Renewable Advisors — Energy Transition & AI Power Infrastructure Advisory

The Thesis

Why the energy transition moved to the data center.

"Twenty years in capital markets — most of it financing the energy transition, starting on Morgan Stanley's Clean Energy team. I've never been loyal to a fuel. I've been loyal to the transition. And the transition is now being won or lost at the data center." — Gary LaDrido, Founder & Principal

For two decades, the energy transition was a supply story — build more wind and solar, retire coal. We largely solved cheap, intermittent power. What we never solved was firm, around-the-clock, carbon-free power, because there was no buyer big enough or motivated enough to finance it.

Then AI arrived. U.S. electricity demand had been essentially flat for twenty years — efficiency cancelled out growth. AI broke that overnight. For the first time in a generation we have steep, concentrated load growth, and the grid can't build fast enough to meet it. Power, not chips, became the binding constraint.

Here's the part people miss: the hyperscalers are now the largest, best-capitalized, most committed buyers of clean energy on Earth. The company that will finance the next phase of the transition — nuclear, SMRs, long-duration storage, carbon capture, even fusion — turned out to be a data-center company. It's why Microsoft is restarting Three Mile Island and signing fusion offtakes, and why Google and Amazon are funding small modular reactors. The transition's next chapter isn't written by subsidies anymore; it's underwritten by AI demand. That's what it means to say it's being decided at the data center — and Renewable Advisors is built precisely at that intersection.

The logic — five beats

Supply-side → demand-side

The first twenty years made clean generation cheap. The unsolved problem was always demand big enough to pull firm clean power into existence.

Demand inflected for the first time in a generation

Two decades of flat U.S. power demand, then AI snapped the curve upward — steep, and concentrated in a handful of campuses. A structural break, not a cycle.

The constraint flipped from chips to power

Interconnection queues run years; transmission runs a decade. Whoever can deliver power at gigawatt scale, fast, controls the pace of AI itself.

The hyperscaler became the transition's banker

Trillion-dollar balance sheets, the deepest net-zero commitments, and an existential need for firm power — now the marginal buyer financing nuclear, SMRs, storage, and carbon capture.

The bridge is the strategy, not a compromise

Gas-backed, behind-the-meter power gets compute online today — and the cash flows, sites, and hyperscaler relationships become the platform on which clean firm power deploys tomorrow.

"The data center turned out to be the buyer the energy transition was always waiting for." I didn't change my thesis. I followed it to where it's now being decided — and to where capital formation, my craft, has the most leverage. — Gary LaDrido

Pillar 04 · Co-Investment

Co-Invest — direct exposure, pari passu with the GP.

We raise dedicated co-investment vehicles (SPVs) that invest alongside best-in-class general partners — typically 1–10% of a given allocation — giving family offices clean, fee-efficient access to individual deals on the same terms as the lead.

Aligned with the GP

Our SPV invests pari passu alongside the sponsor — same instrument, same economics, same timing.

Right-sized allocations

We take 1–10% of a deal's allocation, sized to complement the GP's process rather than compete with it.

Built for family offices

LPs gain single-asset exposure with transparency and control they don't get in a blind-pool fund.

Efficient structure

A purpose-built SPV per opportunity — streamlined governance, clear reporting, and aligned fee economics.

"A successful capital raise requires two ingredients: an exceptional value proposition, and an effective, rigorous and regimented process. We represent only the highest-quality opportunities and apply all of our energy, creativity and knowledge to deliver a successful outcome." — Renewable Advisors

The SPV placement process

An illustrative ~12-week timeline from mandate to funding. Creating demand and momentum accelerates the raise — competition for capital is at an all-time high.

Phase 1

Weeks 1–2

Define objectives

Review operations & performance
Preliminary valuation
Set strategy & timeline

Phase 2

Weeks 2–4

Prepare materials

Define investment points
Marketing deck & presentation
Target list; launch data room

Phase 3

Weeks 4–7

Go to market

Distribute teasers
Roadshows; data-room access
Preliminary indications

Phase 4

Weeks 7–10

Due diligence

Select on price & terms
Management presentations
Diligence visits

Phase 5

Weeks 10–12

Close

Evaluate bids & fit
Negotiate final offers
Execute & fund

Note: time periods are averages and can vary materially in any given transaction.

Discuss a co-investment

Pillar 05 · Flagship

AI Power Infrastructure.

Power has become the binding constraint of the AI era. RA works across the data-center and AI-factory value chain as an investor and financial advisor — and as the capital partner that raises dedicated SPVs and structures financing for the developers and projects driving the buildout. Our role is advisory — and, on the co-invest side, principal: we raise SPVs to invest alongside the developers we serve, never to compete with them.

"Power is the bottleneck of the AI era. I went where the transition is actually being financed — the data center."— Gary LaDrido

Investor

Raises dedicated SPVs to invest alongside developers and sponsors — pari passu — giving family offices direct exposure to AI-power assets.

Financial Advisor

Helps sponsors raise capital and structure transactions, drawing on deep capital-markets, M&A and private-placement expertise.

Development Capital

Partners with developers who have a financing need — raising purpose-built SPVs and structuring development and construction capital alongside them. We capitalize developers; we don't compete with them.

Strategic Solutions

Advises on the power, siting, and offtake constraints — structuring the right capital and connecting sponsors to power, offtake, and technology partners across our network.

Thesis · Why Now

A once-in-a-generation infrastructure super-cycle.

The U.S. is entering an infrastructure super-cycle comparable to the railroad and fiber-optic build-outs of prior eras — power-constrained, capital-hungry, and structurally re-rating the value of power and digital infrastructure. Early, well-positioned capital and development capability can capture outsized returns.

$32.7T

Baseline U.S. infrastructure need through 2050 (up to $42T desired)

+116%

U.S. data-center investment growth, $53.2B (2024) → $118.4B (2027)

$371B

Annual U.S. power infrastructure investment by 2050, up from $153B

~87%

Of forecast power investment is renewables plus transmission & distribution

Source: PwC — Global Infrastructure Trends. Figures are estimates for illustrative purposes.

Market Size

One of the largest infrastructure build-outs in history.

Analysts now headline the global data-center and AI build-out as a multi-trillion-dollar race through 2030 — and power is roughly a quarter of every dollar spent. That is the slice RA is built around.

~$7T

Global data-center capex by 2030 — the "$7 trillion race" (McKinsey)

~$2.9T

Global data-center capex, 2025–2028 (~$1.6T hardware + ~$1.3T infra) — Morgan Stanley

~25%

Of AI spend goes to power "energizers" — generation, transmission, cooling

$5.3T

Hyperscaler AI & data-center capex, 2025–2030 (Goldman Sachs)

A decade of accelerating need — 2026 to 2036

Estimated annual global AI & data-center infrastructure investment, with the power share highlighted.

Total AI & data-center investment Power infrastructure (~25% — "energizers")

Illustrative projection blending published estimates from McKinsey, Morgan Stanley, Goldman Sachs, and JLL; power share derived from McKinsey's ~25% "energizers" allocation. Cumulative forecasts are routinely revised upward. For information only.

Headline estimates compared — by firm

Cumulative spend (note differing scopes & timeframes); the green segment marks the ~25% power share.

Our Perspective

Where we think the AI power market is actually being decided.

Four theses that frame how we underwrite, advise, and invest. Click to read.

The Power Constraint Thesis

Why power — not chips — is the real bottleneck on AI.

▾

Our proprietary angle

The popular story is that AI is constrained by compute. On the ground, it is constrained by electrons. Interconnection queues in the major U.S. markets now stretch for years, large-transformer lead times have blown out, and new transmission cannot be permitted and built on an AI timeline. Meanwhile a single hyperscale campus can require hundreds of megawatts — increasingly gigawatts — of firm, 24/7 power. The gating item is no longer the chip order; it is energized capacity at the rack.

That inversion changes how infrastructure should be valued. Power that can be delivered quickly — behind-the-meter and on-site generation that bypasses the interconnection queue — commands a scarcity premium over power that is merely cheap but years away. Speed-to-power becomes the differentiator, and the projects that solve it capture the economics.

The bottleneck is therefore also the opportunity. We underwrite projects on the three things that actually de-risk delivery: secured power, interconnection position, and efficiency. Capital and development capability aimed squarely at the constraint is where outsized, durable returns are made.

Offtake as Infrastructure

The contract — not the concrete — is the new unit of financing.

▾

Capital markets × infrastructure

In traditional power, the power-purchase agreement was what made a project bankable: a long-dated, creditworthy contract that lenders and equity could underwrite. In AI power, that role is being filled by the offtake between developer and hyperscaler — long-tenor, investment-grade counterparty, often take-or-pay. A signed hyperscaler offtake converts a speculative megawatt into a contracted, financeable cash flow.

Once the offtake is the financeable asset, the work becomes capital-markets work: structuring the contract, allocating construction and operating risk, tranching the cash flow across project debt, structured equity, and credit, and placing each tranche with the capital that prices it best. The physical build matters, but the contract is what gets financed.

This is precisely where our advisory sits — translating an offtake into a capital structure, and a capital structure into committed capital. It is the intersection of the capital markets and the infrastructure, which is exactly the seam RA was built to work.

Market Map: Who's Financing What

A field guide to who is deploying capital across AI power.

▾

Market intelligence · updated periodically

The capital base for AI power is crowded and moving fast, and the players do not all want the same risk. Hyperscalers (Microsoft, Google, Amazon, Meta, and now Oracle and xAI) are securing power directly — corporate PPAs, nuclear restarts, and small modular reactors — to lock in firm supply. Developers and IPPs are building the gas, renewables, and storage that sit behind the meter for those campuses.

Utilities are racing to expand interconnection and transmission, the binding constraint on time-to-power. Infrastructure and credit platforms — KKR, Brookfield, Blue Owl, and the large debt funds — are financing platforms and portfolios at scale. And a distinct nuclear cohort, from restarts to SMRs, is being underwritten for the firm, 24/7 profile AI training loads demand.

Knowing who is doing which deal, on what terms, and where the capital is moving next is an advantage in itself. Tracking that map — and the relationships behind it — is core to how we serve developers, GPs, and family offices.

The Behind-the-Meter Shift

Why data centers are leaving the grid — and what it means.

▾

Explainer

With grid interconnection a multi-year wait, developers are increasingly self-supplying: on-site gas turbines and engines for speed, paired with solar, storage, fuel cells, and — over time — clean firm options including nuclear. industry estimates (Jefferies, citing McKinsey) put roughly 25–33% of net new data-center generation as met behind the meter through 2030, and project trackers already show a quarter or more of planned U.S. capacity carrying BTM plans.

The implications ripple outward. It reshapes renewable procurement — the choice between corporate PPAs and on-site generation. It changes grid planning, as load first defects from the grid and then, in many cases, reconnects in hybrid configurations. And it reshapes the fuel mix: gas wins on speed today, with clean firm power gaining as it becomes deliverable.

For the energy transition, BTM is both threat and opportunity — it can sideline some grid-scale renewables while creating large new demand for on-site clean power and storage. We help developers and capital partners navigate that shift on the right side of it.

Sector Focus

Infrastructure sectors powering the AI build-out.

Data Centers & AI Factories

Hyperscale campuses and edge facilities — the core demand driver of the cycle.

Power Generation

Gas, nuclear and on-site generation to meet the ~150% projected rise in electricity demand.

Renewables & Storage

Solar, wind and battery storage — ~87% of forecast power investment alongside T&D.

Grid & Transmission

Interconnection and T&D capacity — the binding constraint on time-to-power.

Cooling & Critical Infra

High-density cooling, water systems and critical digital infrastructure.

Digital Infra Technology

DCIM, grid-edge control and permitting software that de-risk delivery.

Development Capital

Capital for the developers building AI campuses & edge.

▸

Hyperscaler campuses

We raise SPVs and structure development and construction capital for sponsors building large-scale, power-secured campuses — funding land control, power procurement, and interconnection through to energization.

▸

Edge data centers

We capitalize distributed, latency-sensitive projects — paired with on-site generation, storage, and grid-edge solutions — alongside the developers delivering them.

▸

Backing time-to-power

We prioritize projects that solve power, siting, and delivery — the binding constraints — and bring the capital that gets them online.

Capital across the value chain

We arrange and place capital at every stage — development, construction and operating — from early site control through stabilized, income-producing assets.

Through SPVs, we invest alongside leading developers, operators and capital providers, structuring equity, structured equity and project-level capital, and underwrite the power constraint directly: prioritizing projects with secured power, interconnection and efficiency advantages.

Strategic Solutions

Solving the power, siting & efficiency constraints.

Beyond capital, we help sponsors tackle the operational bottlenecks — connecting projects to the DCIM, grid-edge control, and energy-optimization partners that improve efficiency and speed interconnection, and structuring those solutions into the deal. The kinds of outcomes these solutions can unlock:

15–30%

Cooling-cost reduction

10–20%+

Compute-energy savings

~60 days

Typical deployment, no rip-and-replace

$100K–$1M

Per-site monthly cost of interconnection delay addressed

Illustrative outcomes from operating-technology solutions we help source and structure; results vary by site.

IPPsBehind-the-meterHyperscale campusesEdge data centersGrid & transmissionOn-site / hybrid generationEnergy storageDCIM & grid-edge control

Discuss financing for your project

On Brand · Decarbonization

Decarbonization: from NIMBY blocker to deal enabler.

This is core Renewable Advisors territory. Local resistance to data centers usually centers on land, water, grid strain, and emissions. A credible path to low- or zero-carbon operations reframes a project from community burden to climate catalyst — easing permitting, building goodwill, and unlocking incentives. It's also where renewables expertise creates the most value in the AI buildout.

~460 TWh

Global data-center electricity use in 2024 (IEA)

1,000+ TWh

Projected by 2030 — more than double (IEA base case)

~1,300 TWh

Projected by 2035 — the demand renewables can help meet

24/7 CFE

Carbon-free matching — the new hyperscaler procurement standard

How decarbonization counters NIMBY

▸

Environmental framing

Opposition cites fossil-fuel lock-in, water use, and emissions. Committing to 24/7 carbon-free matching (renewables + storage or nuclear), advanced cooling that cuts water, and measurable efficiency gains answers those concerns head-on.

▸

Policy & community leverage

Several states tie tax incentives or approvals to renewable procurement or efficiency targets — Virginia lawmakers, for example, have proposed conditioning data-center tax exemptions on clean-energy commitments (2026 session, unresolved) — and dispute settlements have required 100% renewables and recycled water.

▸

Burden → catalyst

Position the data center as the demand that finances new renewables and grid upgrades. That shifts the local narrative from "burden" to "catalyst for the clean-energy transition."

Where renewables players win

Five opportunities in the AI power race.

Impact

Renewables PPAs & co-location

Hyperscalers are signing large renewable deals and exploring behind-the-meter or co-located solar/wind + storage. Solar-plus-storage pairs especially well with data centers' relatively flat load profiles.

Impact

Efficiency as a multiplier

Cooling is a large share of energy use; liquid cooling and AI-optimized operations deliver big gains. Improving PUE and server utilization cuts total power needed — making every renewable MW go further.

Impact

Hybrid firm power

Pair renewables with nuclear or gas + carbon capture (CCS) for true 24/7 low- or zero-carbon supply: renewables handle daytime peaks, firm sources cover the rest.

Demand flexibility & grid services

Data centers can offer shiftable or curtailable load and ancillary services — helping integrate more renewables onto the grid and earning value for it.

Siting & design

Locating where renewables are abundant or climates are cooler, with water recycling and reuse, reduces both emissions and local resource conflict.

A portfolio approach

Renewables + efficiency + firm clean power together cut emissions and water impact dramatically versus business-as-usual.

Carbon credits

Revenue, not just PR.

▸

High-integrity offsets & removals

Operators can buy quality credits or removals to address residual Scope 2/3 emissions in support of net-zero — or carbon-negative — goals.

▸

Removal built into infrastructure

Emerging models tie biogenic or engineered removal directly to operations — moving beyond paper offsets toward verifiable, high-quality credits that appeal to investors and regulators.

▸

Bundling opportunity

As scrutiny on additionality and permanence rises, renewables developers can bundle credits with clean-power supply or participate in linked markets.

Government incentives & tax credits

Clean Electricity PTC / ITC (45Y / 48E)Technology-neutral credits for renewables, storage, and clean generation; transferability allows monetization even with limited tax appetite.

Zero-Emission Nuclear PTC (45U)Supports nuclear restarts and new builds that can power data centers.

Carbon Sequestration Credit (45Q)Valuable for gas + CCS and direct air capture paired with data centers — helping create clean-firm power.

Efficient Buildings Deduction (179D)Deductions for efficiency upgrades — advanced cooling, controls, and more.

State & DOE programsData-center tax exemptions (often tied to clean-energy or jobs), plus DOE loans and grants for clean infrastructure. Credits are stackable and have created active transfer markets.

Practical takeaway: position around additionality and measurable decarbonization — new renewable capacity directly attributable to data-center demand, or efficiency projects with quantified savings. Bundle PPAs with credit generation and efficiency services, and use decarbonization to clear NIMBY hurdles and unlock incentives — making projects more financeable and approvable.

Sources: IEA, BRG, and U.S. policy documents. Tax-credit provisions (including IRA credits) have been modified by recent legislation; this is general information, not tax or legal advice — confirm current eligibility with qualified advisors. Notably, the 2025 One Big Beautiful Bill Act phased out or restricted several of these — ending the 45Y / 48E credits for new wind and solar after 2027 and the 179D deduction after mid-2026.

Opportunity Map

Picks & shovels: the AI-adjacent buildout.

Hyperscalers are spending hundreds of billions a year, and much of it flows not to chips or models but to the enablers that make data centers buildable and operable at scale. Demand is structural, multi-year, and protected by specialized expertise and long equipment lead times — and several of these sectors carry a real efficiency or resource-impact angle.

Impact · efficiency

Cooling & Thermal Management

Liquid, immersion, and CRAC cooling and heat-rejection systems — demand is exploding as air cooling hits its limit on high-density AI racks. A premier picks-and-shovels play, and a direct lever on energy and water efficiency.

Examples: Vertiv, nVent, HVAC specialists

Impact · efficiency

Electrical Equipment & Power Distribution

Switchgear, UPS, busways, PDUs, internal-distribution transformers, and enclosures that move massive power through the facility — efficiency here cuts losses across the whole load.

Examples: Eaton, nVent Electric, Schneider Electric

Construction & EPC

Site prep, foundations, mechanical and electrical installation, and turnkey builds — the civil, mechanical, and electrical contractors delivering campuses on AI timelines, often with record backlogs.

Examples: Sterling Infrastructure, EMCOR, Comfort Systems USA

Networking & Interconnects

High-speed Ethernet switches, optical transceivers, cabling, and SerDes for rack-to-rack and intra-rack communication inside AI clusters.

Examples: Arista, Broadcom, Marvell, Astera Labs, Credo

Real Estate · Colocation · REITs

Land acquisition, campus development, and leasing of powered shell or turnkey space to hyperscalers — frequently on long-dated, investment-grade leases.

Examples: Equinix, Digital Realty, private developers & REITs

Servers, Racks & Enclosures

Custom AI-optimized servers, GPU racks, and structural components supplied to hyperscalers and ODMs across the hardware stack.

Examples: ODMs & rack / enclosure suppliers

Specialized Semiconductors

Beyond the main GPUs: power semiconductors, high-bandwidth memory (HBM), custom ASICs, and advanced packaging that support the full stack.

Examples: Power-chip makers, foundry / OSAT suppliers

Fiber Optics & Connectivity

High-speed fiber and transceivers moving data between and within campuses — benefiting from both the core buildout and edge / interconnect needs.

Examples: Fiber networks & transceiver suppliers

Impact · water

Water Management & Treatment

Recycling and treatment systems for water-intensive evaporative and liquid cooling — a growing regulatory and efficiency focus, and a clear resource-impact opportunity.

Examples: Water treatment & recycling specialists

Specialized Manufacturing

Corrugated hoses, metal fabrication, modular components, and backup systems — traditional industrials pivoting to data-center demand, often with reshoring upside.

Examples: Industrial & modular-component manufacturers

Why it matters: these enablers offer more defensive, structural exposure to AI than volatile chip or model names — essential infrastructure that gets built regardless of the hype cycle. Cooling, electrical, and construction have been standout performers, and the efficiency and water angles make several of them genuine impact investments.

Company names are illustrative examples of each sector, not investment recommendations or an indication of holdings.

Behind-the-Meter · Clean Firm Power

Nuclear: the clean firm answer to AI's power problem.

Of every power source in the AI buildout, nuclear sits closest to the energy transition's original promise — carbon-free, high-density, dispatchable baseload, ideal for co-location with data centers. After decades on the sidelines, it is being pulled back to the center by hyperscaler demand, deregulation, and national-security urgency.

~90%+

Capacity factor, vs. ~25–40% for wind & solar

>10 GW

Big-tech nuclear contracted or targeted in roughly the past year

First-ever

U.S. reactor restarts from retirement — Three Mile Island & Palisades — the fastest path to firm power

10–20+ yr

PPAs underwriting new and restarted capacity

Why nuclear fits the load

▸

Carbon-free baseload

High-density, dispatchable 24/7 power that aligns directly with hyperscalers' net-zero commitments — the "clean firm" power the transition never financed.

▸

Small footprint, co-locatable

Far less land per MW, and the option of behind-the-meter or dedicated supply that reduces grid dependency and transmission loss.

▸

Revenue certainty

Long-dated PPAs with investment-grade hyperscaler counterparties de-risk development and make new capacity financeable.

Recent big-tech nuclear deals

Microsoft × Constellation20-yr PPA for ~835 MW from restarting Three Mile Island Unit 1 (Crane Clean Energy Center), targeted 2028 — Microsoft takes essentially all output.

Google × Kairos PowerPPA for up to 500 MW from SMRs (Hermes 2 design), first units targeted ~2030; TVA deal makes it the first U.S. utility PPA with an advanced reactor.

Amazon$650M data-center campus beside the Susquehanna plant, plus backing X-energy for up to 5 GW of new SMRs by 2039 — the largest SMR deployment target to date — with partners including Energy Northwest and Dominion.

MetaFollowing its 1–4 GW new-nuclear RFP, has locked up to ~6.6 GW via deals with TerraPower (up to eight Natrium reactors), Oklo, and Vistra (existing-plant extensions).

OracleDesigning a 1 GW+ data center powered by a trio of SMRs, with building permits already secured.

RestartsPalisades (Michigan, ~800 MW) — the first U.S. plant to restart from decommissioning, backed by a $1.52B DOE loan guarantee, targeted for early 2026.

Driver 01

Deregulation

Executive orders are accelerating federal permitting for AI data centers and energy infrastructure — qualifying projects (e.g., >100 MW incremental load or >$500M capex) get expedited review and potential federal support. Add NRC reform, co-location on DOE sites, loan guarantees, and faster SMR licensing, plus active state incentives (Texas, New York).

Driver 02

National Security

AI leadership is framed as economic and military priority — and reliable, scalable power is its foundation. Advanced nuclear is designated strategic infrastructure, with certain data centers and reactors tied to defense. For institutions, that creates a policy moat around domestic nuclear and data-center assets.

Driver 03

The Investor Lens

Long-term PPAs provide revenue certainty; co-location cuts transmission losses and speeds deployment; nuclear becomes a scarce, high-value resource. Risks remain — high capex, long new-build lead times, NIMBY, supply chain, and SMR execution — but restarts shorten timelines and policy support is accelerating.

Bottom line: nuclear is emerging as the linchpin of reliable, low-carbon power for the AI buildout. Deregulation is shortening timelines and de-risking projects; national-security framing elevates it beyond pure commercials. The opportunity spans the value chain — developers, fuel, components, and restarts — and rewards careful diligence on timelines, offtake security, and policy durability.

Sources: IEA, company announcements, and U.S. policy documents (2025–2026). Deal terms and timelines are as announced and subject to change.

Behind-the-Meter · Storage

BESS: turning constrained power into dispatchable certainty.

AI clusters create highly variable, high-density loads — rack densities of 30–100+ kW with rapid swings. Traditional UPS only rides through for seconds. Battery Energy Storage Systems (BESS) are becoming table stakes for power-certain AI infrastructure, converting intermittent or constrained power into reliable, dispatchable supply — directly affecting whether a campus gets financed, permitted, and leased.

30–100+ kW

Per-rack densities in AI clusters, with rapid power swings

99.999%+

Uptime BESS helps secure — beyond short UPS ride-through

1.5 GWh

Sodium-ion supply deal targeting AI data centers (Energy Vault × Peak Energy)

10+ GWh

BESS secured by Texas developer Greenflash Infrastructure for data centers

Market context & growth drivers

U.S. and global BESS deployments are hitting records, with AI data centers among the fastest-growing segments alongside utility-scale. Three themes recur across the research:

Power certainty

Millisecond response for GPU load swings, backup autonomy during outages, and grid support — the reliability AI workloads can't run without.

Hybrid models

Frequently paired with gas, renewables, or emerging long-duration storage (LDES, 100+ hour duration) to meet true 24/7 needs.

Market tailwinds

Faster interconnection, revenue from grid services, and policy support — though some clean-energy incentives now face change.

Why BESS matters for AI power

▸

Fast-response backup & ride-through

Instantaneous backup during grid outages or generation shortfalls — essential for the mission-critical uptime hyperscalers demand.

▸

Load smoothing & peak shaving

Buffers the extreme, rapid power fluctuations GPUs create during training and inference, protecting both equipment and the grid connection.

▸

Grid services & interconnection support

Frequency regulation, voltage support, and demand management that help secure grid connections or reduce costly upgrades.

▸

Hybrid enablement

Pairs with renewables (firming), behind-the-meter gas, fuel cells, or nuclear, and the grid — building a reliable "private grid" or co-located microgrid.

▸

Sustainability & revenue

Cuts diesel reliance — Microsoft has committed to phase out petroleum-diesel backup by 2030 — and can earn revenue in energy markets.

Where BESS makes a deal stronger

The "missing link" for power certainty in hybrid or constrained-grid sites — smoothing volatility and enabling reliable operation on partial grid or renewable input.

Improves bankability — stronger reliability metrics and lower outage risk, non-negotiable for mission-critical AI workloads.

Enables Bring-Your-Own-Power and microgrid strategies — faster deployment in power-constrained areas, and potential incentives.

Lowers system cost through arbitrage, grid-services revenue, and deferred grid upgrades.

Where weak BESS breaks or delays it

Insufficient capacity, duration, or response speed fails uptime or load-smoothing needs — forcing redesigns or walk-aways.

High upfront cost or technology risk (degradation, large lithium-ion safety) can kill economics if not offset by revenue or savings.

Poor integration with the generation mix leads to interconnection denials or delays.

In competitive site selection, no robust BESS/hybrid plan makes a location uncompetitive.

The ecosystem

Vendors & integrators: Fluence, Tesla (Megapack), Energy Vault & Peak Energy (sodium-ion), ZincFive, Schneider Electric, Eaton, EPC Power, Vertiv, and Wärtsilä (hybrid engine + storage).

Demand drivers: Google, Meta, and Microsoft via direct projects and PPAs, plus colocation providers like Equinix and Iron Mountain integrating on-site or co-located storage. U.S. storage deployments hit records in 2025, with data centers a key growth vector.

Research we track

The market literature, at a glance

MarketsandMarkets — U.S. Data Center BESS, 2025–2030~$274.5M (2025) → $527.8M (2030), a 14.0% CAGR; covers Li-ion vs. sodium-ion, hyperscale/colo/edge, and behind- vs. front-of-the-meter.

Meticulous Research — BESS for Data Centers, 2026–2036Global market ~$4.38B (2025) → ~$4.96B (2026); lithium-ion dominant, sodium-ion emerging, North America leading.

IDTechEx — Battery Storage for Data Centers & C&I, 2026–2036Forecasts ~5x growth in C&I BESS demand; benchmarks technologies, players, and suitability for data-center reliability and fast response.

Wood Mackenzie / ACP — U.S. Energy Storage Monitor (Q1 2026)Record U.S. deployments (~3.3 GW / 8.4 GWh, up ~54% over the prior Q1 record), with data centers a key C&I driver — and Google and Meta moving into long-duration storage for 24/7 carbon-free goals.

Fluence / DNV; Wärtsilä 2026 Energy Storage Outlook; EIAIndustry research on BESS for AI scalability, power quality, and replacing diesel backup (e.g., the Google St. Ghislain case study), with data centers cited as a major demand driver.

Company, project, and market-size references are illustrative of the sector, drawn from public reports (MarketsandMarkets, Meticulous Research, IDTechEx, Wood Mackenzie / ACP, Fluence, DNV, Wärtsilä, EIA, ESS News); not investment recommendations. Figures are as reported, reflect differing scopes (e.g., U.S. vs. global, hardware vs. full system), and are subject to change.

Capital and capability for a once-in-a-generation infrastructure cycle.

An independent advisor built for sustainable capital markets.

▸ Sector focus

▸ Aligned structures

▸ Institutional reach

Why the energy transition moved to the data center.

Supply-side → demand-side

Demand inflected for the first time in a generation

The constraint flipped from chips to power

The hyperscaler became the transition's banker

The bridge is the strategy, not a compromise

Five pillars. One partner.

Energy Transition

GP Advisory

Sustainable Infrastructure

Co-Invest

AI Power Infrastructure

Let's talk.

Co-Invest — direct exposure, pari passu with the GP.

Aligned with the GP

Right-sized allocations

Built for family offices

Efficient structure

The SPV placement process

Define objectives

Prepare materials

Go to market

Due diligence

Close

AI Power Infrastructure.

Investor

Financial Advisor

Development Capital

Strategic Solutions

A once-in-a-generation infrastructure super-cycle.

One of the largest infrastructure build-outs in history.

A decade of accelerating need — 2026 to 2036

Headline estimates compared — by firm

Where we think the AI power market is actually being decided.

The Power Constraint Thesis

Offtake as Infrastructure

Market Map: Who's Financing What

The Behind-the-Meter Shift

Infrastructure sectors powering the AI build-out.

Data Centers & AI Factories

Power Generation

Renewables & Storage

Grid & Transmission

Cooling & Critical Infra

Digital Infra Technology

Capital for the developers building AI campuses & edge.

Hyperscaler campuses

Edge data centers

Backing time-to-power

Solving the power, siting & efficiency constraints.

Decarbonization: from NIMBY blocker to deal enabler.

Environmental framing

Policy & community leverage

Burden → catalyst

Five opportunities in the AI power race.

Renewables PPAs & co-location

Efficiency as a multiplier

Hybrid firm power

Demand flexibility & grid services

Siting & design

A portfolio approach

Revenue, not just PR.

High-integrity offsets & removals

Removal built into infrastructure

Bundling opportunity

Government incentives & tax credits

Picks & shovels: the AI-adjacent buildout.

Cooling & Thermal Management

Electrical Equipment & Power Distribution

Construction & EPC

Networking & Interconnects

Real Estate · Colocation · REITs

Servers, Racks & Enclosures

Specialized Semiconductors

Fiber Optics & Connectivity