Trigo

Snapshot

Trigo is an Israeli computer-vision company that turns a conventional supermarket into a checkout-free store without replacing a single shelf or scale. Shoppers scan a QR code at the entrance, grab what they want from the regular aisles, and walk out; a lattice of ceiling cameras and Trigo's recognition stack bills the right card for the right basket. The company has raised over $200 million since 2017 and has shipped the technology into live grocery stores for Tesco (UK), REWE (Germany), ALDI Nord (Netherlands), Netto (Germany), Shufersal (Israel), Auchan (France) and others. It matters now because Amazon spent most of 2024 ripping Just Walk Out out of its own full-size Fresh stores, leaving the thesis — that computer vision can replace cashiers at scale — up for grabs. Trigo is the credible non-Amazon answer, and the attempt is being tested in European banners that cannot afford to be wrong.

Founding story

Trigo was founded in 2017 in Tel Aviv by brothers Michael Gabay (CEO) and Daniel Gabay (CTO) with Yuval Alaluf. Michael had come out of Unit 8200, Israel's signals-intelligence corps that has produced a disproportionate share of the country's computer-vision founders; Daniel's background was in machine-learning research. The initial pitch was explicit: Amazon Go had just opened in Seattle and the Gabays bet that every grocer in Europe would need the same capability but would never buy it from Amazon. Rather than build out their own stores, they would sell retrofit technology to retailers who already had the real estate and the loyalty cards. Early capital came from Red Dot, Hetz Ventures, and Vertex, who understood that Israel's edge in multi-camera tracking (built up through border-surveillance work) could be repointed at consumer retail.

How it works

A Trigo store looks ordinary. Ceilings hold a dense array of standard-definition RGB cameras — typically one every two to three meters, numbering in the hundreds for a 500-square-meter urban format and over a thousand for a large box — connected to on-prem GPU servers. Trigo's models fuse the camera streams into a single 3D scene, tracking each shopper as a skeletal pose and each product as an object. When a hand picks up a bag of pasta, the system associates that SKU with that shopper's virtual cart. Weight-sensing shelves are optional and used sparingly, generally only for high-shrink categories (deli, wine, infant formula) where visual verification is weak. Payment is tokenised to the account that scanned in. The architecture is deliberately edge-heavy: data stays on the retailer's premises for GDPR and for latency, with only cart events and anonymised analytics hitting the cloud. Trigo's installation playbook has compressed meaningfully — according to the company's own 2022 and 2024 statements, retrofit calibration dropped from weeks to days as its auto-mapping tools matured.

Product and business overview

The core offering is EasyOut, the autonomous-checkout runtime. On top of it sit three paid modules: StoreOS, an operations layer that provides live planogram compliance, out-of-stock alerts, and shrink monitoring; ShopperDNA, a privacy-preserving behavioural-analytics tool that tells grocers how shoppers actually move through aisles; and a dynamic-pricing module piloted with Shufersal. The company has also licensed the stack as a "smart cart" option for retailers who do not want ceiling cameras, repackaging the recognition models for in-cart hardware.

Business model and pricing

Trigo sells to grocers under a hybrid structure: an upfront installation fee covering cameras, servers, and store calibration, plus a recurring per-store software fee tied to transaction volume. The company has not published prices, but trade coverage from 2022–2023 placed installation at roughly €0.5–1 million for a 300–500 m² urban store, with ongoing fees in the low six figures per store per year. That is cheap next to the labour it replaces in high-wage European markets — a 24/7 convenience store on three shifts of cashiers runs well above €300,000 in staff cost — but expensive next to a self-checkout kiosk at €15,000. The sell is not labour replacement in isolation; it is loss prevention, longer trading hours, and richer first-party data.

Traction over time

Date	Milestone
2017	Founded in Tel Aviv.
2019 Sep	$22M Series A led by Red Dot and Vertex (per Calcalist, Sep 2019).
2020 Jun	$7M top-up; first Shufersal pilot in Israel.
2021 Jan	$60M Series B led by 83North, Tesco joining as strategic investor.
2021 Oct	First live Tesco GetGo store opens in central London (Oct 2021).
2021 Oct	First autonomous REWE store opens in Cologne.
2022 Jul	ALDI Nord opens the Netherlands' first frictionless store with Trigo.
2022 Oct	$100M round led by SoftBank Vision Fund 2 (per TechCrunch, Oct 2022).
2023–2024	Deployments at Netto (Germany), Auchan (France), additional REWE sites.
2025	Company reports over 50 live stores across 8 European countries (per BusinessWire).

Headcount reached roughly 230 by late 2023 per LinkedIn, concentrated in Israeli R&D with a thinner EU customer-success footprint. Trigo has not published revenue; CB Insights last pegged its valuation at ~$500 million post-money following the 2022 round.

Market analysis

The addressable segment is European and UK grocery, where the top-20 chains run roughly 170,000 stores and employ over 3 million people. Grocery labour has structurally tightened: UK retail vacancies rose 15% between 2019 and 2023 (per BRC); German retail unions settled 2024 wage deals at 7%+. Autonomous retail has a theoretical TAM calculated by Grand View Research of $21.4 billion by 2030, growing 20% CAGR, though those forecasts predate the Amazon Just Walk Out pullback. The more honest frame: Trigo competes for the convenience, forecourt, and corporate-campus formats where cashier cost is highest, store footprint is smallest, and shrink is a real line item. That is a few tens of thousands of stores globally — a meaningful business, but not a trillion-dollar one.

Competitive intel

Amazon Just Walk Out is the elephant: having pulled it from its own Fresh stores in April 2024, Amazon pivoted to selling it to third parties (airports, stadiums, campuses). It remains the technical reference but has lost first-party conviction. AiFi (US, ~$75M raised, per Teardown's own prior coverage) focuses on nano-stores and smart carts and signed deals with Carrefour and Verizon; it is Trigo's closest Western rival. Grabango (US, $73M raised) shut down in May 2024 after failing to close a C round, removing one competitor and vindicating Trigo's European geography. Standard AI (US) pivoted away from cashierless to adjacent retail analytics after burning through most of its $230M. Zippin (US, ~$33M) targets small stadium formats and remains niche. On smart carts specifically, Instacart's Caper (acquired 2021) is the US baseline, and Shopic (Israel, ~$55M) is Trigo's local sibling with a growing European footprint. Trigo's real moat is the live installed base in Tier-1 European banners; computer-vision research is commoditising, but a REWE operations team is not going to re-tender a working store.

History and evolution

The company's arc has three phases. Phase one, 2017–2020, was lab mode: hiring researchers, winning the Shufersal pilot, and quietly building the ceiling-camera system in a Tel Aviv showcase. Phase two, 2021–2022, was the European banner-signing sprint; the Tesco GetGo opening in High Holborn in October 2021 and the REWE Cologne opening weeks later were the sales catalysts, and the 2022 SoftBank round rode that momentum. Phase three, 2023–present, has been a grind: the macro turned, retail IT budgets contracted, and the gap between pilot stores and chain-wide rollout widened. Several banner relationships that started as flagship pilots have not grown to the second or third store. Trigo has responded by diversifying — pushing the StoreOS and ShopperDNA layers to retailers that are not ready to go fully autonomous — and by cutting headcount through 2024 to extend runway. Reports in Israeli press through 2024 and 2025 described a company that is still live and shipping but is a leaner operation than the 2022 pitch deck implied.

What people say

The case for. Shoppers who have used the Tesco GetGo store report a genuinely novel experience — grab, walk out, receipt arrives — and Tesco itself called early results "encouraging" in trade press. REWE has publicly stated that its Cologne autonomous store generated higher basket sizes than comparable staffed stores, a claim repeated by Business Wire in October 2021. German retail analysts have praised the vision quality: Trigo's cameras read SKUs correctly on messy shelves where Just Walk Out's reportedly struggled. The second recurring praise point is the retrofit model — grocers do not need to build a new store to try it.

The complaints. The chief complaint, aired repeatedly in trade press and in a lengthy 2024 Calcalist feature, is that the pilots stopped being pilots and started being flagship PR assets. Chain-wide rollouts have not materialised; several European banners publicly committed to "dozens" of stores and ended with one or two. The cost per store remains too high for the median convenience format. Employees on Glassdoor, as of mid-2025, describe a 2024 reorganisation and layoffs; the recurring theme is that commercial velocity lagged technical capability. Short-sellers (there is no public float to short) and skeptical analysts note that Amazon's own retreat is less a tailwind than it looks — Amazon pulled the technology because operating-cost economics were worse than cashiers, not because customers disliked it.

Outlook: the open question

Does retrofit computer-vision checkout ever cross from flagship-store novelty into chain-wide default, and does Trigo capture that transition before its enterprise sales cycle outlasts its runway? For yes: Trigo needs at least two Tier-1 European banners to commit to triple-digit store rollouts at prices that still fund a software margin, and it needs StoreOS and ShopperDNA to carry stand-alone revenue so the business is not binary on autonomous checkout. For no: the pattern of the last three years — strong pilot, no rollout, cost per store that defies compression — continues, and the category looks like one Amazon tested, found unprofitable, and left for someone else to prove unprofitable twice.

How to attack it

The attack is to collapse the two things Trigo bundles: the recognition model and the ceiling-camera deployment. A new entrant could ship a smart-cart-only stack that uses a $300 bolt-on camera-and-scale rig on existing shopping carts, licence the model to the retailer, and skip the construction project entirely. Shopic has started down this road in Israel; Instacart's Caper owns the US. The wedge is a modular retrofit where the retailer pays per cart, not per store, and the entrant collects a recurring software fee on roughly one million carts across European mid-market grocers. The core weakness in Trigo's position that invites attack is the installation cost — €500,000+ per store with a six-month calibration tail is a hard sell to a Tesco Express or REWE To Go franchisee operating on thin margins. A second weakness is the data-locality constraint: because models run on-premises, Trigo cannot amortise learning across the fleet the way a cloud-native rival could. A third is channel: Trigo sells direct to grocery CIOs with 18-month cycles, while a smart-cart entrant could sell through loyalty-platform partners (Dunnhumby, Eagle Eye) with a far shorter path to pilot. Finally, Trigo's price list is tied to installation hardware; a software-only pricing model (per basket, per month) would be lower-risk for the retailer and easier to expand in the way SaaS has beaten perpetual-licence enterprise software across every other category.

Adjacent-segment play

The same multi-camera 3D-tracking stack has three natural adjacencies. Convenience and forecourt is the most obvious — gas-station c-stores run on a single cashier per shift and lose 2–3% of inventory to shrink; Trigo's economics work better there than in grocery. Couche-Tard, Circle K, and ampm are the buyers; EG Group is already piloting with another vendor. Stadiums and venues is the Amazon pivot route and the one with the fewest incumbents, though volumes are seasonal; Rebel Foods and Mashgin dominate the North American stadium stack, but Europe is open. Warehousing and dark stores is the under-appreciated adjacency: the same tracking model that bills a shopper can audit a picker's pick-rate and shrink inside a micro-fulfilment centre. Ocado's robotic fulfilment and the Picnic model (see today's companion page) both tolerate human picking in some workflows, and no vendor currently owns the "warehouse-shopper tracking" category. The wedge does generalise, but not to US big-box grocery, where cashier wages are lower and self-checkout has already captured most of the labour savings. On the other end, luxury retail is unlikely to adopt — the whole point there is attentive human service. The practical adjacent play is the one Trigo has signalled with StoreOS: become the operations/analytics layer on top of any store, autonomous or not, and let autonomous checkout be a premium SKU rather than the whole business.

Sources and further reading

• Trigo (company) — Wikipedia (reference, accessed Nov 2026)
• REWE Group, Viola Growth investment takes Trigo's total funding over $100M — BusinessWire, June 2021
• Trigo powers fully autonomous Tesco grocery store in Central London — BusinessWire, October 2021
• Trigo raises $100M to expand its Amazon-style cashier-free store technology — TechCrunch, October 2022
• Trigo, ALDI Nord open Netherlands's first AI-powered frictionless supermarket — DutchNews/BusinessWire, July 2022
• Amazon pulls Just Walk Out from Fresh stores — Grabango shuts down, May 2024
• Tesco-backed checkout-free tech firm Trigo secures $60m funding — The Grocer, 2021
• Israeli store digitalization co Trigo raises $100m — Globes, October 2022