The most revealing moment in an autonomous store is not when somebody walks out without paying at a register. It is ten seconds earlier, when they pick up a sandwich, reconsider, put it back and choose a salad. A conventional sales system records one purchase. AiFi wants to understand the small choreography that produced it: the hesitation, the shelf, the traffic around it and the final basket - without asking the shopper to scan anything.
That ambition has carried the San Francisco company a long way from its original billing as an Amazon Go rival. Founded in 2016, AiFi built a camera-led system that follows anonymous shoppers and products through a store. By September 2025, it said its technology had been deployed in more than 300 autonomous stores. The network ranges from Żabka convenience shops in Poland to concessions inside the LA Clippers' Intuit Dome.
The sharper way to understand AiFi in 2026 is not as a company deleting cashiers. It is as an enterprise software and infrastructure business trying to give physical operators the kind of observability online businesses take for granted. A website owner can see a funnel. A stadium manager often sees a queue after it forms. AiFi's cameras turn movement into structured events, building a live 3D digital twin that can be searched, measured and connected to other systems.
A receipt assembled in the ceiling
In the familiar version of the product, a customer taps a payment card or opens an app at the entrance. Overhead cameras see a set of skeletal keypoints rather than needing to identify a face. Computer vision links the person's movement to products lifted from shelves. If a bottle goes back, it leaves the virtual basket. When the shopper exits, the system charges the payment method and sends a receipt.
The experience is deliberately uneventful. That is the point. Nobody has to line up, find a barcode or summon an employee because a self-checkout machine disapproves of the bagging area. For a convenience store, the promise is 24-hour access without staffing a till around the clock. For a stadium, it is throughput during the brutally short interval when thousands of people all want a drink.
AiFi packages that capability in several shapes. Refresh retrofits an existing store. Build covers a new, customized location. To-Go is a modular shop that can be dropped into an unconventional footprint. The platform can work camera-first and add RFID where clothing, irregular products or a retailer's preferred journey make radio tags useful. It connects to payments, point of sale, ERP and inventory systems rather than demanding that a customer abandon the rest of its stack.
“Autonomous shopping will become so commonplace it will simply be shopping.”João Diogo Falcão, CEO of AiFi
The second product hiding inside the first
Checkout was a useful wedge because everybody understands the irritation of a line. But once software can reliably connect people, objects, places and time, the basket is only one possible output. The same system can reveal a blocked aisle, an empty shelf, a suspicious pattern, a poor display, a kitchen bottleneck or the precise moment a concourse turns into a traffic jam.
AiFi calls this broader layer spatial intelligence. At its center is eddie, short for Event Driven Data Interpretation Engine. Built on Microsoft Azure, the engine converts streams of camera activity into data that software can query. Microsoft's 2026 account of the partnership said the platform was analyzing more than 60 million human interactions each month. It also reported customer results including 25 to 40 percent reductions in shrink and venue throughput gains of as much as 300 percent. Those are customer outcomes, not a universal guarantee, but they explain the size of the prize.
This is where AiFi moves into a much larger, fuzzier market. Retail analytics companies measure traffic. Loss-prevention systems flag theft. Computer-vision platforms watch factories. Airport technology vendors manage identity and flow. AiFi is proposing one perception layer that can support several of those jobs, with APIs and partner integrations carrying the insight into whatever tool an operator already uses.
Who buys it, and why
The buyer is a business that owns complicated square footage. Grocery and convenience groups use AiFi to automate payment, monitor shelves and operate more hours. Stadiums and arenas use it to move fans through concessions before play resumes. Universities and workplaces can run unattended markets. Hospitals can keep gift shops accessible to nurses and visitors at odd hours. Quick-service restaurants and travel hubs are newer territory, where the company pitches workflow, dwell, identity and flow analytics.
A 2025 rollout with Lori's Gifts makes the value concrete. The hospital retailer connected AiFi to its ERP, synchronized sales data and used in-store tablets to collect email addresses for receipts. The cleverness is not a futuristic shop floating apart from the organization. It is the opposite: a new front end tucked into old accounting and inventory routines.
Build the basket in the background, extend opening hours and inspect product interaction.
Process compressed demand while measuring traffic and improving concession layouts.
Find bottlenecks, compliance gaps and inefficient use of a physical environment.
Combine anonymous flow with optional identity systems for contactless corridors.
AiFi makes money as a B2B technology provider. Public pricing is not posted because the job depends on the footprint, camera and compute requirements, integrations and rollout format. The commercial relationship spans software, edge and cloud infrastructure, deployment work and continuing support. That makes it closer to enterprise infrastructure than a downloadable software subscription, even though recurring platform economics sit inside the model.
A crowded aisle
AiFi does not own the idea of walking out. Amazon sells Just Walk Out to third parties. Zippin, Trigo, Standard AI and PIXEVIA pursue overlapping ground, while smart carts, RFID gates and ordinary self-checkout attack the same labor and waiting problem from different angles. Grabango, once a prominent rival, ceased operations in 2024 - a reminder that clever demos do not automatically create durable retail economics.
AiFi's differentiation is practical rather than magical. It emphasizes off-the-shelf cameras, no mandatory weight sensors, privacy-minded keypoint tracking and a system that can bend around an existing shop. Retailers can choose gated, app, card or hybrid entry. They can add RFID or, in specific travel scenarios, an identity partner. They keep their own storefront and integrate their own systems. For retailers wary of handing their store data to Amazon, independence is itself part of the product.
Flexibility also creates difficulty. Real stores contain occlusion, children, crowds, reflective packaging, misplaced products and customers who do surprising things. Every extra integration makes the solution more useful and the delivery job more complex. AiFi's advantage will be decided less by whether computer vision works in a clean demonstration than by whether deployments remain accurate, economical and supportable across hundreds of messy rooms.
The checkout line is the visible pain. The deeper business is giving a room a memory.The strategic shift from autonomous retail to spatial intelligence
The company behind the cameras
AiFi was started by computer-vision researchers Steve Gu and Ying Zheng, a married couple with experience at Google and Apple. The company now describes João Diogo Falcão as its technical co-founder. Falcão, whose background includes robotics at Cornell and cyber-physical systems at Carnegie Mellon, led the proprietary platform as CTO before becoming CEO in September 2025. Steve Carlin, recruited in 2022 to professionalize growth and partnerships, returned to TransLink Capital and stayed on as an adviser.
The corporate structure mirrors the product's geography. AiFi says it is fully remote, with more than 100 employees representing 22 nationalities and a workforce and customer base spanning 16 countries. Its investor list mixes venture firms with companies that can help distribute or validate the technology: Verizon, Qualcomm and HP on the infrastructure side; ALDI, Żabka and REWE on the retail side. A $65 million Series B in 2022 was the defining round. A reported $6.7 million extension followed in 2024.
By 2025, AiFi said it had served more than 9 million shoppers and processed 7.1 million orders. It ranked No. 326 on the Inc. 5000 after reporting 1,208 percent revenue growth over three years. The numbers describe a private company with real deployment experience, but not a settled category. Autonomous checkout still competes with a cheap human register and a familiar self-checkout kiosk. Spatial intelligence must prove that more data leads to better decisions, not simply a more beautiful dashboard.
The next five feet
Recent projects show where AiFi is testing the boundaries. At Co-op Live in Manchester, a four-gate autonomous market joins the platform to Boxbar's self-serve draught system, including Guinness. The camera system has to understand not just a can lifted from a shelf but a pour from a tap, then make the entire visit feel like one transaction. At Orlando International Airport, work with Paravision and Embross explores a contactless corridor that can connect identity and movement without making every traveler stop.
These are less like stores than coordinated environments. The software has to decide what happened, attach it to the right session and tell another system what to do. If AiFi succeeds, the most valuable output will not always be a receipt. It might be an alert that prevents an empty shelf, evidence that redesigns a queue, or an answer to a question nobody could previously ask without standing in the room with a clipboard.
The company's mission is unusually literal: make the physical world accessible and useful as data. That future will require restraint as well as accuracy. A space that can be searched can also be surveilled. AiFi's anonymous-by-design approach and optional identity layer draw an important boundary, but customers, regulators and shoppers will keep testing it. The technology becomes more interesting as it sees more. The company becomes more trustworthy when it knows what not to remember.
For now, the cleanest demonstration still fits in a shopping trip. Enter. Take a sandwich. Put it back. Choose the salad. Leave. The receipt appears, and somewhere behind it the store has learned a little more about itself.