How advanced conductoring can help the USA win the AI race

By Jason Huang

The AI data center boom is creating unprecedented demand for electricity — and access to power has become the biggest constraint on America's AI race. The nation's grid simply wasn't built to deliver power at the speed or scale these systems require. But there's a solution that can unlock transmission capacity faster and more cost-effectively than any alternative: advanced conductoring.

The scale of the transmission challenge is already evident. In 2024, US data centers used 183 terawatt-hours (TWh), more than 4% of national electricity use — equivalent to Pakistan’s entire annual demand. By 2030, that consumption could surge 133% to 426 TWh. Already in Northern Virginia, the world’s largest data center market, wait times for electricity connection now stretch up to seven years. And across the nation, more than 2,600 gigawatts (GW) of energy and storage projects sit in interconnection queues — about twice the size of the entire existing grid.

As the United States Department of Energy’s (DOE) Speed to Power Initiative makes clear, transmission capacity — not just new generation — will define America’s competitiveness in the AI era. To understand how the US can meet this moment, it helps to look at the three strategies developers and utilities are pursuing today, and why only one can scale fast enough.

Three ways developers are working to meet the demand surge

Today’s AI-driven load growth has forced data center developers and utilities to rethink how and where they source power. Three approaches have emerged, each with strengths and some with more limits than others:

●     Solution A: Co-locating with existing generators. Some developers are siting facilities directly next to power plants, such as Microsoft reactivating a retired nuclear facility. This approach provides direct power without transmission bottlenecks, but opportunities are limited: It doesn't scale and can only serve a fraction of planned capacity.

●     Solution B: Bringing your own power. Others are developing data centers with dedicated power generation, including natural gas plants or utility-scale solar and wind. This offers more control and can accelerate timelines. But it’s only a partial fix, especially as data centers look to develop closer to major urban centers to reduce latency and get closer to where compute needs to happen.

●     Solution C: Connecting "here" and "there" through transmission. In most cases, data centers are in one place and generators are in another. The grid is already strained, and congestion will intensify as AI loads rise. Even when new generation is built, limited transmission capacity determines how much power actually reaches high-growth regions.

Why fast access to power has become the defining constraint

Data center demand is rising so fast that it’s reshaping electricity load profiles and exposing long-standing structural weaknesses in the grid. This accelerating growth explains why today’s developer strategies — co-location, dedicated generation, and long-distance siting — can’t fully keep pace with the scale or speed of AI-driven demand.

Power density inside facilities is exploding, with rack-level loads jumping from 8–15 kilowatt (kW) to 80–100kW+ or more. At the same time, over 80 GW of data center capacity is under development and could come online by 2030. This rapid buildout is colliding with grid congestion that has already pushed wholesale electricity prices in major data center regions up as much as 267% over the past five years. In the PJM electricity market, increased demands contributed to a $9.3 billion price spike in the 2025–26 cycle, adding an estimated $16–$18 to some monthly residential bills, while average US residential prices rose 6% year over year as of August 2025 than during the same month a year earlier.

Meanwhile, reliability pressures are mounting. DOE officials warn that blackout risks could rise sharply by 2030 if firm power sources retire faster than replacements come online. And with America’s transmission capacity expanding at less than 1% annually — far short of the 4–7% growth needed — the system is increasingly unable to move power to the locations where data center clusters are forming. 

At its core, solving these grid challenges comes down to solving a basic equation: new data centers need power, new generation must be built, and transmission must connect the two.

●     New data centers create massive new loads that need power. AI demand is overwhelming transmission systems faster than utilities can respond.

●     New generation must be built to meet skyrocketing demand. But generation alone doesn’t solve the problem if that power can’t reach high-growth regions.

●     Current and new power must be able to reach data centers. Both new-build transmission and reconductoring of existing transmission rights of way (ROWs) need to de-bottleneck the grid to support data center load growth and overall grid reliability.Solutions A and B solve only slices of this equation. Solution C — expanding transmission capacity — is the only approach that solves the full system challenge.

Why advanced conductors are the fastest, most scalable solution

If the US wants to build transmission capacity at the speed AI requires, it needs solutions that scale quickly, cost-effectively, and within the constraints of existing infrastructure. This is where advanced conductors shine.

A combination of new build (important but slow) and reconductoring existing lines (cheaper, faster, easier) will be required. But reconductoring is where the greatest near-term impact lies — and where advanced conductor technology (such as aluminum encapsulated carbon core, or AECC) provides.

There's no other viable alternative that unlocks so much new capacity so efficiently. AECC offers 2-3x the capacity of traditional ACSR while reducing line losses by up to half and limiting thermal sag at high temperatures. Much of this capacity can be unlocked without replacing or retrofitting existing structures, dramatically shortening timelines and minimizing environmental and permitting hurdles, lowering total project costs by 30–40%.

Advanced conductors deliver urgent “speed to power.” By expanding capacity on existing corridors, advanced conductors shorten data center interconnection timelines and strengthen grid reliability as AI load grows.

They improve the economics of new builds, too. Conductors make up just ~5% of project costs, yet they enable fewer and shorter structures — reducing steel, foundations, and labor. These efficiencies can lower total project costs by 10–20% while creating headroom for future growth.

And AECC delivers long-term systemwide savings. Nationwide adoption could cut consumer electricity costs by an estimated $2.2 billion annually through line loss reduction alone. Reconductoring with advanced conductors could enable 4x the interzonal transmission capacity expansion vs. greenfield new build transmission alone. That's massive and exactly what the grid needs in this acute moment of surging demand from AI data center buildout.

All these attributes make reconductoring with advanced conductors the best near-term way to unlock large-scale capacity. It provides the speed, the scale, and the cost efficiency utilities and developers need right now.

Meeting AI’s energy challenge with reliable, cost-effective power

Data center developers and electric utilities should be shouting from the rooftops to reconductor America’s transmission grid with advanced conductors. Copper may wire the racks inside data centers, but AECC aluminum needs to power them.

Advanced conductors are the only technology that can deliver transmission capacity fast enough and efficiently enough to keep America competitive in the AI era.

About the Author

Jason Huang is founder and CEO of TS Conductor, a US-based leader in award-winning, next-generation advanced conductors for the world's power grids. After earning his MS in materials science and engineering from UCLA and then both his MBA and PhD from The Ohio State University, Huang went on to roles with US government agencies such as NIST, to work on sensitive American defense aircraft like the F22 and F35, and to a series of technical and senior roles with companies such as Owens Corning, BAE Systems, Solvay, and CTC Global.