Every question we’re asked in every demo.

Cliffcenter is an AI-native operating system for data center predevelopment. It ships as five solutions that share one regulatory graph: Queue Intelligence (large-load interconnection), Incentive Optimization (state incentive filings), BTM Workflow (behind-the-meter deal structuring), Site Intelligence (BTM-viable site index), and Predevelopment Services (AI-augmented consulting delivery).

Each solution is defensible on its own and earns its first customers before the platform becomes obvious. You do not need to buy the full platform to buy any single wedge. Over time, the shared regulatory graph compounds the value of each additional solution, so an operator with three wedges gets the fourth for marginal cost. See /research/wedge-thesis for the full analysis.

No. Cliffcenter is the regulatory graph below the law firm and the engineering consultant. Our customers continue to work with counsel and engineering — but instead of paying them to track dockets, read orders, and move PDFs, they pay them for the judgment calls that actually require a human. We make their work faster, not their jobs redundant.

Request a demo at /contact?intent=demo and we'll walk you through a loaded workspace on either our sample data or yours. No NDA required for the sample; yours sits under a standard mutual NDA.

SOC 2 Type II audit is in progress with a Big 4 firm, target completion Q3 2026. In the meantime, we can provide a SOC 2 readiness statement, a summary of controls, and a signed security questionnaire to prospective enterprise customers under NDA. See /security for our public posture.

All customer data lives in AWS us-east-1 and us-west-2, with point-in-time recovery and cross-region backup. Per-tenant logical database isolation means cross-customer queries are impossible by construction. Enterprise customers can request VPC or on-prem deployment for sensitive transactions.

No. We never train public or shared models on customer workflow data. Per-tenant fine-tuning is only performed under signed agreement with data scoped to that tenant. The form-mapping engine is trained on public filings, synthetic examples, and explicitly-opted-in public contributions.

Customer workflow data is retained for the life of the subscription plus 90 days. You can export and delete your data at any time via the admin console or a written request to privacy@cliffcenter.com. Deletion is immediate and irreversible.

Yes. Our standard DPA is available on request and incorporates the usual GDPR and CCPA data handling commitments. Enterprise customers often add specific requirements; we accept most without friction.

We maintain a structured schema of 427 programs across 50 states, updated weekly by a combination of automated monitoring (every state comptroller page, every economic development agency press release, every legislative tracker) and a small analyst team that triages changes into the schema. Every program record cites its source statute or application URL and carries a last-verified timestamp. State briefs for Virginia, Texas, Georgia, Arizona, Ohio, Tennessee, and Iowa are published under /solutions/incentive-optimization.

Yes. Atlas is the entry point for a single-site operator — $2,200 per month, full database access, unlimited candidate projects, weekly refresh. When the first incentive application comes up, most customers graduate to Filings at $15,000 flat fee per filing.

Usually not. Most customers use us alongside an existing brokerage or consultant — we handle the paperwork and compliance, they handle the relationship with local agencies. We reduce the consultant's scope and cost meaningfully without replacing the relationship.

LandGate covers energy infrastructure broadly and has a solid BTM white paper from April 2026. Transect focuses on environmental due diligence. Athena is an internal Cushman & Wakefield platform not available to external buyers. Cliffcenter integrates plant retirement schedules (EIA), interconnection capacity (ISO/RTO), remediation status (EPA), zoning (county GIS), and water availability into a single schema scoped specifically to data center viability. Most customers use us alongside LandGate or Transect, not instead.

ISO queue data, EIA plant data, and EPA environmental data refresh nightly. County zoning and parcel data refresh weekly. NRC licensing and federal register notices refresh hourly. Every record carries a last-verified timestamp and a source URL. See /solutions/site-intelligence/coal, /nuclear, and /brownfield for curated public samples.

A weighted blend of POI voltage, queue headroom, watershed access, zoning classification, remediation status, local ordinance risk, and sponsor intent. The default weights reflect our analyst team's judgment; customers can tune the weights against their own project thesis via the Diligence tier.

You should not trust us as the only set of eyes on a $500M deal. You should trust us as the tool that surfaces risk faster than your in-house team or your outside counsel can maintain manually. The sign-off stays with the humans who have the license and the liability. We reduce the cost of keeping the humans informed.

Internal platform teams are great at serving a single company's pipeline. They are not positioned to maintain a regulatory graph that spans 50 states, 7 ISOs, and hundreds of utilities — and they certainly don't have the velocity to encode a new FERC order within 24 hours. Most of the hyperscalers we talk to run an internal platform and license Cliffcenter to cover the parts that don't belong inside a single company.

The opposite is more likely. The regulatory landscape has been rewriting faster every year, and the rate of change is accelerating. A dedicated software company tracking the graph has a structural velocity advantage over internal teams. If the landscape stabilizes on a 10-year horizon — which we don't believe it will — the product naturally evolves into a transactional platform for the stable workflow.

We sell alongside big firms, not against them. Our wedge is the work the big firms won't staff at your price point and the work your in-house team needs software for. Latham, Baker Botts, and Gibson Dunn all have customers who use Cliffcenter; we don't displace their billable surface, we compress the parts they don't want to bill for.

On Enterprise plans, yes. A handful of customers run Cliffcenter as a white-label layer inside their own platform. We ship an admin console, a tenancy model, and a support escalation path that supports the white-label configuration.

Every artifact the platform produces cites its source. Every filing links back to the underlying regulatory text and the inputs from the customer's workspace. Audit logs are append-only and tamper-evident. If our AI can't trace a claim to a source, it flags the claim and refuses to emit it. The 'fidelity floor' is 100%, not 95% — we would rather skip a field than guess it.

ChatGPT doesn't know the PUCT 58481 draft rule exists. It doesn't have the ERCOT base case, the FERC docket history, the 427-program incentive atlas, or the ISO tariff structures encoded as first-class objects. It can summarize a public document, but it cannot generate an interconnection study request package or model an upgrade cost envelope. The product is the regulatory graph, not the LLM.