2026
It's 2026. The food supply chain is still broken.
We have better software, connected commerce, and modern logistics tools. Yet the same failures repeat because the underlying participants still do not share a reliable operating signal.
- —Customers order groceries and get called mid-delivery: “Item out of stock, can we substitute?”
- —Restaurants promise items on their menu they can't consistently source.
- —30–50% of food in supply chains becomes waste — spoilage, expiry, overstock.
- —Store owners work 60-hour weeks managing inventory on guesswork, not data.
- —Restaurants give up 30% of revenue to platforms just to reach customers.
- —Drivers sit idle waiting for orders, earning close to minimum wage.
- —Ethnic groceries can't scale: pricing is opaque and distribution is fragmented.
- —This shouldn't still be happening.
I spent three years studying this problem on the ground. I moved to Boston and started visiting ethnic grocery stores, restaurants and delivery operations. I wanted one answer: why is this still broken in 2026? I sat with owners running seven different systems to operate one store. I watched them throw away products they ordered on a guess. I saw them turn down delivery because they couldn't guarantee what was on the shelf. The system is optimized for everyone except the people who grow, prepare and serve food.
My background
12+ years inside food supply chains. Operationally, not theoretically.
- —Worked directly with farmers — production constraints, seasonality, quality.
- —Managed manufacturing — COGS, scaling, quality control.
- —Coordinated cross-border distribution — where delays happen and why logistics fail.
- —Built products from field to customer — every break point in the system.
Here's what most people never see: when real demand data flows back through the system, everything changes. When suppliers know what will actually sell, waste drops. When stores can see customer behavior, they optimize shelf space. When farmers have visibility into demand, they grow the right quantities. When pricing is transparent, fairness emerges.
What I studied here
I didn't arrive in Boston and build. I studied the problem first.
Platforms
Analyzed DoorDash, Instacart, Uber Eats and Weee!. Their model is commission extraction. Their weakness is that they don't care about ethnic food quality. Their limit is that asset-light means they can't guarantee anything.
20+ stores
Visited 20+ ethnic grocery stores across Boston and surrounding areas and asked owners three questions: what's your biggest problem, why can't you scale, what would make this work?
Restaurants
Researched kitchen costs, labor, food waste and delivery commission economics. A restaurant can't open a second location when the platform takes 30% of revenue.
Logistics
Studied dispatch, driver pay and idle time. The inefficiency is structural, not accidental.
One product
Traced a single product — pomegranate — through five importers and three distributors, all competing to sell to the same stores.
Everything confirmed one thing: the system is fragmented, data doesn't flow, and everyone is guessing.
Why this still exists in 2026
The system optimizes for middlemen, not for food.
Platforms are commission machines
They take 30% and stop caring. If substitutions happen, they already got paid. If quality fails, the customer blames the store. Zero incentive to fix anything.
Distributors extract margin
Importer → distributor (20–30%). Distributor → store (20–30%). The farmer gets squeezed, the store runs thin, the customer overpays.
Stores operate alone
No visibility into real demand, no coordination with suppliers, no volume leverage. Managing by feeling, not data — so waste, stockouts and lost revenue.
The opportunity
A $50B+ market with no infrastructure built for it.
$50B+
US ethnic grocery market
$20B+
Sales still happening offline
0
Platforms built for ethnic food
- —Instacart treats ethnic like general grocery — high substitution, low margin.
- —DoorDash treats it like restaurants — 30% commission kills the profit.
- —Weee! focuses on Asian and leaves Middle Eastern, African and Caribbean gaps.
- —Amazon Fresh tried and failed.
Why now: connected commerce and inventory tools are finally practical for independent operators. Once that operating record is reliable, forecasting can become useful. The missing piece is coordination across the whole network.
Unit economics
The model we are building toward.
Illustrative model. No store has yet operated on Kolshee — these are the economics we are building to prove.
Per store, per month
- Average delivery order
- $25
- KOLSHEE commission (15%)
- $3.75
- Payment processing (2%)
- $0.50
- Revenue per order
- $4.25
- Monthly orders per store (40/day)
- 1,200
- Gross profit per store / month
- $4,800
At scale (65% net margin)
- 100 stores — monthly revenue
- $510K
- 100 stores — monthly net profit
- $330K
- 500 stores — annual revenue
- $30.6M
- 500 stores — annual net profit
- $19.8M
Why I'm building this
Not to disrupt food. Because I'm tired of watching good people lose.
- —Farmers producing and guessing.
- —Distributors warehousing and hoping nothing spoils.
- —Store owners working impossible hours managing chaos.
- —Restaurants giving up 30% to platforms that don't care.
- —Drivers waiting around for minimum wage.
- —Customers ordering and receiving substitutes.
- —Ethnic food unable to scale despite enormous demand.
In 2026 we have the technology to coordinate all of this. We just haven't built it. I'm building it.
