Dark Kitchens 2.0 — Humans vs Robots

The first deal I ever championed at firstminute was Mission Food, now called Taster, a dark kitchen start-up founded by Anton Soulier, former Head of Deliveroo France. We ended up not doing the deal because, at the time, we felt it was more a real estate play than a tech play. Two years later, Anton has proved us wrong: he has raised $13m to date from Tier 1 VCs and has delivered more than 400,000 meals across 11 kitchens by using best-in-class technology.

Today, I’m still bullish on dark kitchens, and I’m particularly excited about automated dark kitchens.

Why Dark Kitchens?

The delivery market is growing rapidly. In 2018 report “Is the Kitchen Dead?“ UBS projected that the $35 billion meal delivery economy would grow to $365 billion by 2030. However, traditional brick-and-mortar restaurants are not well-equipped to handle delivery orders partly due to 1) retail restaurants being too expensive to build and scale quickly around a city; 2) restaurants margins being too thin to handle the platform fee from the delivery apps; and 3) the delivery experience conflicting with the in-store experience.

This new paradigm has opened up opportunities for new solutions: delivery-only kitchens, known as “cloud”, “ghost” or “dark kitchens” that use a combination of advanced and cheaper food preparation, lower labour costs, underused and cheaper real estate (because space is dedicated to cooking-only and is located away from the high-street), and AI-driven optimisation to increase output and decrease overheads.

A competitive landscape

In the past two years, dark kitchens across Asia, the US and Europe have seen various permutations. Food delivery services, such as Deliveroo Editions (one of the pioneers of the dark kitchen space), UberEats and India-based Swiggy, leverage the abundance of data they have to identify areas of unmet demand to then operate their own network of cloud kitchens in those areas, usually in under-utilised spaces, such as warehouses and shipping containers, on the behalf of existing restaurant brands. Similarly, Softbank-backed ParkJockey (rebranded as Reef Technology) is making a bet on under-utilised car parks to host mobile kitchens.

Some players don’t operate restaurants at all and instead focus on real estate and rent out equipped cooking spaces, such as City Storage that operates Uber co-founder Travis Kalanick’s CloudKitchens in the US, which rents out to QSRs (quick-service restaurants) such as Sweetgreen, and Karma Kitchen in the UK. Similarly, Kitchen United, a Google-backed business, which plans to open hundreds of delivery kitchens across the US, charges a monthly membership fee that includes the premises, back-of-house services such as dishwashing and access to its technology system for processing online orders from a range of delivery apps.

Other start-ups offer different combinations of facilities and services. London-based Kitopi is a ‘Kitchen-as-a-Service’ (KaaS) platform that enables restaurants to open delivery-only locations by providing the necessary managed infrastructure and software with minimal capex and time. As one of Kalanick’s largest tenants, Kitopi operates 80 restaurants across 35 kitchens.

While “virtual restaurants”, such as Taster and Keatz, create their own virtual brands and work closely with delivery platforms (ie. Deliveroo, UberEats and Glovo) to identify areas of unmet demand, whether in location or cuisine, and then rent out space from an outfit such as CloudKitchens or Karma.

One of Keatz’s virtual brands

Dark Kitchens 2.0 — automated dark kitchens

However, in the past six months, many dark kitchen companies have had to scale back as growth across Europe and the US has been limited. The main problem faced by operators is that, to ensure rapid delivery, operators have to be local to their customer base, resulting in small catchment areas, meaning low order volumes (on average 50 orders per day per brand).

Low order volumes combined to high labor costs, and an increasingly short supply of low-wage labor with high churn due to the gig economy, have made it challenging for traditional dark kitchens to scale and make unit economics work.

I think automation and robotics can help resolve this problem. While automation in the restaurant industry is still very much in its infancy, the landscape seems to be moving rapidly, with several prototypes already in operation. London-based Karakuri have built their DK-One robot that’s meshing machine learning, optics, and sensors to help QSRs automate the process of dispensing ingredients and measuring precise quantities. firstminute invested in the brilliant Barney Wragg and Simon Watt alongside Founders Factory and Ocado, who led the round and made their first investment into a technology start-up.

In the dark kitchen space, SF-based Halo Kitchens is one of the first players looking to use robotics to automatise food preparation. By operating their own network of cloud kitchens with an automation machine, Halo Kitchens hopes to drive economies of scale with optimised unit economics through lower labor costs (as they will need only one prep chef instead of four for traditional dark kitchens), more efficient use of kitchen space, less food waste, and more standardised quality across menu items.

The intersection of meal delivery services, AI, data, and robotics automation is creating the new wave of disruption in food preparation. With their ability to optimise and automate, robotic virtual kitchens may ultimately be the winners of the food war.