Four years ago, we decided not to fight the hardware. In May 2026, Stellantis made the same call public — at industrial scale.
Four years ago we incorporated Bonobo Services OÜ in Tallinn to bring uNobo Connect to the European market through Country Partners. The platform had been designed from day one around a clear product decision: to be hardware-agnostic. The reason was operational. uNobo Connect was built for operators who are not manufacturers — for those who buy the hardware rather than producing it. For a European non-manufacturer operator, tying yourself to a single manufacturer means tying your model. Chinese machines cost between 40% and 60% less than premium European ones. The incentives for the operator to buy direct from the factory were aggressive. They still are. Differentiation, for those who don’t manufacture, has to come from the layer above the hardware, not from the hardware itself.
The decision was to build the layer on top.
Stellantis confirmed it at industrial scale
That same month, Stellantis made public, at its scale, the same reading. The Villaverde plant in Madrid hands its production to Leapmotor — a Chinese EV maker — from the first half of 2028. Figueruelas, in Zaragoza, will build an Opel SUV co-developed with Leapmotor. Poissy, in France, closes as a car plant and is converted into a logistics centre, losing 800,000 vehicles of European production. And Stellantis also announced a new joint venture with Dongfeng to distribute Voyah, a Chinese premium brand, across Europe. The group’s net loss in 2025: €22.3 billion.
Stellantis is a visible case, but not an isolated one. It reflects a logic that any European non-manufacturer operator has been applying for years because the arithmetic pushes you there: hardware has stopped being the viable space for differentiation for those who do not manufacture. In EVs, in consumer electronics, in vending, in much of manufacturing machinery, China holds a cost and electronics advantage that the European non-manufacturer operator cannot ignore. Fighting that battle in 2026 is fighting a war that was decided a decade ago. The next ten years are played on a different layer.
The question, then, is not whether Europe will keep buying Chinese hardware. It will, and increasingly so. The question is what Europe builds on top. And that is exactly where the vending sector is still missing the conversation.
The sector is debating the wrong thing
In vending, two debates dominate trade shows and sector announcements. One is which European manufacturer is more reliable — a respectable debate, but a late one, because Hunan holds a generational lead. The other is whether SmartLink, the new communication protocol EVA presented at Venditalia in May 2026, will standardise the sector. SmartLink is a legitimate effort from the European side, in draft since late 2025, with a working demo on 7 May in Rimini and a realistic horizon for industrial adoption in 2027–2028. The protocol is welcome. But both debates — the European manufacturer and the protocol — are technical. And the end customer is not waiting for either.
The consumer already voted
Because the European consumer already voted — and not in a single market. McDonald’s operates virtually without traditional tills across its European restaurants: the order goes in by digital kiosk in Madrid, Berlin, Paris or Rome alike. More than 40% of global sales in the chain’s main markets already comes through app, kiosk or delivery. Inditex has removed or reduced physical tills across much of Zara and the group’s other brands throughout Europe: the customer pays for themselves, with their phone or their card, with no cashier in between. And underneath, Lidl, Carrefour, IKEA and Albert Heijn have made self-checkout a pan-European standard.
That consumer — the same one who passes through one or more of those formats every week — already knows what good unattended retail looks like. And in good unattended retail, payment is not a ritual. There is no keypad for entering numeric codes, no digging into a pocket for the exact coin, no bending down to retrieve a product that landed badly in a drawer. When that same consumer walks up to a traditional vending machine, the gap between what they already know and what they find is what decides whether they come back or not. For a machine manufacturer the central debate is legitimate: which chassis is more rigid, which telemetry finer, which protocol more open. That is what they get to debate from their model. For a European non-manufacturer operator the central debate is different: closing that gap. And that gap is not solved by the chassis. It has to come from the layer above.
What we built instead
That is what we decided not to do when we designed uNobo Connect. The platform was built not to tie itself to a manufacturer, because the European operator changes supplier every five or six years — and their commerce layer cannot change with them. Today uNobo runs on TCN, on Vendlife, on EasyTouch, on Jofemar, on Necta, on Bianchi. On Asian machines it comes in through the manufacturer’s API or SDK; it replaces the native sales app and leaves the screen showing the operator’s brand, catalogue, language and wallet. The hardware controllers stay with the manufacturer, because the mechanics are good. The layer on top is European end to end. On European machines with an MDB 4.x bus, it comes in through the bus and does the same from another technical door.
The platform was also designed not to charge the operator a percentage of every consumer sale. It is an alignment decision: the payment processor wins when there are more transactions; the operator needs to win when they sell better per transaction. A licence per machine, not a transaction commission. The gateway is agnostic — MONEI, Redsys, Nexi, Adyen, or whatever each market asks for — because the European operator works across different markets with different preferences, and tying them to a single gateway ties their hands. And the data lives on European infrastructure, with European sub-processors, with no international transfers. It is not marketing — it is what the serious operator starts to demand when their end clients are public administrations, hospitals, universities or chains with supplier policies.
On top of those design decisions sits the commerce layer that the sector has spent fifteen years confusing with telemetry: cross-machine combos, a consumer wallet with balance and traceability, ecommerce-style suggested selling, points-based loyalty, biometric age verification integrating Veriff or, in the Spanish market, Bouncer Digital. Composed with European sub-processors — not compliance merely asserted by Bonobo.
Four years in production, one market at a time
uNobo Connect has been in real production for four years through the pioneer Country Partner, uRetail S.L., in Spain. More than two hundred connected stores, eight hundred machines and one hundred and eighty operators in the network. That operational depth is not built in six months, nor with a funding round — it is built by operating, with a serious Country Partner in the first market, gathering data, adjusting, failing, correcting. It is the hardest asset for any later entrant to replicate.
Bonobo operates from Tallinn with European sub-processors. The data does not leave the EU. uRetail S.L. is today the only active Country Partner in the network. Portugal, Italy, France and Germany are the priority markets for adding a Country Partner in the coming months, and the rest of Europe is equally open, with no partner assigned in each market.
If you lead an automated retail network in one of those markets and this analysis resonates, the conversation with whoever wants to be first in their market is a different one from the conversation we will have once the network is complete. Write to info@bonoboservices.com or use the form at bonoboservices.com.
Tomorrow there will be more Country Partners. Today, there are not.
Román Suárez — Founder, Bonobo Services OÜ