How much does Light actually cost in 2026?

Licensing model: Quote-based SaaS subscription; no published price list, tiers, or pricing page exists as of mid-2026 (light.inc/pricing returns nothing), and no third-party pricing benchmarks, Vendr data, or practitioner-reported deal figures were found. Pricing structure (per entity, per user, per module) is itself undocumented publicly.

This is a genuine blank: no vendor list prices, no marketplace benchmarks, and no practitioner reports were found, so every dollar figure for Light is unconfirmed. The closest defensible anchors are indirect. First, the AI-native GL category lands roughly $20K-$40K/yr for typical multi-entity buyers (Rillet's Vendr median ~$28K/yr is the only hard benchmark in the cohort). Second, Light's customer base skews earlier-stage European scale-ups than Rillet's or Campfire's, and its founder messaging attacks the '$50,000 and five months' cost of legacy expansion, both of which suggest entry pricing at or below the category anchor. Treat any budget figure as low confidence until quoted, and note that the embedded expense management and virtual cards may partially offset a Ramp/Brex-class subscription elsewhere in the stack.

No public implementation fee data exists. The vendor markets go-lives in 'weeks, not months' with a deployment team it is actively expanding, which implies a vendor-led model with materially lighter scope than suite ERPs; whether it is billed separately, bundled, or waived is unknown. Budget internal time for entity setup, master data, bank connections, and tax-code configuration per jurisdiction, and get the fee, timeline, and named resources in writing.

Costs buyers commonly miss

• ▪The surrounding stack largely stays: billing/rev-rec tooling (Light shows no rev-rec engine), payroll, FP&A, and any operational systems remain separate subscriptions.
• ▪Multi-jurisdiction onboarding effort: per-country tax codes, e-invoicing registrations, and local bank connections take internal time even when the software supports them.
• ▪US-stack integration gaps may force manual imports or middleware until the US connector catalog matures.
• ▪Currency conversion and payment execution fees through the embedded payment rails; ask for the fee schedule.
• ▪Renewal repricing risk standard to venture-stage vendors, compounded here by the smallest war chest in the category.
• ▪Parallel-run and auditor-onboarding costs, likely higher in the US where no audit firm has public Light experience.

No renewal history exists; the company only began charging commercially around 2025. The structural risks are the same as the rest of the cohort (early-adopter discounts resetting, complexity-based repricing) plus one specific to Light: with $43M raised against rivals holding $100M+, a future down-round, acquisition, or pivot is a live scenario, and renewal terms could shift with it. Multi-year rate locks and contractual exit terms matter more here than anywhere else in this category.

• ▪First-US-logo status: reference, logo, and case-study participation in exchange for pricing and onboarding concessions.
• ▪Competitive quotes from Rillet, Campfire, and DualEntry; Light must beat better-funded rivals on price or terms.
• ▪Multi-year rate lock with a defined renewal uplift cap.
• ▪Contractual data-export formats, exit assistance, and source-data escrow, given the vendor's stage.
• ▪Committed implementation timeline and named deployment resources in the order form.
• ▪Fee schedule transparency on embedded payments and FX before signing.

Negotiation note: Leverage is high in the US: Light is buying its first American logos, which usually means aggressive pricing, white-glove onboarding, and flexible terms for referenceable customers. The counterparty risk is equally real, so trade reference participation for contractual protections: rate locks, data-export terms, and committed support SLAs.

Confidence here is low, and we would rather say so than fake precision: pricing is quote-based and public data is thin. Treat every number on this page as a rough anchor to open a negotiation, not a benchmark to hold a vendor to. Every figure traces to the sources below, last reviewed 2026-07-12.