# Edgify Financial Model

Edge AI computer vision software that identifies non-barcoded produce at retail point-of-sale, trained locally on the device without cloud dependency.

- Canonical: https://finamodel.com/startups/edgify
- Excel download: https://finamodel.com/startup-models/edgify.xlsx
- Category: AI/ML
- Model type: SaaS ARR / Valuation
- Funding round: Seed
- Funding: $6.5M
- Founded: 2020
- Geography: Not in deck (deck language is English; retail context appears global).
- Customer: B2B

## About the company

Edgify delivers edge AI software that identifies non-barcoded produce on existing retail point-of-sale hardware. The model trains locally from customer transactions, then shares weights across a store's devices without uploading store data to the cloud.

It supports cashier lanes, self-checkout, and PC scales, with a one-day-to-two-week integration followed by a four-to-six-week training phase. Edgify claims 99.98% recognition accuracy, 0.006-second recognition speed, and checkout-time reductions of up to 75%.

The ARR model is built around chain contracts, stores, and PoS devices, with a revenue ramp while each site learns. It tests sales-cycle length, devices per store, implementation capacity, device licensing, churn, expansion, software margins, engineering and sales staffing, cash burn, and runway.

## What's included

- 5-year monthly revenue build with stage-appropriate growth assumptions
- Full P&L, headcount plan, and operating-expense schedule
- Cash-flow statement, runway, and burn-rate tracking
- Valuation via exit multiple with a DCF cross-check
- Returns analysis with MOIC and IRR
- Unit economics including CAC, LTV, payback, and cohort retention

## Product & value proposition

Edge AI software that runs on existing retail PoS hardware (lane cashier, self-checkout, PC scale) and uses computer vision to identify fresh produce (fruit & veg, bakery, fresh produce) without barcodes. Key differentiator: the model trains itself locally on each device using real customer transactions, then shares model weights (not data) across all PoS units in a store and across stores - no cloud upload required. Accuracy claimed: 99.98% / "100% over time".

Three deployment use cases:
1. Lane PoS (cashier)
2. Self-Checkout (SCO)
3. PC Scale (F&V / Dairy)

Implementation timeline:
- Software integration: 1 day – 2 weeks (remote, no new infrastructure)
- Dormant/training phase: 4–6 weeks
- Active phase launch: 1 week
- Results review: 1 hour

Two proprietary IP components:
- Edgify Edge Training Loop - software loop that trains on any edge device with minimal processing overhead
- Edgify Collaborative Controller - aggregates models from all edge devices, combines into optimised model, redistributes

## Traction & metrics

Quantitative performance claims (product, not financial):
- Accuracy (market standard baseline): 65%
- Accuracy (Edgify): 99.98%
- Checkout time reduction: up to 75%
- Recognition speed: 0.006 seconds per item
- Time to full produce recognition in store: ~30 days of shopping

## Competition / moat

Competitive framing vs. cloud-based computer vision:
- Incumbents require millions of images sent to cloud for training; costly AI infrastructure and ongoing retraining.
- Edgify: trains on-device, no data leaves the store, continuous retraining on actual store conditions (lighting, angles, produce freshness stages), no cloud costs.
- Moat claims: on-device federated/distributed learning eliminates cloud dependency; proprietary collaborative controller IP; privacy-by-design (no data transfer); model improves automatically per store conditions.

No named competitors called out.

## Recommended financial model

- **Archetype + why:** B2B SaaS ARR model with a per-device / per-store subscription unit. The product is software deployed on existing hardware - recurring licensing is the natural monetisation. The key value metric is number of PoS devices (or stores) under contract, making a seat/device-based SaaS ARR model the right archetype.

- **Forecast horizon & granularity:** 3 years monthly (Year 1–2) then annual (Year 3); monthly granularity needed to model ramp from dormant → active phase (~6–8 weeks per deployment) and cohort-based ARR recognition.

- **Key drivers & assumptions:**

| Driver | Value |
| -- | -- |
| Average PoS devices per store | 5–10 |
| Annual license fee per device | $X (unknown) |
| Average store count at contract signing (chain deal) | 10–50 stores |
| Sales cycle length | 3–6 months |
| Dormant + active phase ramp (weeks to revenue) | 6–9 weeks per site |
| Gross margin (software only, no hardware COGS) | 70–80% |
| Logo churn (annual) | 5–10% |
| Net revenue retention | 105–115% |
| Implementation/onboarding fee (one-time) | small relative to ARR |
| Checkout time reduction driving ROI for customer | 75% |

- **Scenarios (Base / Bull / Bear - which variables flex):**
  - Base: Moderate enterprise deal velocity; 1–2 chain deals per quarter in Year 1; device count mid-range.
  - Bull: Faster chain penetration (3–4 deals/quarter); higher device count per store; upsell to PC scale units.
  - Bear: Long sales cycles (9–12 months); pilot-heavy before conversion; low initial device count commitments.
  - Primary flex variables: new logos signed per quarter, avg devices per contract, ACV per device, ramp time to active phase.

- **Required sheets / outputs:**
  1. Assumptions dashboard (all drivers in one place)
  2. ARR build (new ARR, expansion ARR, churned ARR, net new ARR, closing ARR - by cohort/quarter)
  3. Revenue schedule (MRR → recognised revenue, accounting for dormant ramp period)
  4. P&L (Revenue, gross profit, S&M, R&D, G&A, EBITDA)
  5. Headcount plan (sales, implementation, engineering)
  6. Cash / runway (if raise size known - currently not in deck)
  7. Scenario toggle (Base / Bull / Bear)

## Frequently asked questions

### Is the Edgify financial model free?

Yes. The Edgify model is a free Excel download with live formulas.

### Can I change the assumptions?

Yes. The workbook is editable and its live formulas recalculate when assumptions change.