memo

Forge builds autonomous kitchen cells that drop into an existing restaurant in one day and replace two line cooks per shift. The cells are sold as a service at $6,000 per month, with no capex on the operator's balance sheet.

This memo accompanies the Forge Seed pitch deck and covers the same fifteen sections in long form. Skim the index; read deeper where the deck is thin.

Forge is raising $3M at $15M post-money to scale to twelve cells deployed and a ten-engineer team in eighteen months, bridging to a Series A at month 18-24.

American restaurants cannot hire line cooks. There are roughly 180,000 unfilled line-cook positions across US fast-casual today, and labor cost has risen 28% since 2021 in major metros.

The problem compounds. Annual turnover sits at 75% per cook station, which means every line is in a state of constant retraining. Output suffers, reviews suffer, repeat visits drop, and the labor bill keeps climbing.

Operators have tried the standard playbooks. Higher wages do not solve a labor pool that is shrinking faster than wages can rise. Off-shore staffing does not apply. Service-counter automation helps the front of house but leaves the kitchen line untouched.

Forge ships a self-contained robotic cell that takes over one cook station. It wheels into an existing kitchen on a shipping pallet, plugs into a single 240V line, and replaces two line-cook shifts.

The operator keeps the recipe, the menu, and the customer experience. Forge runs the equipment. The cell goes from purchase order to first plate in fourteen days, and from there it does not call out, does not turn over, and does not get worse with overtime.

The business model is Robot-as-a-Service: $6,000 per cell per month, all-in, on 24-month contracts. Operators see ROI in month one against the labor cost the cell replaces.

The Forge cell has three pieces that matter.

The manipulator arm is a six-axis robot tuned for the chaos of a kitchen line, with end-effectors that handle wok, tongs, ladle, and sheet pan without re-tooling.

The sensing stack uses computer vision and thermal imaging to read the food and the pan in real time, tracking doneness and oil temperature without thermometers or scheduled timers.

The Recipe OS translates a chain's existing SOPs into cell actions, so the chain does not have to re-engineer its menu or retrain its operators. Together, these three pieces turn two cook stations into one cell that never calls in sick.

Deployment is engineered to look like installing a new piece of kitchen equipment, because that is what restaurants understand.

A Forge tech does a one-day site survey, picks the right cell type for the menu and floor plan, and confirms electrical and exhaust readiness. The cell ships within ten days and gets wheeled in during a single closed shift.

Forge chef-engineers encode the chain's recipes into the cell in 72 hours, validated against a live tasting with the chain's own QA team. The cell then goes hot during a closed pilot shift with a Forge tech onsite, and only joins the daily line once the operator signs off. Remote ops monitors every cell every shift after that, 24/7.

Restaurants today pay for kitchen labor that is expensive, unreliable, and difficult to scale. The Forge cell replaces that labor with equipment.

Versus retrofit competitors like Miso Robotics, Forge ships a purpose-built cell, not a bolted-on arm. Versus traditional automation projects, Forge is a 14-day install, not a 9-month integration. Versus more cooks, Forge is a fixed monthly cost that does not turn over.

The operator's math is simple. Two line cooks fully loaded run $7,000 per month in a major metro. One Forge cell costs $6,000 per month. The cell pays for itself in labor savings the first month, and the operator does not have to recruit, train, schedule, or replace.

Five features make the Forge cell viable in a real restaurant rather than a lab.

Drop-in install means one closed shift of kitchen downtime, not a structural retrofit. Recipe-agnostic onboarding means the cell works for the chain's existing menu, not a Forge-designed menu. Remote ops 24/7 means a real engineering team watches every cell, every shift, and intervenes before a service is at risk.

Uptime in production pilots has averaged 98% across the last two quarters. Per-cell pricing is all-in: $6,000 per month covers hardware service, parts, software, and remote ops. No surprise invoices.

Total addressable market is the US restaurant industry at $920B in annual spend, of which roughly $38B is the line-cook labor cost that Forge is positioned to replace.

Serviceable addressable market is the top 100 US fast-casual chains, which collectively run more than 80,000 stores. At an average of two cells per store, that is 160,000 cell slots, or roughly $11.5B in annual contract value at current pricing.

Serviceable obtainable market over the next five years is roughly $2.1B, representing 5% adoption by the top 100 chains. Forge's plan reaches 700 active cells in year five, well inside that envelope.

Revenue is straightforward. Each cell rents at $6,000 per month on a 24-month contract. The unit economics are an industrial recurring-revenue business with a hardware-services gross margin profile.

Per-cell COGS runs about $2,110 per month: $1,460 in cell depreciation (on a $70k build cost amortized over 48 months), $400 in maintenance and truck rolls, and $250 in remote-ops allocation. That yields a 65% gross margin per cell, which is the ceiling for hardware-services hybrid businesses.

Contracts are priced for the operator's ROI, not Forge's GM. The next pricing lever is multi-cell discounts and category-specific cell types, both expansionary not promotional.

Go-to-market is structured around the way fast-casual chains actually buy equipment.

The entry deal is one to three pilot stores at the regional chain level, with the contract running through the chain's HQ procurement team. Industry events — NRA Show, FSTEC — place Forge in front of every decision maker once a year and have driven the existing pilot pipeline.

For scale, Forge will lease cells through foodservice integrators like Welbilt and Heritage, which already hold the operator relationship and the financing infrastructure. Andre Park personally runs the ops conversation in every founder-led sale; conversion is four times higher than cold-sales-rep teams across the existing pipeline of nine production conversations.

The Forge team is purpose-built for the kitchen-cell problem.

Mia Voss spent nine years at Apple as an industrial designer, most recently leading the Vision Pro thermal cell. Before Apple she cooked professionally. She runs the company.

Andre Park ran East Coast kitchen operations for Sweetgreen through their hyper-growth from 30 to 180 stores. He knows what a line cook does, what franchisees will pay, and what breaks under load.

Dr. Hannah Wei is the technical co-founder, an MIT PhD in dexterous manipulation and ex-Boston Dynamics on the Atlas team. She owns the part of the cell that has to actually pick up a soft tofu cube without breaking it.

Forge's defensibility is the combination of three things competitors cannot copy quickly.

The cell is purpose-built. Miso Robotics and Chef Robotics bolt arms onto existing kitchen lines, which means they inherit the chaos of those lines and require months of integration. The Forge cell is one piece of equipment with a known shape that drops into a known footprint.

The Recipe OS lets Forge serve any chain. Competitors lock to one or two cuisine types per cell. Forge wok cells, pasta cells, and grill cells share 80% of their architecture; only the end-effector and the cell-side OS plug-ins change.

The team. Apple industrial design, Sweetgreen operations, Boston Dynamics manipulation. No competitor has the cross-stack.

Forge has built deliberately over two years to reach this point. The first prototype wok cell was on the bench in Q1 2024. The first pilot was signed Q3 2024. The first cell was live in a commercial kitchen Q1 2025, and three paid pilots have been running across wok, pasta, and grill cells since Q3 2025.

The next 18 months are about closing the Seed, scaling engineering from three to ten, and deploying twelve cells across the first non-pilot chain contract. The Series A close at $1M+ ARR is the trigger for the equipment-financing facility that funds fleet growth.

By year five, Forge is operating 700 cells across more than 50 chains, running on a balance of equity and asset-backed debt, and approaching EBITDA breakeven.

Year five revenue lands at $37.8M with 700 active cells, 65% gross margin, and EBITDA approaching breakeven. EBITDA-positive sits in year six on the current ramp.

The forecast is conservative on cell-deployment pace and aggressive on capex efficiency. The Seed round covers Y1 capex of $630k (nine new cells) plus the first eighteen months of operating loss; an equipment-financing facility is required by mid-Y2 to fund the fleet beyond that point.

Cash burn averages $2-3M per year through Y4, then drops below $1M in Y5 as gross profit catches operating cost. The Series A is a bridge, not a rescue.

Forge is raising $3M at $15M post-money on priced equity. The round closes in September 2026 and funds the next eighteen months: twelve cells deployed, a ten-engineer team, and the first non-pilot chain contract live in production.

Use of funds: 45% engineering team build, 20% Y1 cell capex, 15% GTM team, 10% deployment ops, 10% G&A and working-capital buffer. The Series A at month 18-24 unlocks the equipment-financing facility that scales the fleet from 12 to 45 cells and beyond.

What we want from this round are partners who understand both hardware and services, who can help us pick the right foodservice integrators, and who will be in the room when we hire the next ten engineers.