The numbers are stark. US data center power demand stood at 366 TWh in 2025. By 2030, that figure is projected to nearly double to 728 TWh — a scale of load growth not seen since the 1980s.
Infrastructure demand is expected to climb from 75.8 GW in 2026 to 134.4 GW by the end of the decade. The US energy system is staring down a transformation of historic proportions.
But here is what the 2026 US Datacenters and Energy Outlook makes clear, and what too many headlines miss: the crisis is not about generation capacity. The US has enough power plants in the pipeline.
The crisis is about delivery. The transmission grid — the vast network of wires, substations, and interconnections that moves electricity from where it is generated to where it is consumed — is simply not built for what is coming.
For policymakers, this reframes the entire debate. Approving new solar farms and nuclear PPAs is necessary but not sufficient.
The real question — the one that will define whether the US data center boom strengthens or strains the national grid — is whether Washington and state regulators can modernise the infrastructure that connects generation to load. And the window to act is 2026.
“Just because we accelerate the processes upfront for generation and perhaps load interconnection, without sufficient grid and reliability upgrades to the system, we can’t actually operate.” — FERC Commissioner Judy Chang, January 2026
The 53-Month Problem
At the heart of the transmission crisis is a number that should alarm every policymaker who reads this report: 53 months.
That is the current national average time to process an interconnection request — the formal procedure by which a new power project gains approval to connect to the grid. More than four years. In a sector where demand is doubling in five, that is not a queue. It is a wall.
The interconnection backlog has become one of the most significant structural barriers in the US energy system. Thousands of gigawatts of generation capacity — solar, wind, gas, nuclear — sit in queues waiting for approval to connect.
Meanwhile, the grid they would connect to is itself ageing. An estimated 70% of US grid infrastructure is approaching the end of its operational life, even as it is being asked to carry unprecedented new load.
The result is a dangerous mismatch. Generation investment is racing ahead. Grid investment is not keeping pace.
And the interconnection process — designed for a different era — is the bottleneck through which everything must pass.
KEY FIGURES FROM THE REPORT
53 months — Average national interconnection processing time
728 TWh — Projected data center power demand by 2030
70% — US grid infrastructure approaching end of operational life
40+ GW — Solar contracted by hyperscalers in 2025 alone
~160% — Jump in gas generator interconnection requests (year-over-year)
FERC: The Pressure Point
The Federal Energy Regulatory Commission sits at the centre of this challenge. Its decisions on interconnection rules, transmission planning, and cost allocation will shape how — and whether — the US grid can absorb the coming surge.
The 2026 outlook identifies FERC as the critical regulatory variable, and recent developments confirm this assessment.
The Department of Energy has directed FERC to finalise a large-load interconnection rulemaking by 30 April 2026. This deadline is expected to attract substantive and contested commentary from across the sector.
The stakes are high: the rules that emerge will govern how large energy consumers — data centres chief among them — connect to the grid for years to come.
FERC Commissioner Judy Chang put the stakes plainly in January 2026: accelerating interconnection at the front end means nothing if the grid itself cannot handle the resulting load.
The physical infrastructure — transmission lines, substations, distribution networks — must be upgraded in parallel with the regulatory processes designed to speed new connections. One without the other fails.
The FERC rulemaking deadline of 30 April 2026 is not just a regulatory milestone. It is a moment that will define the grid’s capacity to absorb data centre demand for the next decade.
The Cost Allocation Battle: Who Pays?
Behind the regulatory process lies a question that is as much political as technical: who pays for the grid upgrades that data centre growth demands?
The answer is far from settled, and the 2026 outlook surfaces a fault line that policymakers cannot afford to ignore.
FERC’s current proposal adopts a 100% participant funding model. Under this approach, large load customers — including data centre developers — would bear the full cost of the network upgrades their projects trigger.
This represents a sharp departure from the traditional model, in which grid upgrade costs are socialised across the broader ratepayer base.
The consequences of this shift are already visible at the household level.
In Washington DC, Pepco residential customers saw monthly bills rise by $21 beginning in June 2025, with approximately half of that increase attributable to capacity market price increases driven by data centre demand.
As hyperscaler energy consumption scales, so does the pressure on ordinary ratepayers — unless regulatory frameworks explicitly address cost distribution.
This is not merely a fairness argument. It is a political sustainability argument. Grid modernisation requires sustained public investment and regulatory support.
That support erodes quickly if ordinary consumers see their energy bills climb to subsidise the infrastructure needs of trillion-dollar technology companies.
Policymakers who want durable grid reform need to design cost allocation frameworks that the public can understand and accept.
BYOP: A Symptom, Not a Solution
One of the most telling findings in the 2026 outlook is the emergence of Bring-Your-Own-Power (BYOP) strategies among data centre developers.
Facing long interconnection queues and uncertain grid capacity, hyperscalers and developers are increasingly exploring behind-the-meter generation — deploying mobile gas turbines, pursuing co-location with existing power plants, and structuring deals that allow them to operate independently of the public grid.
On one level, this is rational private sector problem-solving. On another, it is a warning signal that policymakers should take seriously.
When the largest energy consumers in the country begin opting out of the shared grid infrastructure, it raises fundamental questions about the long-term viability and equity of that infrastructure.
BYOP solutions may get individual data centres online faster. But they do not solve — and may deepen — the underlying grid crisis.
They shift costs onto remaining grid users, reduce the commercial incentive for grid operators to invest in modernisation, and fragment the energy system in ways that create new reliability risks.
The decisions being made now about generation, transmission, and distribution infrastructure are expected to endure for decades — potentially longer than the operational life of the data centres they serve — and will ultimately be paid for by customers, whether they are hyperscalers or households.
When the private sector starts building its own power plants to bypass the grid, it is not a market innovation story. It is a policy failure warning signal.
The 2026 Policy Window
The 2026 US Datacenters and Energy Outlook does not frame its findings as a prediction. It frames them as a choice. The load growth is coming.
The generation investment is following. What remains undetermined is whether the regulatory and infrastructure frameworks governing the grid can keep pace — and whether policymakers treat this year as the inflection point it is.
Several critical decisions converge in 2026. The FERC large-load interconnection rulemaking must be finalised by April.
FERC Order 1920, which established new long-term transmission planning requirements, faces compliance deadlines that will test whether utilities can translate rules into action.
The SPEED Act, advancing in Congress, proposes reforms to accelerate transmission permitting — but its passage and implementation remain uncertain.
The report’s implicit message for policymakers is this: transmission is no longer a downstream infrastructure problem. It is a first-order energy policy priority.
Without faster permitting, more investment in grid modernisation, and clearer frameworks for cost allocation, almost every other 2026 energy objective becomes harder — from integrating nuclear capacity to scaling renewables, from securing supply chains to maintaining grid reliability during peak demand periods.
The US has the energy. The generation pipeline is full. The technology is ready. The capital is willing.
What is needed now is the policy infrastructure to match — the regulatory reforms, investment frameworks, and political will to build a grid capable of carrying the energy economy of the next decade. That work cannot wait for 2027.
Also Read
Ames & Gough Warns of Rising US Construction Insurance Costs
US Construction Begins Race to Find 456,000 Workers by 2027
