Underwriting and dispatch software for second-life battery systems that let GPU data centers add power before grid upgrades land.

1. Factory-scale production reduces the biggest historical blocker: reused EV batteries no longer have to be sourced and engineered one project at a time.
2. AI-driven load growth is creating customers who value speed-to-power more than perfect greenfield economics, making second-life systems newly attractive.
3. Large new financing means the ecosystem can support a dedicated software and warranty layer instead of relying on one-off engineering services.
4. The market narrative has shifted from recycling to independent energy, which creates budget owners in infrastructure teams rather than sustainability teams.

Catalyst. Moment's new funding and factory buildout, combined with explicit framing around AI data-center power bottlenecks, suggest reused EV batteries are becoming available at the exact moment data centers need non-grid capacity fast.

Build a software layer that turns heterogeneous used EV battery inventory into a financeable, operable power system for constrained data-center sites. The product ingests battery health and provenance data from second-life factories, scores each cohort for specific duty cycles, and produces permitting, insurer, and lender documentation that today must be assembled manually. After deployment, it runs dispatch for peak shaving, ride-through, and limited backup while continuously tracking degradation against the original underwriting assumptions. Over time, the company builds the best dataset on how reused battery cohorts perform under real AI-load patterns, enabling warranties and financing products competitors cannot price confidently.

What's different. Most energy-management software assumes new, uniform battery packs; most second-life players sell hardware or project integration. This company owns the hard middle: cohort-level underwriting, safety documentation, and runtime optimization purpose-built for reused batteries in high-value power-constrained environments. Its moat is a growing performance dataset linking battery provenance, duty cycle, degradation, and project outcomes, which becomes the basis for warranties and financing approvals.

Jobs to be done

flowchart LR
  Buyer[GPU colo operator] --> Pain[Utility upgrade delayed; revenue at risk]
  Pain --> Product[Second-life battery underwriting plus dispatch]
  Product --> Outcome[Faster site energization with bankable storage]

Executive takeaways

• Supply-side proof is now real: Moment says its Series B took total funding above $100M and its DOE-backed U.S. factory plan targets 1 GWh of repurposed storage, so second-life batteries are moving from pilot rhetoric toward industrial supply [2][4].
• The bottleneck is bankability, not chemistry alone: certification, safety review, and battery-valuation data keep showing up as gating issues before reused packs can be financed or permitted at critical-load sites [6][8][9][24][31].
• AI and data-center power scarcity make the beachhead unusually urgent; primary and independent sources now frame batteries as part of the answer to constrained data-center growth [1][3][29][30][31][32][35].
• Competition is real but fragmented between hardware-led second-life vendors and easier-to-insure first-life or diesel substitutes, leaving room for a neutral underwriting and evidence layer [2][8][19][21][27].
• Willingness to pay should come from avoided delay, diesel substitution, and demand-charge relief rather than from a generic sustainability budget [7][12][14].
• The beachhead SAM is likely in the tens of millions rather than billions, so venture-scale upside depends on expanding the same control plane into fleets, industrial campuses, and financing/warranty partners [9][10][11][14][35].

Market definition

North American software control layer for second-life battery energy storage deployed behind the meter at power-constrained sites, with a beachhead in 1-10 MW AI/GPU colocation and adjacent microgrid projects. The category includes battery cohort scoring, safety and insurer/AHJ evidence packs, and dispatch/monitoring software for reused EV packs; it excludes battery manufacturing, recycling-only services, generic EMS for homogeneous new batteries, and utility-scale merchant storage [2][3][13][17][18][35].

Customer and buyer

The ICP is a North American data-center operator or microgrid EPC facing a signed compute customer and a delayed utility upgrade. The daily user is the infrastructure/energy team and project EPC; the economic buyer is likely the VP Infrastructure, COO, or development head because the spend sits inside capacity delivery and resilience rather than a sustainability budget. Procurement friction comes from insurer, fire-code, and engineering review as much as from software security [1][2][14][31][32][35][36].

Buying triggers

• A signed AI or colocation contract lands before utility capacity is available, putting revenue timing at risk. [1][2][35]
• The buyer needs a faster alternative to diesel or wants to shave expensive peaks while waiting for grid reinforcement. [7][12][14]
• A project team wants to use reused batteries but gets blocked by insurer, AHJ, or internal safety review requests. [6][8][31]

Willingness to pay

Willingness to pay is strongest when the product attaches to an already-justified infrastructure spend: peak shaving, diesel replacement, or financed battery projects. Fetched evidence shows buyers already accept leasing and Battery-as-a-Service structures and value avoided demand charges and resilience benefits, which supports a workflow-plus-runtime software layer sold into the same project budget [7][12][13][14]. [7][12][13][14]

Category dynamics

Growth signal Trade research cited by Energy-Storage.News values the recycling plus second-life battery market at $45B by 2030, while data-center battery deployments are also scaling into larger projects.

Validation signals

• Moment’s 2026 Series B brought total capital above $100M, a strong signal that investors see second-life batteries as scalable infrastructure rather than bespoke pilots.
• Moment’s DOE-backed factory plan and UL 1974 certification milestone show supply-chain and certification infrastructure are forming in North America.
• Moment’s SFU collaboration explicitly ties second-life batteries to growing data-center power needs.
• Connected Energy has investor backing plus OEM relationships with Volvo and other partners, reinforcing that second-life supply and deployment ecosystems are maturing.
• B2U has put multiple second-life systems into operation, including a Honda-battery project, indicating non-trivial commercial deployment is already happening.
• Google, Amazon-adjacent utility procurement, and older Nissan/Eaton deployments all show batteries are moving closer to core data-center power architecture.

Regulatory & technical constraints

• Lithium-ion batteries inside AI data centers create fire-risk and code-review complexity that can delay approval even when the technical case is sound.
• Repurposing-process credibility matters; UL 1974-style evidence is becoming a meaningful bankability signal for reused batteries.
• EPA guidance makes clear that lithium-ion batteries require careful end-of-life handling because improper management can harm people and the environment.
• Reliable second-life underwriting depends on battery telemetry and state-of-health data, not just nominal nameplate specs.
• Dispatch software must integrate with site loads, tariffs, grid services, and other energy assets, which raises implementation complexity.

Competition splits into hardware-led second-life vendors (Moment, Connected Energy, B2U, Smartville, RePurpose), first-life BESS substitutes, and the default of waiting for grid upgrades or using diesel. Most current players make money by supplying systems or battery inventory; fewer are neutral across suppliers, and fewer still specialize in insurer/AHJ-ready underwriting for critical-load sites [2][8][17][18][19][21][27][29].

Why incumbents do not win by default

• New-battery BESS vendors. First-life BESS vendors are easier to insure because their packs are homogeneous, but their default motion is to sell hardware capacity, not to build a neutral underwriting layer that makes heterogeneous reused packs bankable across multiple suppliers.
• Second-life hardware vendors. Moment, Connected Energy, B2U, and Smartville prove the category exists, but they are economically pushed toward selling their own systems or inventory; that leaves room for software that is supplier-neutral and optimized for documentation, insurance, and runtime evidence rather than metal.
• Generic energy software and in-house builds. Horizontal EMS or internal tooling can dispatch batteries, but they do not come with repurposing-process credibility, battery valuation logic, or ready-made safety packages for reused EV packs in high-scrutiny facilities.
• Wait-for-grid and diesel substitutes. The status quo often wins because it is familiar, not because it is superior. Battery systems that bridge grid delays or replace diesel become attractive when lost revenue, emissions, and operating complexity are explicit enough to justify a faster alternative.

Second-life battery supply is becoming industrialized just as North American GPU colocation operators face delayed utility upgrades and acute pressure to energize revenue-generating AI capacity faster. The proposed company should enter with a narrow software wedge: cohort-level battery underwriting, insurer and AHJ evidence packs, and dispatch controls for 1-10 MW behind-the-meter projects that use reused EV batteries for peak shaving, ride-through, and limited backup. The first customer is a colo operator or microgrid EPC working a live project where a signed tenant or utility delay makes months of lost capacity expensive enough to justify a paid pilot. Research supports the urgency and channel logic, but not full market certainty: the modeled SAM is about $37.5M and year-3 SOM about $3.0M, so the investment case depends on expanding the same control plane into fleet depots, industrial campuses, and financing or warranty partners. The company should therefore avoid broad energy-management positioning and avoid competing on hardware, full EPC scope, or generic sustainability analytics. The near-term proof point is not deployed megawatt-hours alone; it is whether standardized evidence and runtime data can convert second-life batteries from bespoke engineering risk into a repeatable project approval path. The biggest disconfirming risk is that critical-load buyers still default to first-life batteries or diesel because insurer, AHJ, and owner-engineer scrutiny remains too high for reused packs. Key gaps from the inputs are the exact documentation set required for approval, the commercial terms for multi-supplier telemetry access, and whether software can be sold as a distinct line item rather than only bundled into project economics.

Problem

• GPU colocation operators and microgrid EPCs lose months of revenue or project margin when utility upgrades lag but second-life battery projects still require bespoke safety, performance, and permitting work.
• Existing alternatives such as first-life BESS, diesel, or internal spreadsheet-based studies win by default because reused EV packs are harder to underwrite, insure, and operate consistently across heterogeneous cohorts.

Solution

• Build a supplier-neutral underwriting layer that ingests battery provenance and state-of-health data, scores cohorts against specific duty cycles, and produces standardized insurer, AHJ, and owner-engineer evidence packs.
• Add shadow-mode then live dispatch software for peak shaving, ride-through, and limited backup so realized degradation and runtime performance feed back into future underwriting and financing decisions.

Why we win

• The company sits in the high-friction middle that hardware vendors and generic EMS providers under-serve: making heterogeneous reused batteries bankable for critical but time-sensitive projects.
• A growing dataset linking battery provenance, duty cycle, approval outcomes, and realized degradation can become a differentiated underwriting and warranty asset across suppliers and sites.
• The first buyer has a concrete budget trigger because lost AI capacity, diesel exposure, and delayed tenant revenue are easier to justify than a generic sustainability software purchase.

Milestones

flowchart LR
  Wedge[Delayed GPU colo power-gap project] --> MVP[Underwriting workspace plus evidence pack]
  MVP --> Proof[Approved pilot and shadow-dispatch data]
  Proof --> Expansion[Live dispatch then fleets campuses finance partners]

Founding team

Experiment roadmap

Risk assessment

What must be true

• At least 30% of target beachhead accounts have a live or near-term project where utility delay makes temporary or behind-the-meter storage economically urgent.
• Insurer, AHJ, and owner-engineer reviews can be standardized enough that the first evidence-pack template materially reduces bespoke project work.
• At least two second-life suppliers will provide provenance and telemetry data on terms that preserve supplier neutrality.
• Buyers will fund the product from capacity-delivery or resilience budgets at roughly the researched $250k blended year-1 value, not treat it as optional software.
• The underwriting and runtime layer will expand into larger adjacent constrained-site segments quickly enough to matter before the beachhead saturates.

Open diligence questions

• What exact document set and third-party evidence are required today for insurer and AHJ approval on a reused-battery data-center project?
• How often do delayed 1-10 MW colo projects choose diesel or first-life batteries instead of second-life storage, and why?
• Can the startup get multi-supplier battery data without exclusive hardware relationships or margin-sharing that destroys neutrality?
• Is the real buyer the colo operator, the EPC, or a financing partner once a live project reaches procurement?
• Which early use case clears fastest in practice: peak shaving, ride-through, or limited backup?

Model sanity

• Revenue engine. Base case revenue is driven by reaching about 11 active projects by Y3 exit at a $250K blended annual ARPU, which implies roughly $2.8M exit ARR and $2.28M recognized Y3 revenue.
• Must go right. The company must convert early paid pilots into repeatable production deployments fast enough to stay on the 3 / 4 / 6 customer-add path while keeping supplier integrations reusable.
• Model breaks if. If sales cycles slip and pricing falls toward the downside case, Y3 EBITDA worsens to about -$521K and the cash low point compresses to roughly $549K.
• Next-round proof. The next financing is justified once the company shows 4 to 6 production deployments, two supplier integrations, one referenceable data-center case, and adjacent-segment pull by month 24.

Scenarios

Sensitivity

flowchart LR
  Leads --> PaidPilots
  PaidPilots --> ProductionProjects
  ProductionProjects --> Revenue
  Revenue --> GrossProfit
  GrossProfit --> Cash

Flags: Revenue remains concentrated in a small number of deployments, so one delayed or lost project would move Y3 results materially. · The beachhead SAM is only about $37.5M, so venture-scale upside still depends on expansion into fleet, industrial-campus, and financing or warranty workflows. · The model assumes software can hold a standalone or clearly attributable $250K blended annual project value even though research notes buyers may prefer bundling into EPC or financing economics. · Holding 70% gross margin requires resisting bespoke field-service and certification work from leaking into COGS as projects scale.

• Safety and permitting drag. Authorities, insurers, or customers may treat second-life batteries as too risky for critical infrastructure sites. Mitigation: Start with peak-shaving and ride-through use cases rather than full backup, and ship standardized AHJ and insurer documentation backed by third-party testing partners.
• Inconsistent battery supply. Used EV battery cohorts may vary too much in chemistry, health, and provenance for repeatable deployments. Mitigation: Limit supported cohorts early, integrate directly with factory-grade testing data, and only underwrite projects that fit validated duty cycles.
• Hardware vendors close the gap. New-battery BESS vendors or second-life manufacturers may build basic software and squeeze the company out of projects. Mitigation: Own the cross-supplier underwriting dataset and financing-grade analytics layer, and partner with hardware providers instead of competing on metal.