The real bottleneck in the energy transition: grids, storage, and policy

Deployment won. Integration begins.

After a decade defined by questions like “Will solar and batteries work at scale?”, the answer has arrived emphatically. Global solar PV capacity nearly hit 3 TW by the end of 2025, with around 698 GW added in that year alone. Storage is no longer a pilot story either: utilities, developers, and new chemistries are moving from press releases to shovels in the ground. And yet, bottlenecks are multiplying. The limiting factor is no longer the technology—it’s everything around it: wires, market rules, siting and permitting, and the ability to align explosive demand from data centers and electrification with variable, clean supply.

The recent news cycle underscores this pivot. India’s storage market is shifting from tender paperwork to execution. A Nevada utility is reportedly rerouting 75% of local supply to data centers, jolting 49,000 residents into considering solar-plus-storage as a necessity. In the U.S., a policy whiplash around the 30% residential solar tax credit is forcing homeowners and installers to race a new deadline. And a 500 MWh sodium‑ion deal in California hints at a broader push to diversify storage supply chains. The next phase of the transition is an integration challenge, not just a deployment story.

Solar’s scale flips the challenge

When the world adds nearly 700 GW of PV in a single year, system effects show up quickly:

• Curtailment rises: More midday megawatts collide with fixed grid capacity and inflexible demand. Regions from California to parts of Europe have seen frequent negative prices and growing curtailment as penetration climbs.
• Value cannibalization intensifies: Wholesale prices sink during solar peaks, eroding merchant revenues and stressing business models built on energy-only markets.
• Transmission becomes critical path: The best solar (and wind) resources are often far from demand centers. Without new high-voltage lines and stronger distribution networks, additional PV adds less incremental system value.

In short, adding more panels isn’t the hard part. Extracting their value when and where it’s needed is.

Storage is catching up—but the rules must reward flexibility

The storage learning curve is increasingly evident in the headlines:

• India’s pivot from tenders to execution: Market leaders there are emphasizing grid readiness, transmission build‑out, and long‑term supply resilience. That means less focus on headline procurement volumes and more on interconnection, siting, and operations.
• Chemistry diversification: Alsym Energy’s 500 MWh sodium‑ion project with Juniper, starting in California and expanding to other states, signals momentum behind non‑lithium options. Sodium‑ion generally trades lower energy density for abundant materials, thermal robustness, and potential cost advantages for stationary use—healthy attributes for a grid under strain.
• U.S. scale and pipeline: Grid‑scale battery capacity in the U.S. surpassed 16 GW in 2024 and is slated to more than double by 2026 based on project lists. But the growth is uneven, often stymied by interconnection queues, transformer shortages, and opaque revenue streams.

Storage can solve a lot—shifting solar from day to evening, providing fast frequency response, replacing gas peakers, and deferring wires upgrades. Yet many markets still price only a sliver of that value. Two fixes stand out:

• Reform product definitions: Align compensation with services that systems actually need (e.g., multi‑hour capacity products, flexible ramping, and fast frequency response), not just energy arbitrage.
• Normalize co‑location: Hybrid solar‑plus‑storage should see streamlined interconnection and settlement so developers can optimize operations without administrative penalty.

Grids: the new critical path

Transmission and distribution are where many clean megawatts go to stall. By the end of 2023, U.S. interconnection queues had swelled to well over two terawatts of generation and storage, with average timelines stretching to five years or more for complex projects. Similar bottlenecks exist in Europe and emerging markets.

FERC’s recent transmission planning overhaul (Order No. 1920) is a step forward: it requires long‑term, scenario‑based planning and clearer cost allocation. But even good planning needs execution capacity—steel, siting, labor, transformers—and social license.

On the distribution side, utilities face a dual challenge: electrification pushes loads up (EV charging, heat pumps), while behind‑the‑meter resources push flows in both directions. Hosting capacity maps, non‑wires alternatives, and modern distribution management systems (DMS/DERMS) can wring more headroom out of existing networks. Still, in many fast‑growing regions, there’s no substitute for upgrading substations, reconductoring feeders, and adding transformers.

Key priorities:

• Build regionally significant transmission tailored to resource basins (e.g., solar‑rich deserts, wind corridors) with standardized permitting milestones.
• Invest in advanced grid ops: dynamic line ratings, topology optimization, and grid‑forming inverters to enhance stability with fewer synchronous machines.
• Use targeted non‑wires solutions to bridge near‑term bottlenecks while large lines are built.

New loads are not just bigger—they’re peakier and pickier

Data centers epitomize the new load challenge. According to recent reporting, a Nevada utility is shifting 75% of available power for 49,000 Lake Tahoe residents to serve data centers, a drastic reallocation that makes resilience a household concern. Globally, the IEA has projected data centers could consume 620–1,050 TWh by 2026—on par with a mid‑sized industrialized nation or two.

Two things make this surge uniquely hard to integrate:

• Concentration: Compute campuses seek multi‑hundred‑megawatt interconnections, often in clusters, overwhelming local grids.
• Temporal rigidity—until it isn’t: AI training can be scheduled, but inferencing is more real‑time. Aligning both with variable renewables requires market signals and operator practices that barely exist today.

What would good integration look like?

• Make location a grid resource: Incentivize siting near surplus renewables and available transmission. Use congestion pricing to steer loads.
• Require flexibility as a condition of interconnection: On‑site batteries, thermal storage, demand response commitments, and the ability to modulate non‑critical compute.
• Treat backup as a grid asset: Pay data centers for firming and fast response if their on‑site storage and generators meet emissions and performance standards.

Residential electrification is evolving too. With the reported rerouting of power in Nevada, homeowners are treating rooftop solar and batteries less as green add‑ons and more as resilience infrastructure. Policy turbulence can amplify this behavior: despite moves by the current administration to repeal the 30% residential solar tax credit, some homeowners may still qualify under specific circumstances until July 4, 2026—a deadline now shaping installation schedules and supply chains. Policy stability matters as much as policy generosity.

Market design is lagging technology

If a system has abundant zero‑marginal‑cost energy at noon and scarcity at 8 p.m., energy‑only pricing will swing violently. That volatility doesn’t necessarily build the mix of assets we need. Three design gaps recur across regions:

1. Long‑duration signals: Four‑hour batteries are scaling, but systems also need 8–12 hours and seasonal flexibility. Absent clear products, these assets struggle to finance. Contract‑for‑difference (CfD)‑like structures or availability‑based payments can help.

2. Locational value: Congestion is local; so should be compensation. Granular nodal or zonal pricing, coupled with flexible interconnection and dynamic tariffs, helps DERs and storage show up where they matter most.

3. Performance‑based utility incentives: Reward utilities for procuring non‑wires alternatives, reducing curtailment, and integrating DERs cost‑effectively, not just for capital spend.

The integration playbook

• Build transmission faster and smarter: Standardize permitting timelines, pre‑permit corridors, adopt advanced reconductoring, and accelerate grid‑enhancing technologies. Use long‑term planning that explicitly values curtailment reduction.
• Treat flexibility as a first‑class product: Procure multi‑hour capacity, flexible ramping, and fast frequency response; stack values transparently for storage and demand response.
• Make data centers flexible by design: Require on‑site storage, thermal buffers, and demand modulation; pay for verified services; align interconnection queues with siting near surplus renewables.
• Modernize distribution operations: Invest in DERMS, dynamic hosting capacity, and non‑wires alternatives; standardize interconnection for solar‑plus‑storage and V2G.
• Diversify storage supply chains: Encourage sodium‑ion, flow batteries, and thermal storage alongside lithium‑ion to reduce material risk and expand duration options. The 500 MWh sodium‑ion deployment in California is an early marker.
• Stabilize policy signals: Avoid boom‑bust cycles around tax credits; when changes occur, set clear transition windows (like the July 4, 2026 eligibility path) to protect consumers and installers.
• Align finance with system needs: Use long‑term offtakes for storage and flexibility, not just energy; deploy public credit enhancements to derisk first‑wave long‑duration projects.

The bottom line

The clean‑energy boom has shifted from “Can we build it?” to “Can we integrate it?” Solar is racing toward three terawatts, batteries are scaling and diversifying, and demand from data centers and electrification is arriving in surges. The constraint is now policy design, grid capacity, and market architecture. Get those right—and the technology we already have will decarbonize faster, cheaper, and more reliably than most expect. Get them wrong—and we’ll keep adding clean megawatts that too often go unused at noon and missed at night.