Roughly 200,000 venues pour at volume. None of them can see the loss in real time.

The honest market story leads with the serviceable wedge and treats the full industry universe as a ceiling, never as an addressable base. The Bureau of Labor Statistics counts roughly 696,000 privately owned food-service and drinking establishments under NAICS 722, the universe of restaurants, bars, caterers, and other eating places. Most of them, the fast-food counters, coffee shops, and juice bars, pour little or no liquor and would never buy a liquor-inventory mat. Presenting that whole universe as the market would be a top-down overstatement. We do not.

The segment that actually pours at volume is far smaller. Dedicated drinking places under NAICS 72241 plus the large share of full-service restaurants that run a real bar program form the serviceable market. The five-year obtainable share is a low single-digit slice of it, and it reconciles to the bottom-up revenue model within a healthy margin. The market exists and is large. The question that matters is execution, not whether the room is big enough.

The Canada and Caribbean extension sometimes cited alongside the US figure is [pending source verification to Statistics Canada and a named Caribbean source] and is excluded from the headline until it is sourced.

Each layer narrows from a category ceiling to a model-reconciled target. Every figure is either tied to a named source or explicitly flagged as a management estimate. Nothing is asserted without a citation or a flag.

flowchart TB
  TAM["TAM: ~696,000 US food and drink establishments<br/>BLS NAICS 722 (government data)<br/>category ceiling, not addressable base"]
  SAM["SAM: ~200,000 liquor-serving venues, $200K+ annual liquor sales<br/>management estimate, flagged"]
  SOM["SOM (Year 5): 6,000 venues, 3.0% of SAM<br/>reconciles to the revenue model"]
  TAM --> SAM --> SOM

The category frame is well established in the trade. Bar and restaurant operators commonly lose on the order of 20 percent of profit to shrinkage, and an acceptable shrinkage rate is generally cited as under 2 percent of inventory. Shrinkage is not only theft; it is over-pouring, spillage, and unrecorded comps, the everyday leakage no clipboard count catches in time.

The 20-percent and sub-2-percent figures are Sculpture Hospitality and WISK (trade and industry tier). The ~$162B figure is a USDA-derived restaurant waste anchor cited across the trade press; it is a waste number, broader than shrinkage. The $55B annual-loss and 22-percent-shrinkage figures that circulate in beverage marketing material are excluded here: they could not be tied to a named, dated source, so they are not shipped.

The per-venue economics are sourced to Klynkz deployment results, which Klynkz sources before they ship. The macro context above is presented only where it carries a citation; the per-venue recovery is the figure that carries the case.

The category today is fragmented, software-led, and hardware-light. The tools that exist still depend on a human to capture the data. The competitor the feature grid usually omits is the status quo: a large share of independent bars still run a manual weekly count because it is free and universal. That do-nothing manual count is the true incumbent, and the answer to it is the per-venue recovery, not a feature war.

One issued patent, US 11,983,670, already extends the sensing system to solids, liquids, chemicals, and parts. That makes the same mat technically credible in medical supply, government and military, chemical inventory, and retail parts. The discipline that matters: this is granted optionality backed by issued IP, not the basis of the valuation. The core story is hospitality, and the model and the valuation stand on hospitality alone.

The defensible exit band on the comparables that change hands today is roughly 7x to 11x Year-5 revenue, anchored low by the most relevant transaction and high by the closest public analog. 12x is optimistic, not conservative, and 20x is not supported by any current comp, so it is retired. The full treatment, with the implied valuation range, lives on the exit and valuation pages.

See the exit page for the full comparable set and the implied valuation range, and the valuation page for how the multiple maps onto the raise.

Four forces have converged to open this wedge at once. None of them was true a few years ago, and together they turn a manual, tolerated cost into a problem an automatic system can finally solve at scale.

The raise funds distribution into a proven, protected position. The hardware is deployed, the patents are issued, and the white space is open, which is lower-risk capital than funding invention.