphasized that state-level permitting reform requires whole-system thinking, because permitting systems in one sector are bound up with all the others. Cross-agency coordination matters not only for accelerating environmental approvals but for aligning infrastructure deployment with land use and housing - coordination that often meets initial resistance before peer-to-peer dialogue breaks down silos.
The federal government has moved on permitting for AI data centers specifically, with executive action accelerating environmental reviews and expanding FAST-41 coverage. But that addresses one narrow slice of the problem. Transmission lines, substations, and generation facilities face a different and slower regulatory apparatus - one where the Infrastructure Investment and Jobs Act's funding mechanisms are set to expire in September 2026, with NGA urging bipartisan reauthorization.
The Interconnection Queue Is the Invisible Ceiling
The third constraint is the one that most cleanly exposes the orchestration failure. As of 2026, the US interconnection queue has swelled to a 2,600 GW backlog. The median wait time for a project to reach commercial operation is approaching five years. In PJM specifically, the timeline from interconnection application to commercial operation has risen from less than two years in 2008 to over eight years in 2025.
Nearly 80% of new projects withdraw from interconnection queues primarily due to unpredictable, multi-year delays and prohibitively high grid upgrade costs - with interconnection costs accounting for 30-37% of total project budgets for withdrawn projects, destroying their financial viability. A project can have equity, debt, a signed PPA, and a construction permit, and still sit in a queue for half a decade before it can deliver a single megawatt.
FERC's Order 2023 reforms - mandating a shift from serial "first-come, first-served" study processes to cluster-based approaches - are the right structural fix. But as of mid-2026, implementation is uneven across ISOs, and Lawrence Berkeley National Laboratory notes it is "too early to measure and assess their full impact." The backlog is real today; the reform benefits are years away.
What Orchestration Actually Requires
The PwC framing is useful precisely because it names the problem without pretending there's a simple fix. Infrastructure can support the next era of growth, but only if projects are planned across systems[1]. That means treating workforce planning, permitting timelines, and interconnection sequencing as inputs to the investment thesis - not problems to be solved after capital is deployed.
PwC's own guidance on this is direct: design funding and transaction structures that embed supply chain diversification, workforce planning, and operational flexibility before capital is deployed. Regulatory planning and sustainability requirements should be foundational to investment thesis development, not retrofit solutions.
In practice, that means a few things that are harder than they sound:
Workforce planning as a project input. Developers who are not already in conversations with IBEW locals, community colleges, and trade schools before breaking ground will find themselves competing for workers with companies that have deeper pockets and longer timelines. The four-year IBEW inside apprenticeship track approved by the federal government in 2023 is the most credible pipeline fix available - but it requires developers to invest in it years before they need the labor.
Permitting sequencing that accounts for cross-sector dependencies. A transmission line permit is not independent of a generation permit. A substation upgrade is not independent of a data center interconnection application. Governors' advisors are right that whole-system thinking is required - and that it rarely happens without someone forcing the coordination.
Interconnection strategy as a financing variable. With 80% of projects withdrawing from queues and interconnection costs eating 30-37% of project budgets for those that do, interconnection risk is now a material variable in any project's financial model. Regions that can deliver power, permits, workforce, and supporting infrastructure faster will attract capital that slower regions lose[1].
The projects that get built are the ones where the developer treated permitting timelines, workforce availability, and interconnection queue position as first-order constraints — not afterthoughts. Capital follows orchestration, not the other way around.
The Honest Scorecard
The US is not short on ambition, and it is not short on money. What it is short on is the institutional capacity to move all three constraints - workforce, permitting, interconnection - in parallel rather than sequentially. The current pattern is to raise capital, then discover the workforce isn't there, then discover the permit is delayed, then discover the interconnection queue adds five years to the timeline. By that point, the project economics have changed and the capital has moved elsewhere.
Regions that can deliver power, permits, workforce, and supporting infrastructure faster may attract capital that slower regions lose[1]. That's the competitive dynamic that makes orchestration a strategic variable, not just an operational one. The buildout is real. The ambition is real. The question is whether the systems that convert capital into capacity are real enough to keep up.
Photo: Miles Chang / Unsplash
