Batteries Are Jumping the Data Center Power Queue

A new data center can be built in twelve to twenty-four months. Getting the grid to actually connect it now takes roughly three times as long. According to Lawrence Berkeley National Laboratory’s “Queued Up” research, the average US interconnection request took 55 months from application to operation in 2024, up from 36 months in 2015 and just 22 months in 2008. In PJM, the grid operator covering much of the mid-Atlantic, large data center projects that entered service in 2025 averaged more than seven years from application to power-on. For a buyer who has already signed a lease, hired a GC, and told a board the facility ships next year, that gap is not an inconvenience — it is the single biggest threat to the project’s return. A small but fast-growing group of operators has found a way to shrink it: put a battery behind the meter and start operating years before the utility finishes its own upgrades.

The Queue Math That’s Pushing Buyers Toward Batteries

The scale of the backlog explains why this is suddenly a board-level conversation rather than an engineering footnote. LBNL puts the total active US interconnection queue at roughly 2,290 GW, with more than 700 GW of that withdrawn in 2024 alone as projects gave up or got restructured. In ERCOT, large-load interconnection applications hit 198 GW in a single 2026 quarter, with about 86 GW still under active review. Queues in Northern Virginia and Silicon Valley, two of the densest data center markets in the world, are now running seven-plus years in some cases. None of this is abstract to a buyer: FERC’s 2026 large-load interconnection rules were written specifically because this backlog had become a national capacity problem, not a regional one. The rules help, but they don’t make a queue position you already hold move faster — which is why developers are increasingly looking for a way to get power without waiting for the queue to clear.

How a Battery Buys Years, Not Just Megawatts

The clearest public example is Aligned Data Centers’ Hillsboro, Oregon campus, where developer Calibrant Energy — backed by Macquarie Asset Management — installed a 31 MW / 62 MWh front-of-meter battery system under Portland General Electric’s Dispatchable Standby Generation program. The battery covers roughly 43% of the 72 MW facility’s peak load, letting the site draw less from the grid at peak and begin full operations years before a conventional substation upgrade would have cleared. Aligned CEO Andrew Schaap has said the arrangement brings the facility online “years earlier than would be possible with traditional utility upgrades.” The economics explain why operators are willing to take this on: delay on a facility this size can cost on the order of $80 million a year in lost colocation revenue, against a 62 MWh battery system that runs roughly $13-20 million to build. Advocates for the approach argue that math can produce a seven-to-tenfold return on the hardware alone, before counting any revenue the battery earns from grid services. A separate Portland-area program, GridCARE’s DeFlex initiative, is targeting 80 MW of incremental capacity at Hillsboro in 2026 and more than 400 MW through 2029 using similar flexible-connection logic — a sign this isn’t a one-off deal but an emerging category of utility program.

The Money Is Already Moving

Three independent market estimates published in the past week point the same direction. A Research and Markets report released October 15 sizes the AI data center energy storage market at $4.1-6.0 billion in annual revenue by 2030, up from roughly $1.2 billion in 2025 — a 28-38% compound annual growth rate — using Goldman Sachs’ estimate of 122 GW of data center power demand as its top-down anchor. Jefferies separately estimates hyperscaler BESS deployments could reach 20 GW cumulatively through 2035. A second Research and Markets report, published October 19 and covering the broader commercial and industrial storage market, forecasts the data-center-and-C&I battery segment growing roughly fivefold to $21 billion by 2036. Mercom Capital Group tracked more than $11 billion in corporate investment and nearly $3 billion in private equity flowing into the energy storage sector in the first nine months of 2025 alone. Buyers should treat these as directional, not precise — they come from vendor-adjacent research shops and press releases, and the underlying attachment-rate assumptions vary widely between the bear, base, and bull scenarios each report publishes. But directionally, the capital is clearly moving toward treating storage as core data center infrastructure rather than a backup afterthought.

Workload Volatility Is the Second Reason Batteries Are Showing Up

Interconnection speed isn’t the only driver. Joint research from NVIDIA, Microsoft, and OpenAI has documented that AI training and inference workloads cause rack-level power draw to swing from 30% to 100% of capacity within milliseconds — fast enough to stress transformers and trip protective relays that were engineered for the slower, steadier load profiles of traditional enterprise compute. A battery sized for interconnection speed can also absorb that millisecond-scale volatility, smoothing the swings before they ever reach the grid. That matters increasingly as more utilities adopt the kind of curtailment and flexible-load arrangements covered in our look at data center curtailment economics: a facility that can flatten its own demand spikes is a much easier counterparty for a utility to sign a large-load tariff with, and a much less risky one to prioritize in an interconnection queue.

What Buyers Should Not Skip Past

The caveats here matter as much as the headline numbers. Battery chemistry costs still vary enormously depending on who’s selling and how fast you need it: NREL pegs 2025 utility-scale four-hour lithium iron phosphate systems around $210 per kWh, while Wood Mackenzie’s post-tariff pricing for large front-of-meter systems runs $230-320 per kWh, and smaller commercial-and-industrial systems can run $250-450 per kWh. Sodium-ion chemistry is being pitched as a cheaper, non-flammable alternative with auxiliary power reductions of up to 97% versus some lithium systems, but the sector’s own recent history is a warning sign for buyers doing diligence on vendors: sodium-ion startup Natron Energy folded in 2025 despite real technical promise, and buyers evaluating any newer chemistry vendor should ask hard questions about balance sheet durability, not just cell-level economics. It’s also worth noting that batteries aren’t the only lever here — a December 2025 whitepaper from Princeton’s ZERO Lab, Camus Energy, and encoord modeled a 500 MW data center using flexible grid connection plus bring-your-own-capacity arrangements reaching full operation three to five years faster without a battery at all, saving an estimated $78 million per GW through flexible connection alone and $404 million per GW when combined with bring-your-own-capacity. For some sites, a software-and-tariff fix from the utility will beat a capital-intensive battery build; for others, particularly where the utility has no flexible-connection program on offer, a battery is the only lever a buyer actually controls.

None of this changes the fundamental fact that grid capacity, not chip supply, is now the binding constraint on how fast a buyer can turn a signed lease into revenue-generating capacity. A battery doesn’t get a buyer out of the interconnection queue — it lets them start earning revenue while they’re still in it, which for a large facility can be worth tens of millions of dollars a year. Before committing capital to a battery program, buyers should ask three questions of any site: what the actual queue position and realistic wait time are, whether the local utility has a flexible or standby-generation tariff program like PGE’s that would even allow a battery-backed interconnection, and what the vendor’s balance sheet and delivery track record look like beyond the press release. Get those three answers before signing, and a battery stops being a hedge against uncertainty and starts being a modeled, defensible part of the capacity plan. For a structured way to work through site and power risk before you commit, see the TechInfraHub data center buyer’s toolkit.

Written by

Raajeev Ratra

Data Center Infrastructure Expert | 15+ Years in DC Design, Operations & Project Management

Raajeev is a seasoned data center professional with hands-on experience in hyperscale facilities, colocation design, power & cooling infrastructure, and global DC operations. He shares practical insights to help engineers and IT leaders build better infrastructure.

Connect on LinkedIn →