Lender-side underwriting OS for powered-land and data-center projects that turns site diligence into IC-ready credit memos in days.

1. A 95% reduction in diligence time makes speed a competitive requirement for anyone trying to finance scarce powered-land opportunities.
2. More than 100 projects across 15 countries and an initial hyperscaler customer imply the volume and urgency of critical-infrastructure siting are outgrowing bespoke lender review processes.
3. Once site sourcing, power assessment, and early design are becoming structured machine work, lenders can buy a purpose-built underwriting layer instead of relying only on consultant packets.
4. Category investors are signaling that AI which produces real infrastructure work product is fundable and credible, which lowers buyer resistance to an output-oriented underwriting tool.

Catalyst. Build's 95% diligence-time reduction, first hyperscaler customer, and 100-project cross-border footprint show project velocity is accelerating faster than lender underwriting workflows can absorb.

The product is a lender-side control plane for physical-infrastructure underwriting. It pulls sponsor materials, utility and parcel data, environmental and zoning records, transmission and water constraints, and early design assumptions into a single deal graph, then generates a red-flag view for credit and investment committees. Instead of waiting weeks for a stitched-together consultant packet, associates get an IC memo draft, missing-data checklist, and scenario analysis showing which assumptions break the loan case. After a term sheet is issued, the system keeps monitoring permitting, power, and design drift so lenders can gate drawdowns and covenant updates before problems compound. Over time, the platform learns which early diligence signals predict delay, cost overrun, or deal abandonment across financed sites, creating a proprietary infrastructure risk benchmark.

What's different. Developer tools optimize sponsor velocity, consultant firms sell bespoke studies, and generic AI document tools summarize data rooms. This company is built for the capital side of the table: it turns site facts into underwriting opinions, evidence trails, and ongoing covenant monitoring. Its moat comes from linking early diligence signals to actual financing outcomes across many infrastructure deals, something sponsor-side software and point consultants do not naturally capture.

Jobs to be done

flowchart LR
  Buyer[Infra credit lead] --> Pain[Slow consultant-heavy site underwriting]
  Pain --> Product[Powered-land underwriting engine]
  Product --> Outcome[Faster approvals and fewer hidden site risks]

Executive takeaways

• This is not generic proptech: data-center development is now a power-, policy-, and capital-constrained underwriting problem, with power availability still the top industry challenge and primary-market vacancy at 1.4% in H2 2025 [14][15][16].
• White space exists on the capital side: Build and Paces already compress sponsor diligence, but their products are optimized for finding and advancing sites rather than giving lenders an independent credit memo and drawdown-control layer [1][3][17][18][19][21].
• Willingness to pay is credible because buyers already purchase this work as a material line item; Index describes a $25,000 data-center desktop diligence PDF that took six weeks, while Build and Paces market 10x speed or 5-day reports [1][3][8][18].
• Private credit is turning into a core enabler of the buildout, which raises the value of faster, auditable underwriting rather than lowering it [10][11][12][13].
• The hard part is trust and fragmented source systems, not model availability; utility processes, county zoning changes, and environmental screening still need to be stitched into a defensible decision trail [7][9][28][29][30][35].

Market definition

Defined market: lender-side workflow software and outcome services that convert sponsor materials, parcel/power/fiber/water datasets, and regulatory records into IC-ready underwriting evidence for data-center and powered-land financings; it sits between raw site-intelligence tools and full project-finance execution [1][4][5][9][10][21][23][24].

Customer and buyer

Day-to-day users are heads of underwriting, investment directors, and credit associates at infrastructure private-credit, project-finance, or digital-infrastructure debt teams. The economic buyer is typically the partner, CIO, or head of credit because sponsor support, draw mechanics, utility risk, and entitlement assumptions materially affect downside protection and committee sign-off [10][11][12][31].

Buying triggers

• A sponsor needs a term sheet before power-delivery, zoning, or fatal-flaw diligence can be independently validated inside a short exclusivity window. [15][16][17][21]
• A construction, bridge, or project-finance process requires lenders to formalize sponsor support, draw conditions, and utility/permitting risk before closing. [10][11][31]
• Expansion into a new geography or grid region makes reusable utility, zoning, and environmental evidence packs suddenly more valuable than ad hoc consultant coordination. [22][28][29][30]

Willingness to pay

Budget already exists in consultant-led diligence, sponsor packaging, and project-finance workstreams. Index describes a single $25,000 data-center desktop diligence PDF that took six weeks, Build sells outcome-based deliverables at fixed upfront pricing, and Paces markets expert-validated reports and power-flow studies delivered in five days. The hurdle is proving independent underwriting value, not inventing a new budget line. [1][8][18][21][25]

Category dynamics

Growth signal 36% YoY primary-market supply growth in 2025, with 1.4% vacancy at year-end and 74.3% of H1 2025 capacity already preleased.

Validation signals

• Build claims 150+ delivered projects across 18 countries and $2T+ collective AUM of firms served, indicating serious institutional appetite for faster infrastructure diligence.
• A tier-1 global data-center developer says Build enabled earlier fatal-flaw detection and materially faster site-selection throughput in EMEA.
• CBRE reports investors still plan to increase allocations and see power availability as the top challenge, reinforcing both demand and pain intensity.
• Private credit and structured financing activity is scaling into data centers quickly, increasing the value of underwriting systems that can keep up.

Regulatory & technical constraints

• Large-load request and interconnection processes are formalizing in core power markets, making utility process compliance part of underwriting rather than an afterthought.
• County-level zoning changes and grandfathering rules can change whether a site can still proceed without a fresh discretionary approval path.
• Environmental screening for flood, wetlands, contamination, and water use is now a core early-stage diligence stream for financeable data-center sites.
• Project-finance structures still require sponsor support, covenant packages, milestone testing, and draw controls that depend on high-confidence diligence inputs.

Competition is strongest in adjacent workflows, not in lender-native powered-land underwriting. Build and Paces optimize sponsor-side speed and site advancement, LandGate and Acres sell data/reporting layers, and Colliers-like advisors monetize services-heavy diligence. The gap remains the neutral, credit-memo-first system that a lender can trust without inheriting the sponsor's bias or a consultant's timeline [1][17][20][21][24][26].

Why incumbents do not win by default

• Sponsor-side development OS vendors. Build and Paces are closest to the workflow, but they are optimized for helping developers find, validate, and advance sites, not for giving lenders an independent risk opinion and post-close monitoring discipline.
• Power and parcel intelligence platforms. LandGate and Acres surface critical parcel, power, fiber, and lending data, but they stop short of synthesizing that data into a lender-grade underwriting narrative with committee-ready mitigants.
• Advisory, legal, and engineering firms. Colliers, BDO, and specialist counsel already win trusted mandates, but the work remains project-by-project, people-heavy, and slower to compound into a reusable risk graph.
• In-house analyst plus specialist consultants. The default substitute is still a sponsor data room, spreadsheets, and sequential specialist reports, which preserves human judgment but is slow, expensive, and difficult to audit consistently across deals.

Data-center financing is becoming a power- and permitting-risk underwriting problem, and private-credit teams are now the bottleneck when sponsors need term sheets inside short exclusivity windows. The proposed company sells a lender-side underwriting OS that ingests sponsor data rooms plus power, zoning, water, environmental, and parcel data to produce IC-ready memos, fatal-flaw flags, and post-close monitoring triggers. The initial beachhead is intentionally narrow: North American infrastructure private-credit teams underwriting sub-$100M powered-land and preconstruction loans for secondary-market developers, starting with Northern Virginia deals where investor activity is dense and utility and zoning processes are explicit. This wedge is attractive because budgets already exist in consultant-led diligence, but the current workflow is fragmented across sponsors, lawyers, engineers, utility calls, and spreadsheet memos. Build, Paces, and LandGate validate that machine-assisted site diligence is real, yet none is positioned as the neutral lender approval layer with committee-grade evidence trails and drawdown controls. The first sale should be an assist-only pilot tied to one active loan pipeline, with pricing that mixes a team license and deal-tied fees so the buyer can reallocate existing diligence spend rather than approve a new software budget. The central risk is not model capability but trust and data coverage: if IC chairs will not sign off on AI-assisted memos or if Northern Virginia and Texas data normalization stays too bespoke, the business becomes services-heavy. The research also leaves two material gaps unresolved—the annual deal volume per beachhead team and the citation plus human-review threshold required for committee acceptance—so the first 90 days must validate both before the company scales hiring or geography.

Problem

• Powered-land and preconstruction loans require independent validation of power, water, zoning, environmental, and permitting assumptions, but lenders still assemble that judgment from sponsor materials, consultant PDFs, and manual utility or legal outreach.
• Short exclusivity windows force a bad tradeoff between speed and rigor: teams either lose the site by moving too slowly or approve a loan with hidden fatal flaws that surface after closing.
• Sponsor-side diligence tools, raw land-intelligence platforms, and advisory firms each solve part of the workflow, but none gives the lender a reusable, neutral, committee-ready opinion plus post-close drift monitoring.

Solution

• An assist-only underwriting engine ingests sponsor data rooms and structured infrastructure data to generate IC-ready memo drafts, red-flag lists, and missing-evidence checklists with direct source citations.
• Market-specific diligence modules cover utility and interconnection, zoning and grandfathering, water, flood and wetlands, and sponsor-support assumptions, starting with Northern Virginia and then Texas.
• Post-term-sheet monitoring tracks permitting, power, and design drift so lenders can gate drawdowns, waivers, and covenant changes before losses compound.

Why we win

• The product is built for lender trust rather than sponsor velocity, so its core output is a reviewable credit opinion with evidence trails, not just a faster site-search workflow.
• A Northern Virginia-first wedge creates faster proof than a broad infrastructure platform because deal density is high, power and zoning are acute, and buyers already feel the cost of waiting on fragmented diligence.
• Hybrid pricing fits how funds already buy work—consultants, counsel, and project-finance diligence—so the company can land inside an existing budget before asking for full platform standardization.
• Repeated use creates a proprietary dataset linking early diligence signals to committee decisions, waivers, delays, and post-close surprises that sponsor-side tools and services firms do not naturally capture.

Milestones

flowchart LR
  Wedge[NoVA lender underwriting wedge] --> MVP[Cited IC memo and red-flag MVP]
  MVP --> Proof[Faster committee-ready decisions and fewer missed fatal flaws]
  Proof --> Expansion[Monitoring plus second-region and adjacent-infrastructure expansion]

Founding team

Experiment roadmap

Risk assessment

What must be true

• At least 10 target funds evaluate 5 or more relevant powered-land or preconstruction financings per year and already buy third-party diligence.
• At least 60% of pilot memos can reach committee circulation without a full consultant rebuild.
• Northern Virginia and Texas data coverage can populate at least 80% of required diligence fields from repeatable sources.
• A hybrid contract converts more than half of successful pilots into annual team agreements funded from existing diligence budgets.
• Build, Paces, LandGate, or major advisory firms do not close the lender-native approval gap before the startup builds an outcome and monitoring moat.

Open diligence questions

• Which named funds in North America currently underwrite 5 or more relevant sub-$100M data-center financings per year?
• What citation depth, review workflow, and liability boundary does an IC chair require before using an AI-assisted memo?
• On 20 recent Northern Virginia and Texas deals, what percentage of power, zoning, environmental, and permitting evidence can be sourced repeatably without bespoke analyst work?
• Who owns the budget for monitoring and drawdown-control workflows after closing: underwriting, asset management, or outside advisers?
• How quickly could Build, Paces, LandGate, or Colliers move from adjacent workflows into lender-facing credit memos and surveillance?

Model sanity

• Revenue engine. Base revenue comes from growing paying credit teams from 4 at Y1 exit to 15 by Q4Y3 while realized annual revenue per team climbs toward roughly $459K as site and monitoring fees attach.
• Must go right. Northern Virginia and Texas coverage plus committee trust must become repeatable enough that 15 FTE can support 15 teams and about 50 sites without dragging gross margin back below the high 60s.
• Model breaks if. If pilot conversion slips and exit blended revenue per team stays closer to $38K per month, the downside case goes cash-negative before H2Y3 and needs a bridge round.
• Next-round proof. The seed-ready proof point is 9 paying teams by Month 24, 10 monitored sites, and a reusable committee-accepted memo workflow that carries the company to near-breakeven in H1Y3.

Scenarios

Sensitivity

flowchart LR
  TargetFunds[Target funds] --> PaidTeams[Paying credit teams]
  PaidTeams --> SiteAttach[Underwritten and monitored sites per team]
  SiteAttach --> Revenue[Hybrid subscription and per-site revenue]
  Revenue --> GrossProfit[Gross profit]
  GrossProfit --> Cash[Cash and runway]

Flags: customersEop counts paying lender teams, so a meaningful share of Y3 revenue still depends on per-site fees and monitored-site volume per team. · The Y2 hiring step-up is deliberate and assumes Texas coverage plus monitoring are built before broad sales scale; if custom analyst work stays above the BP target, even this plan will look too lean. · The downside case requires additional capital before H2Y3 if committee acceptance or hybrid pilot-to-annual conversion slips by roughly one to two quarters. · Cash is modeled from EBITDA and ignores collection timing, prepaid data licenses, and any capitalized software spend.

• Trust barrier. Credit committees may resist relying on AI-generated conclusions for multimillion-dollar infrastructure loans. Mitigation: Start as an analyst co-pilot with every conclusion linked to source evidence and optional human review, then expand autonomy only after proving faster approvals without surprise misses.
• Data coverage holes. Utility, municipal, and environmental data can be fragmented in secondary markets, leaving blind spots in the underwriting graph. Mitigation: Launch in a narrow geography and asset profile with repeatable data coverage, and explicitly flag unsupported assumptions instead of pretending to have full certainty.
• Incumbent bundling. Engineering consultants, appraisal firms, or sponsor-side diligence tools could add basic memo automation once the wedge becomes visible. Mitigation: Own the lender-side benchmark dataset and continuous post-approval monitoring workflow that incumbents do not capture in one auditable system.