Research in progress · BESS & data centers · 29 September 2026

Can compute make a BESS investment stronger?

A research question about customers, capital and contracts - with an important distinction between higher revenue and a better investment.

The working hypothesis

A battery site may offer useful land, power infrastructure and operating expertise for modular data-center capacity. A credible customer and earlier paid operation could add value. Our research asks whether that value survives the extra investment, operating constraints and customer risk.

Status: research draft dated 29 September 2026. The examples are assumption-led, not results from an operating project or evidence of lender approval.

There is a tangible market to investigate.

Modular equipment exists.

Vertiv offers prefabricated infrastructure in 0.5 MW and 1 MW IT building blocks, including power and cooling. This establishes a product class, not a turnkey price or guaranteed project schedule. [1]

Infrastructure leasing is emerging.

Energy Vault announced construction of an initial 8 MW powered-shell campus in Texas, combining BESS and modular infrastructure leased to Crusoe. Operation is targeted for Q1 2027; realised returns are not disclosed. [2]

Flexibility can be contracted.

Google reports 1 GW of demand response in long-term US utility contracts. That supports contractual feasibility at scale, while leaving European site economics and individual customer flexibility to be assessed. [3]

Choose the business before modelling the return.

An investor can own the building, power and cooling while a tenant owns the servers. Alternatively, the investor can also buy computing hardware and sell compute services. These businesses have different capital needs, refresh cycles and customer risks. An infrastructure-only budget cannot support a forecast of GPU-service revenue.

Our worked draft examines 5 MW / 20 MWh BESS plus 1 MW of IT capacity, with tenant-owned servers. This research case is separate from the 9.9 MW BESS / 2 MW data-center Kinetic Shield reference design.

Compare the addition and the location separately.

First compare the combined project with BESS alone: does the extra business earn its extra cost? Then compare co-location with the same businesses at separate sites: does sharing the location actually help? A successful hosting business can raise total returns even when co-location adds very little.

A higher IRR can still require much more equity.

In the draft's fifteen-year illustrative case, adding flexible compute raises modelled after-tax project IRR from 6.1% to 9.6%, while required equity increases from €3.15 million to €9.31 million. These are model outputs under disclosed assumptions, not return targets.

More revealingly, the same-start flexible co-location case has approximately €48,000 less net present value than matched separate assets at the assumed 8% discount rate. The exercise therefore does not yet demonstrate a co-location premium. A verified timing advantage, lower delivered cost or better service contract would need to change that result.

An assumed minimum capacity-payment contract improves modelled borrowing capacity, but the customer's ability to pay remains central. PwC/FfE's BESS financing work similarly highlights the interaction between commercial structure, counterparty credit and debt. [4]

What must be proved before committing capital?

A route to paid operation. Equipment delivery, authorised grid access, permits, cooling, connectivity and customer acceptance must align. Existing BESS connection rights do not automatically authorise an added data-center load.

A benefit that survives real operations. Battery charging and compute demand share an electrical limit. Reserved backup energy reduces trading freedom. Load shifting may create a later rebound. The study must measure service quality, battery wear and retained battery income against BESS-only and inflexible-load alternatives.

Cash the investor can retain. Minimum payments, energy-cost allocation, service penalties, customer failure and lender rights determine how much of the expected income supports financing. Physical proximity does not transfer another company's cash flow to the battery owner.

For investors, the useful output is a decision threshold: the minimum customer commitment, maximum installed cost and connection timing at which proceeding becomes worthwhile. Rejecting an unattractive combination is as valuable as identifying a viable one.

Investor takeaways

Separate the hosting business from the benefit of sharing a site.
Check the additional equity requirement alongside the return.
Contract flexibility and customer payments before counting financing benefits.

Research basis: Making BESS and compute work together, research draft dated 29 September 2026. Research is ongoing.

Research partnership

Help test the investment case.

We are looking for research partners, BESS and data-center operators, infrastructure suppliers and finance specialists. Priorities are matched site cases, supplier quotations, measured workload flexibility and independent review of the contracts and financing assumptions.

The next stage is to replace illustrative inputs with evidence and identify where the combination genuinely works. Research is ongoing; no universal return or financing uplift is claimed.

Request research partnership Read the wider market context