BREAKING  Yili Wu turns ordinary shelves into software 30,000 SKUs recognized at 99.7% accuracy SandStar operates across ~12 countries Unmanned store built for roughly 5% of Amazon Go Beijing ↔ Charlotte: one company, two continents Before AI retail, he built a 3D pancake printer BREAKING  Yili Wu turns ordinary shelves into software 30,000 SKUs recognized at 99.7% accuracy SandStar operates across ~12 countries Unmanned store built for roughly 5% of Amazon Go Beijing ↔ Charlotte: one company, two continents Before AI retail, he built a 3D pancake printer
Founder · CEO · SandStar

Yili Wu

He pointed a camera at the shelf and asked a stubborn question: why can't a store see what it sells?

Yili Wu, founder and CEO of SandStar
Yili Wu - the salesman who became an engineer of seeing machines.

Profile · Technology & Retail · Filed from Beijing and Charlotte

The Story

A store that watches itself, priced for the corner shop

Walk into one of Yili Wu's stores and nobody greets you, because the room is already paying attention. Cameras overhead read every item you lift, every shelf you empty, every second you linger. You grab what you want and leave. The receipt finds you later. This is SandStar, the company Wu founded in 2016, and its promise is unglamorous and enormous at once: give physical retail the same nervous system that e-commerce has had for two decades.

For most of the last century, the shelf was blind. A website knew exactly what you clicked, hovered over, abandoned in a cart. The bodega on the corner knew almost nothing. Wu, who spent roughly eight years selling enterprise software at Oracle and IBM, kept bumping into that asymmetry from the outside. His customers were pouring money into digitizing everything online while the offline half of their business ran on guesswork and paper.

His fix was computer vision. SandStar's engine, by the company's account, recognizes more than 30,000 distinct products at 99.7% accuracy. Point it at a vending machine and you get checkout-free snacking. Point it at a room and you get an unmanned store. Point it at a working shop and you get analytics on what shoppers actually touch. Four product lines, one underlying idea: turn the physical world into data without asking the customer to do anything different.

The economics are the quiet punchline. Amazon Go made checkout-free shopping famous and expensive. SandStar's version of an unmanned store reportedly costs about 5% of that build - the same disappearing-cashier magic at a price a franchisee might actually sign for. Wu did not set out to beat Amazon at spectacle. He set out to make the technology cheap enough to be boring, which is how technology wins.

Today SandStar operates across roughly a dozen countries and counts more than twenty Fortune 500 companies among its clients. It runs on two continents at once, with bases in Beijing and Charlotte, North Carolina. For a company that started life under a different name and a founder who started his career carrying a sales quota, that is a long way from the corner store.

By The Numbers
30K+
Products recognized
99.7%
Vision accuracy
~12
Countries served
20+
Fortune 500 clients

SandStar begins from sand and dust, reaching for the stars using the technology gateway.

- Yili Wu, on renaming the company
Origins

Before the stores, there were the pancakes

The tidy version of a founder's story runs in a straight line. Wu's does not. In 2015, a year before SandStar, he and a group of classmates from Tsinghua University's School of Software built a 3D printer that prints pancakes. Not a metaphor - an actual machine that takes a photo or a bit of text and squeezes batter into a Chinese jianbing bearing your portrait, a cartoon, or a city skyline.

It began, as the good ones do, at a class reunion. One of his co-founders later recalled the moment plainly: the idea to make a 3D pancake printer came up over conversation, and to everyone's mild surprise, it worked. Seven Tsinghua graduates, one griddle, and a device that fused a printing-technology company with a pancake chain.

You can dismiss it as a novelty. Wu clearly didn't. The pancake printer taught him the hardest lesson in hardware - that shipping a physical product to real customers is a different sport from writing software - and it put him in a room full of engineers who liked strange, tangible problems. When he turned that instinct toward retail, he already knew how to make cameras and machines behave in the wild.

He is, by training, one of the early graduates of Tsinghua's software school, class of 2003. By trade, before founding anything, he was a salesman - and a very good one. As the top performer leading IBM China's e-commerce sales team, he is credited with driving the group to roughly 80% market share. That combination, an engineer's education and a closer's instincts, turns out to be the whole story.

Brick-and-mortar operations could provide a better shopping experience if they underwent digital transformations.

- Yili Wu, on the gap that started SandStar
The Contrarian Move

He sold it backwards, and that is why it sold

Most startups begin at home. Prove it locally, earn a few reference customers, then expand. Wu inverted the map. Rather than starting in China, he aimed SandStar's pitch at corporate headquarters in the United States and Europe - the rooms where global rollouts actually get decided.

It sounds reckless until it works. In 2018 he demonstrated SandStar's smart kiosks at an innovation conference in Atlanta, in front of the chief innovation officers of large companies. One of those conversations became a contract with Coca-Cola. Selling top-down meant that a single yes could unlock deployments across many markets at once.

This is where the IBM years quietly pay off. Enterprise sales is not about the product demo; it is about knowing which door to knock on and how trust compounds. Wu had spent the better part of a decade learning exactly that. He brought a closer's map to a computer-vision company, and it turned a young startup into a vendor for more than twenty of the world's largest brands.

By 2020, roughly 70% of SandStar's clients were international. The company that most of its neighbors had never heard of was, by design, better known in boardrooms abroad than on the street outside its office.

The Arc

From quota to computer vision

2003

Enters Tsinghua University's School of Software as one of its early students.

~2007-2015

Roughly eight years in enterprise software sales at Oracle and IBM; leads IBM China's e-commerce team to top performance.

2015

Co-founds a 3D pancake printer venture with Tsinghua classmates after a chance idea at a reunion.

2016

Founds YI Tunnel, bringing computer vision to physical retail.

2018

Demos smart kiosks in Atlanta; wins a Coca-Cola contract by selling top-down.

2019

Raises Series B (~$25M total) and rebrands YI Tunnel as SandStar for global growth.

2020

Operates across ~12 countries, 70% of clients international, dual HQ in Beijing and Charlotte.

Good To Know

Five things

1

His first startup printed faces and cartoons onto pancakes with a 3D printer.

2

A SandStar unmanned store reportedly costs about 5% of an Amazon Go setup.

3

The name "SandStar" is a mission in a word: from sand and dust, reaching for the stars.

4

He landed flagship deals by flying to customers' HQs instead of starting at home.

5

He runs one company on two continents - Beijing and Charlotte, North Carolina.

The Aspiration

Make the technology boring, then everywhere

Wu's ambition is not to build the most futuristic store on earth. It is to make checkout-free, self-analyzing retail cheap and ordinary enough that a small shop can afford it. That is a harder goal than spectacle, and a more consequential one. Spectacle gets headlines; affordability gets adoption.

The pieces are already in motion. Four product lines that meet retailers where they are - a vending machine, a checkout machine, an unmanned store, a shelf-watching analytics layer - so a business can start small and grow into the full system. A cost structure built to undercut the marquee names. A client list that reads like a global brand directory.

The rebrand from YI Tunnel to SandStar was, in a sense, the thesis stated out loud. A tunnel is a passage from one place to another. A star is a destination you aim at knowing you may never touch it. Wu chose the star. The pancakes were the warm-up. The stores are the point. And the shelf, for the first time in a hundred years, is learning to look back.

Follow The Trail

Where to find more

Quick facts: Yili Wu

Yili Wu is the founder and CEO of SandStar (formerly YI Tunnel), a computer-vision company that turns ordinary shops, shelves and vending machines into checkout-free, self-analyzing retail. A first-generation graduate of Tsinghua University's School of Software, he spent roughly eight years selling enterprise software at Oracle and IBM before betting that brick-and-mortar retail could be digitized the way e-commerce already had been. He founded the company in 2016, sold his first big contracts top-down to Fortune 500 headquarters in the US and Europe, and grew SandStar into a business operating across a dozen countries with dual bases in Beijing and Charlotte, North Carolina.

Role
Founder & CEO at SandStar
Organizations
SandStar, YI Tunnel, Shi Hao Yun Jianbing, IBM, Oracle
Nationality
Chinese
Education
Bachelor's in Computer Software, Tsinghua University, School of Software
Known for
Founded and scaled SandStar (formerly YI Tunnel) into a global computer-vision retail company, Built product lines spanning AI vending machines, AI checkout machines, pure computer-vision unmanned stores and smart store analytics, Computer vision reported to recognize 30,000+ SKUs at 99.7% accuracy

Last updated: