Close Ad

The Spoon

Daily news and analysis about the food tech revolution

Topics

Forget Sidewalk Robots or Drones. In the Future, Food Could Travel to Your Home in Underground Pipes

Delivery & Commerce

Why use a drone or sidewalk delivery robot to deliver packages when you can have them sent directly to your kitchen via a series of tubes?

No, I’m not referring to Ted Stevens’ imagining of the Internet or a plotline from a Steampunk novel, but one startup’s vision of an underground delivery network that would send packages hurling towards their end destination at speeds of up to 75 miles per hour.

That startup is Pipedream Labs, which has a plan to build an underground pipe network for near-instant delivery of physical goods. The idea, which is one of those that is so crazy you can’t figure out if it’s brilliant or stupid, works like this:

The Pipedream delivery system would be a citywide underground delivery network that utilizes pipes and electric-powered delivery pods to shuttle things around at high speeds. It’s essentially a Hyperloop for delivery, only instead of transporting people, it will bring you the latest Amazon package or hamburger from your favorite restaurant.

While the initial plan is to create a “middle mile” network for long-haul delivery across cities, the company’s CTO says they have a vision for eventually delivering products directly into consumers’ homes. He envisions a new kind of home appliance called the Home Portal which would enable “cheap, fast, and environmentally friendly delivery of groceries, food, and packages.”

The delivery infrastructure will be PVC piping, the same kind used by city utilities for plumbing or electrical systems. In fact, the company says they plan on making all infrastructure usable by utilities “if needed” or “in the event that PipeDream migrates to an alternative delivery method (Star Trek Transporter?) or ceases operations”.

Packages would be delivered “intra-district” to different parts of the city and would go to what the founder describes as delivery nodes.

The nodes will utilize delivery portals, vending machine like kiosks that would hand off goods to a customer or to a last-mile delivery person or robot. Portals hand off packages through a hatch and can cache up to 8 delivery pods at a time, allowing it – for a limited amount of time at least – to act like an Amazon storage locker.

Source: Pipedream Labs

Delivery pods are 10.8″ in diameter by 18″ in length and have a theoretical speed of over 110 miles per hour (but will likely move around at a speed of 60 to 75 miles per hour when in operation). They have two sections, a drive section (which includes the motor, electronics, and battery) and a removable cargo carrier section.

Pipedream envisions all sorts of products delivered via their pods, including food. According to the company, the internal capacity has room to carry 95% of grocery items and most any type of prepared food from a restaurant (except for pizza, which the company says they are working on).

The analyst in me looks at an idea like this and says there’s no way it would work. The cost of building out the network, the difficulty of navigating city bureaucracies to get a network deployed, not to mention the many technical challenges of creating an underground system and operating it all seem insurmountably difficult.

But as I think about a world where ever-more products are delivered to our homes, it doesn’t take long to realize we’ll need a variety of creative solutions beyond the status quo. Car delivery doesn’t make sense long-term for small packages, but we also don’t want to live in a dystopia with drone darkened skies or sidewalk robots congesting our walkways. Taking a portion of package delivery underground may make the most sense long term.

Of course, it will take a while before we ever know if Pipedream’s, um, dream comes true. The company has only raised $1.6 million in seed funding so far and would need to tap into utility loan funding to build a network of the size they envision.

But who knows? Maybe Elon Musk will embrace underground delivery the same way he’s helped push underground transportation forward and invest in the company, or a forward-looking city will work with Pipedream to fund an underground delivery network for stuff over the next decade.

Either way, an operating underground delivery network is an interesting new idea and one that might have a future in an increasingly e-commerce-driven world.

Else Labs Announces Pro Kitchen Focused Oliver Fleet As It Pauses Rollout of Home Cooking Robot

Connected Kitchen, Next-Gen Cooking, Robotics, AI & Data

Else Labs, the company behind the countertop home cooking robot called Oliver, announced today the launch of Oliver Fleet, a commercial kitchen reimagining of its original core product.

The new Fleet solution is a respin of its original standalone Oliver home cooking robot into a solution that allows multiple units to be used and managed simultaneously in professional kitchen environments to automate cooking tasks. According to company CEO Khalid Aboujassoum, while the Oliver Fleet units look the same from the outside as the original consumer unit, they’ve been built to withstand the more rugged requirements of the professional kitchen.

“It might look like the household unit from the outside, but the guts of the Oliver Fleet are different,” Aboujassoum said. “The Fleet units are designed for back-to-back cooking, for that harsh environment in the commercial kitchen compared to the household.”

With the pivot to a food service focused solution, Else is pausing the rollout of the home Oliver. According to Aboujassoum, the decision to make the change was largely driven by the supply chain disruptions and associated component shortages and price changes. While some backers of the Indiegogo campaign eager to get their home Oliver may not be happy with the switch, Aboujassoum said the company would give them the option of a full refund, or they can choose to continue to wait until the company restarts the consumer unit rollout.

While the focus on commercial automated cooking comes after a pandemic where restaurant businesses have faced increasing challenges around labor, Aboujassoum told me the company started hearing interest in developing a commercial version of the Oliver before COVID.

“It was an initial modest conversation at an exhibition late in 2019 where the Oliver got the attention of one of the food service companies,” Aboujassoum said. “The composition of the Fleet was born out of these conversations.”

The pandemic put everything on hold, but eventually, Else Labs started to hear more requests as things began to normalize. “As the dust settled, those conversations revived again,” Aboujassoum said. “We started receiving an influx of inbound requests all the way to the CES participation (earlier this year).”

The way Aboujassoum sees it, the Oliver Fleet can help food service companies move away from centralized food production in a central kitchen by pushing the ability to cook from raw ingredients on-site using automation.

“When I talk to (food service) clients, they’ve set up operations where they may have a huge central kitchen with a production plant, and they are shipping to maybe 50 locations,” Aboujassoum said. “We are talking about decentralizing the central kitchen. How much money can you save by deploying the Oliver Fleet and decentralizing the central kitchen? It’s a very transformational proposal.”

Aboujassoum says the Oliver Fleet system is available now and they will have announcements of deployment partners very soon. You can see a video of the Oliver Fleet system in action below.

PizzaHQ Opens to Public With Plans to Deliver 1,500 Robot-Powered Pizzas Per Day

Robotics, AI & Data

The robotic pizza chain of the future envisioned by Darryl Dueltgen and Jason Udrija took a big step forward this week as its first location opened to the public.

The company, which The Spoon first wrote about last year, envisions a modern take on the pizza chain by building a network of robot-powered pizza restaurants tailored for delivery. Its founders started working with Picnic last year to optimize the Seattle startup’s pizza robot to work with their new restaurant concept. Earlier this year, they started delivering pizzas to corporate and education customers and, as of this week, started making pizzas for the public.

When we first talked to PizzaHQ’s founder Jason Udrija, he told us the idea was to build a pizza chain optimized around robotics utilizing a hub and spoke production model. They planned to build a centralized hub to create the raw ingredients and fulfill the orders via distributed fulfillment centers outfitted with Picnic’s pizza-making robots. The company opened its first location in Totowa, New Jersey (in a building once occupied by another pizza restaurant) and has plans to build its centralized production hub in the same city in 2023.

For now, customers can order through their website, an app, and third-party delivery (there is also a pickup area at the front). In the coming weeks, PizzaHQ plans to ramp up the pizza production from about 500 pies per day to 1,500 per day. The company anticipates its next location – with an additional Picnic robot on board – will be able to produce between four and five thousand pizzas per day.

Pizzametry, Pioneering Maker of Pizza-Making Robotic Kiosks, Is Looking for a Buyer

Robotics, AI & Data

Pizzametry, the maker of the industry’s first pizza-making robot, is looking for a buyer.

In an interview with The Spoon, Pizzametry President Jim Benjamin said that the company, which has been working on its pizza robot for close to two decades, has continued operations for the last few years but has reached the point where they think another owner should take the reins to bring the product to market.

“We haven’t shut down, but we’re in a situation where we’re really looking for someone to take over and bring this to market,” Benjamin said.

According to Benjamin, the company made five Pizzametry units, of which two are currently in operation at an ice arena in upstate New York. The units make each pizza entirely from scratch, slicing and cooking the dough, adding sauce and cheese and toppings, and can go from order to boxed pizza in approximately three minutes. Each unit requires electricity and Internet to operate (but no running water) and has a large video screen for advertising (you can watch a Pizzametry making a pizza here).

The company, which has accumulated several patents around pizza automation, is looking for an interested company or individual(s) who would be open to buying their IP, which includes a license to the patents and the proprietary operating and process know-how, as well as the operating units. According to Benjamin, they would help the company design new machines, including a smaller-footprint machine which he believes is necessary to open up additional operating locations and achieve lower overall hardware costs.

The current machines “are the high volume machines that demonstrate the functionality,” explained Benjamin. “But the sweet spot is, instead of a machine with a 150 pizzas capacity, is a machine more like 50 Pizza capacity per day. Something smaller footprint, able to fit in a convenience store or gas station.”

To develop its pizza machines, the company worked closely with design services and automation service firms in Calvary Robotics and D&K Engineering. The company worked with these firms to understand how to build scaled-up and scaled-down versions of the robot, but at this point, it is looking for a new company to invest in building a smaller-footprint, lower-cost machine.

I had a chance to try a pizza made by a Pizzametry robot when the team flew one up to Seattle for the Smart Kitchen Summit in 2018. The pizza was good, but I can see why they feel they need to build a new version with a smaller footprint. The current unit, which has a refrigerator inside to store the ingredients, takes up about 15 square feet, too big to fit in a typical convenience store on the floor of an airport terminal.

Benjamin agrees and believes they could work with the new owner to build a smaller machine.

“The principles of operation that we would transfer to a buyer would stay the same,” Benjamin said. “The patents that we currently have would be in place, but it would just be a smaller footprint.”

Benjamin explained that they could help with everything from the proper sauce viscosity, the dough formula, and pretty much everything else required to run a pizza robot would be involved in what he described as a “technology transfer” process.

While back-of-house pizza robot startups like Picnic and Hyper Robotic are getting traction, some building robotic pizza kiosks have found the road a little rougher. The news of Pizzametry’s interest in finding a buyer comes just a couple of months after the news of Basil Street selling off its assets. For its part, Piestro, one of the other remaining stand-alone pizza kiosk startups, continues to raise capital and partner with others as they work to bring their product to market.

If you are interested in inquiring about the Pizzametry business, you can contact the company via their website.

People Using The Wonder Food Delivery Service Are Bonkers About It

Delivery & Commerce

You’ve probably heard about Wonder by now. The high-profile company founded by Marc Lore has people buzzing with its nearly billion in funding and a model that includes chef-designed meals and a network of vans that cook up the food curbside.

Last month I wrote that Wonder has the opportunity to either reinvent the food delivery business or become a case study like others who have tried and failed to build out fully integrated delivery models.

What I didn’t write about is how much Wonder’s customers seem to love the product. After hearing from people I’ve talked to and reviews I’ve read about it online, it seems the service has an almost fanatical user base.

One person I spoke to told me his family uses Wonder multiple times a week, ordering meals that can range well above a hundred dollars with alcohol included to more affordable middle-of-the-week family meals.

Another wrote via Linkedin, “I live in that one metro market that Wonder delivers to, and it is wildly popular in my town. Last week I couldn’t decide between two Wonder “restaurants” Di Fara from Brooklyn or Mozza from LA. Both pizza spots impossible to get into but not when the Wonder truck pull up to your house and cooks it for you on the spot!”

The responses via Apple’s app store are even more gushing:

“The food was wonderful, and I cannot wait to order again. What a fabulous concept. It really works. Telling everyone that I know about the fabulous Wonder trucks.“

“Impressive how quickly you can be eating a fine restaurant quality meal from the comfort of your own home. Hoping Wonder continues to thrive, and that they keep the bar raised high for their commitment in delivering fine ingredient quality meals.“

Here are a few observations about why Wonder seems to be developing an infatuated following:

Users love the quality of the food. In review after review, people say the food is really good. Almost everyone says the food is as good or better than they could get in a restaurant.

The service is white glove. While some users have said Wonder has a kink or two to work out, it doesn’t seem to matter because the customer service is so strong. When Wonder makes a mistake, someone is there to make it right. People also love the chefs who show up at their door in a white coat and Wonder hat to deliver their meals.

People love access to chef-designed meal concepts. Getting quick access to food that could have been made in far-away and popular restaurants like Di Fara Pizza or J Bird from Jonathan Waxman is something that resonates with reviewers.

Word of mouth and omnipresent delivery vans are reinforcing success. Everyone in the New Jersey metro area Wonder serves seems to know about the service. Word of mouth is extremely strong and people are seeing the Wonder vans buzzing up and down the street.

While I can’t verify how many of the app’s reviews are from Wonder employees or Wonder-friendly people, there are too many positive ones (over three thousand at this point) for the early buzz to be contrived. The strong reviews also seem to reinforce what anectodally appears to be a higher than industry average frequency of usage by Wonder customers. I also imagine strong customer metrics are one of the big reasons the company has continued to raise money in an environment much tougher than it was just 6 months ago.

If you’re lucky enough to live in Wonder’s delivery area, drop us a line and let us know your thoughts. I am still unsure how the model scales nationwide, but there’s no doubt that early results show Marc Lore and his team may be creating something special with Wonder.

From Robotics to Alt Protein, Layoffs and Shutdowns Have Begun to Hit Food Tech

Foodtech

What a difference a year can make.

Last year, it seemed a day didn’t go by without word of a big new funding round in food tech. Alt proteins, restaurant tech, food robotics, food waste, and other sectors benefited from the combination of low-cost capital and a newfound urgency among investors to reinvent our antiquated food system post-pandemic.

But as the economic outlook darkens and investor eagerness gives way to caution, some startups and publicly traded companies in food tech find themselves in situations where they need to pare back their ambitions, tighten their belts to extend their capital runways and, in some cases, shut down altogether.

Here are some examples of bad news from the last couple of weeks:

Ghost Kitchens & Restaurant Tech: NextBite, one of the high-flyers in the virtual restaurant and ghost kitchen space, has laid off employees for the second time this year. The news comes a couple of months after Reef laid off 5% of its workforce. Restaurant payments app startup Sunday is shutting down operations in a number of its markets and laying off a large percentage of its team.

Food-Making Robots: Doordash, the publicly traded food delivery startup, is shutting down Chowbotics, the food robotics startup it acquired last year. The shutdown of the salad-making robot kiosk comes just a couple of months after we learned that Basil Street Robotics, a maker of pizza-making robotic kiosks, had put its assets up for sale.

Delivery Bots: Speaking of robots, the original sidewalk delivery robot startup Starship (which we started covering back in 2016), has recently laid off 11% of its workforce.

Alt-Protein: Last week, we learned that Motif Foodworks, the well-funded alt-protein ingredient spinout of Ginkgo Bioworks, has laid off an undisclosed number of employees.

Fast-Grocery: The ultra-fast grocery business has fallen further from its 2020-2021 highs than perhaps any other sector, as investors realize how much more capital is needed to build out their hyper-local dark store and delivery networks. Over the past few months, we’ve seen startups Buyk and Fridge No More shut down and others, like JOKR and Gorillas, pare back and exit some markets to preserve capital.

While some sectors (like alt protein) seem less impacted than others, the overall food tech market is recalibrating to a new normal after a year where post-money valuations were too high, and capital was still relatively cheap. In 2022, the rising cost of capital and pessimism about the economy has resulted in investors becoming more cautious, meaning new investments and follow-on rounds are harder to come by.

It also means investors are putting much more pressure on their existing portfolio companies to cut costs and get to profitability quickly. That can mean layoffs, bringing in professional manager types (as with NextBite), or exiting from new markets. For publicly traded companies like Doordash, which face the quarterly pressure of reporting to Wall Street, it means experiments that are not a part of their core business (like automated food kiosks) get dropped.

The question facing the food tech industry (and, let’s face it, all tech-adjacent industries) is whether things get worse before they get better. My guess is the answer is yes, as some startups that benefited from the abundance of capital over the last few years will have a harder time raising their next round and will have to work hard to lengthen their runways. In some cases, those runways will run out.

Despite all of this, I remain optimistic about the future of food tech. There are too many inefficiencies and problems with the current food system, and, because of this, innovative companies will always find opportunities to reinvent an industry. And much like we saw with the first dot-com downturn, those companies that find ways to survive in moments of austerity emerge stronger and built for long-term survival.

But before we get to the other side, we would all be wise to prepare ourselves for a rocky year or two.

Will Wonder Reinvent the Food Delivery Biz, or Become Another Cautionary Tale? Only Time Will Tell

Delivery & Commerce

Imagine you wanted to build a complete-from-scratch meal delivery company.

Not just the delivery part like Doordash. I’m talking about building a company that is essentially an entire restaurant and food delivery industry in a box, one that works with big-name chefs to develop new restaurant concepts, builds centralized food production facilities, creates a network of mini-kitchen hubs around a large metro area, and owns the delivery network to get the food to people’s doors.

In other words, everything. If that sounds like a big vision, it is, and it’s exactly what Marc Lore is building with Wonder. The founder of Jet.com and Diapers.com described the concept in a Linkedin post last December:

Our innovative, vertically-integrated approach begins with exclusive menus from the country’s best chefs and restaurants. A central commissary sources high-quality, fresh ingredients and serves as the start of each meal’s journey. Orders are then fired, finished, and plated in our mobile kitchens just steps away from your door, and served as soon as they’re ready — allowing you to experience the food the way it’s meant to be enjoyed. 

Lore is no stranger to big industry-shifting ideas. He created Diapers.com, one of the early pioneers in online baby products (acquired by Amazon) and Jet.com, a discount-pricing based online retailer acquired by Walmart in 2016. He also has plans to build a Utopian city in the American west.

Wonder’s business model is essentially built around a three-layer logistics network, one where the company owns centralized production, a distributed network of mini-kitchens in various neighborhoods, and the final delivery network that drops the food off at the consumer’s doorstep.

If it sounds reminiscent of Zume, it’s because the two ideas are kinda similar. Zume created a vertically integrated food delivery business complete with a robot-powered pizza-making ghost kitchen, mobile food trucks with ovens built-in and a fleet of scooters to deliver pizza to the customer. Zume also raised a bunch of capital – $423 million – before the company ended up laying off 80% of its employees and exiting the food delivery business to focus on sustainable packaging.

Despite the similarity between the two concepts, Zume’s troubles haven’t dissuaded investors from investing in Wonder. Lore has raised $900 for Wonder so far, including a $350 million Series B announced this month.

In fact, if one thing is clear, it’s that Lore is good at raising money. He raised over $800 million for Jet before he sold it to Walmart, and he’s also started an investment company with Alex Rodriguez with plans to raise $500 million. He also is raising $25 billion for the first phase of construction of utopian city called Telosa, with eventual plans to raise $400 billion.

But even for someone as talented as raising money as Lore, you still have to wonder how long investors will stay patient as his company builds out its food delivery business and scales it to other cities. After raising close to a billion dollars, the company is serving one metro market so far, a cluster of neighborhoods in the New Jersey area. While it’s clear that standing up its first metro market will probably be more capital intensive than its second and third market – the restaurant concepts, recipe development and core technology development built now can be leveraged for each new market expansion – the company will still need to build out a three-layer delivery network for each new metro is expands into.

Who knows, maybe Lore and Wonder can generate enough cash flow with its New Jersey business and can cost-control its burn rate to extend the $900 million out for a couple of years to fund another market build-out or two. Still, no matter how frugal the company remains, it’s going to have to go back to investors at some point, and with things getting tougher in a global macroeconomic environment filled with increasing uncertainty, nothing – including the future availability of multi-hundred million funding rounds – is a certainty at this point.

And it’s not just the economy, but an extremely competitive, fast-changing restaurant and food delivery business. I’d argue the restaurant and food delivery business is even more competitive and market-saturated than baby products or discount e-commerce offerings. Wonder is competing not only with well-capitalized competitors in Amazon, UberEats, and DoorDash, but also facing off against a local mom-and-pop restaurant on every street corner that is increasingly relying on digital business models to survive as we emerge from the pandemic.

All that considered, Lore has an amazing track record that would be foolish to ignore. He’s achieved successful exits for his previous companies, selling both Jet and Diapers.com to big entrenched players (though both companies have been subsequently shut down by their new owners). So maybe investors are looking at those two previous exits and are comfortable Lore can pull yet another rabbit out of a hat.

For his sake and theirs, let’s hope he can do it once again, because if he can’t, I suspect we’ll be studying the case of Wonder years from now as another cautionary tale of audacious visions and spent venture capital.

PARC Spinout EverCase Uses Electric & Magnetic Fields to Store Food in Freezers Without Ice Crystals

Delivery & Commerce

If you’ve ever put meat or fish into a freezer, you’ve probably noticed it doesn’t look nearly as fresh once you thaw it out.

That’s because the process of freezing food alters and damages its structure at a cellular level. As the temperature drops, water molecules slow down, and ice crystal embryos form ice nucleation sites. From there, the ice spreads to freeze the entire piece of food. Water within the food expands by up to 9% when frozen, causing food cells to rupture. When frozen food thaws, nutrients and flavors leach out from the food, often in the form of drip loss (that red liquid dropping from a warmed piece of red meat).

But what if you could store and preserve food in a freezer at sub-zero temperatures and avoid the damage incurred by traditional freezing? That’s the idea behind a new startup called EverCase, a spinout from storied research and business incubator Xerox PARC.

The new company, announced on June 15th, is the result of almost a decade of research that started when Dr. Soojin Jun, a professor at the University of Hawai’i at Mānoa, got a three-year research grant from the USDA in 2013 to research the technology dubbed “Supercooling.”

Jun’s Supercooling technology utilizes pulsed electric and oscillating magnetic fields to cause water molecules within food stored at sub-zero temperature to vibrate, inhibiting the formation of ice crystals. The result is food that, when pulled out of a Supercool equipped freezer, has almost the exact look and texture of food that is fresh and not riddled with ice crystals.

Image Above – Left: Drip Loss from thawed traditionally frozen meat. Right: EverCase’s comparison of meat using different preservation techniques

Jun would eventually take his ideas to Xerox PARC where he would get help incubating them and preparing them for commercialization. The end result of that move is EverCase, a new spinout that plans to build systems with Supercooling that can be used in existing freezers.

You can watch a demo in the video below where EverCase shows a piece of frozen meat compared with a piece of meat stored using Supercooling technology.

EverCase Demo Video

The company is pitching nothing less than a revamp of the traditional “cold chain” supply network, where freezers, from packers to the retail storefront (and possibly even to the home), use Supercooling technology. The company’s pitch deck talks of a new category of smart packaging and a new preserved food category of ‘Supercooled foods.’ They also plan to work with refrigeration manufacturers and other OEMs to build Supercooling technology into freezers and refrigerators.

It’s an ambitious plan, but there’s no doubt there could be a market for technology that helps food sidestep some of the downsides of traditional cold chain freezer technology. The company does say its technology is “inexpensive to make,” but it’s still unclear to me what the total cost of upgrades for a restaurant, retailer, or food packer would be. Nevertheless, it is encouraging that EverCase says its technology works with existing freezers because forklift upgrades for the massive amount of installed freezer systems throughout a mature cold-chain network is a non-starter.

The company is headed up by Chris Somogyi, a former co-founder of cell-cultured seafood startup BlueNalu and business development exec at PARC, and other executives from IBM, Xinova, and PARC. According to the announcement, EverCase is in the process of a Series A funding round.

Food Robot Roundup: Grubhub & Cartken Head to College, The Story of a Stir-Frying Robot

Robotics, AI & Data

Happy Monday! We’ve got a round-up of recent food robot news to get your week kicked off right.

This month, Singapore-based food robotics firm Ross Digital announced a $4.2 million ($3 million USD) Series A+ round led by food conglomerate Fraser and Neave. According to the company, which makes robotic arms for serving coffee and cocktails, the new funding will be used for product improvement and expansion into Thailand and Malaysia.

Ross Digital sells unmanned robotic baristas that can serve different types of drinks. An accompanying digital platform called Ross Cloud powers its baristas and provides a suite of restaurant solutions such as mobile app, a point-of-sale system, and an ordering kiosk. The company already has 15 units deployed in Singapore and China and aims to put out 40 robotics arms in the markets by the end of 2022. Some of its clients include Razer (whose venture arm also joined the funding round), CNBC, and Alibaba. 

Ross Digital’s expansion is yet another sign of Singapore’s active food robotics landscape. Last August, Stellar lifestyle and Crown Digital teamed up to launch ELLA, another robotic barista that can serve up to 200 cups of coffee an hour. The two companies plan to bring ELLA to 30 Mass Rapid Transit stations by the end of 2022.

The Story of a Stir-Frying Robot

One of my favorite easy meals to cook is stir fry because I just add everything I like in the pan with oil and call myself a chef. It’d be even easier if I could have a cooking robot to do the work, and if researchers from Idiap Research Institute in Switzerland, the Chinese University of Hong Kong, and Wuhan University have anything to say about it, someday I just might.

That’s because these researchers have taught a robot how to make the stir fry motion. The three labs have been collaborating for about 10 years with a specific interest in teaching robots to prepare food for people. Junjia Liu, one of the researchers, noted that “Food preparation and cooking are two crucial activities in the household, and a robot chef that can follow arbitrary recipes and cook automatically would be practical and bring a new interactive entertainment experience.”

Researchers were able to achieve this feat by decoupling the two arms of the robot into a leader and a follower and teaching them separately through machine learning. The two arms were then combined through general bimanual coordination and movements were subsequently adjusted automatically by giving visual feedback of the contents of the pan. The robot was able to complete the motions of stir-fry, but the paper doesn’t mention whether the robot was successful with heat, which is likely the next step. 

This isn’t the first time robots have dabbled in stir fry. Spyce, the robotic restaurant acquired by Sweetgreen, also served stir fry prepared by robots. Instead of using robotic arms, Spyce stir fry system utilized rotating compartments that cook and dispense stir fry. Spyce later pivoted to a different food robot that involved placing dishes on a conveyor belt that ran underneath dispensers that portioned out warm and cold ingredients. 

Grubhub and Cartken Head to School

Grubhub announced a partnership with Cartken, a maker of self-driving AI-power robotics and delivery operator, to bring robots to college campuses around the US. The partnership will allow Grubhub to leverage Cartken’s artificial intelligence and camera-based navigation and mapping technology. The robots operate at up to 3 miles per hour on campus and can handle various weather conditions including rain and snow, perfect for the Ohio climate they tested in. 

Grubhub and Cartken piloted the robots at Ohio State University this spring and are planning a full rollout in the fall. College campuses are the most saturated area for food delivery startups to test and operate in. Kiwibot announced plans to enroll in 50 college campuses by the end of 2022 and Starship Technologies is already enrolled at over 20 campuses. 

In Grubhub, Cartken finds a partner already well-entrenched on college campuses. The food delivery company already works with more than 250 college campuses across the U.S. where it integrates directly with meal plans so that students can access on and off-campus restaurants for delivery and pickup.

Food Tech News Pod: The Wendyverse, 15-Minute Grocery Struggles & Roku’s Big Food Play

Delivery & Commerce

It’s another weekly food tech new wrap-up.

The stories we discuss this week include:

As always, you can find more episodes of The Spoon Podcast on Apple Podcasts, Spotify, or wherever you get your podcasts. You can also just click play below to listen to today’s show.

Is Roku About To Bring Us Shoppable TV Content Featuring Martha Stewart & Other Culinary Giants?

Connected Kitchen, Delivery & Commerce

Last week, Walmart and Roku announced a deal that would allow TV viewers watching streaming via a Roku device to purchase items – including food items – using their remote.

According to the announcement, the new experience will allow customers to click on and purchase items advertised within the “moments of entertainment” (translation: during an actual show and not an explicit commercial), as well as during commercial breaks during ad-supported programming.

The new integration will allow viewers to click on a shoppable ad and proceed to checkout. The customer’s payment information will be pre-populated from Roku Pay, Roku’s payments platform, and then the customer taps “OK” on the Walmart checkout page to place the order. A Walmart purchase confirmation is emailed to the customer.

By taking shoppable commerce to the TV screen, Walmart is going beyond the shoppable integrations the company has previously done through partnerships with SideChef and Tasty. While the rise of video-centric social media platforms is blurring the lines, TV watching (including streaming) typically is a much different experience than time spent in front of our computers doing activities like online shopping.

For Roku, the move builds on an impressive ad business which saw the company garner the bulk of its $734 million in Q1 non-hardware revenue via advertising sales. The core of the company’s ad business is done via its Oneview Platform, which does ad-targeting of consumers based on their viewing behavior and allows Roku to deliver ads interspersed into shows on its own channel and via streaming partners that do have ad-supported content (like, say, Hulu) that open up ad inventory to Roku.

This deal is interesting on its own, but it becomes much more intriguing considering Roku’s recent big bet on original cooking content featuring names like Martha Stewart, Emeril Lagasse, and Christopher Kimball. That news was unveiled in May at Roku’s 2022 NewFronts when the streaming company announced co-production deals with Marquee Brands and Milk Street Studios to produce seven new streaming shows exclusively for the Roku Channel.

Why are the two deals, when considered together, much more interesting than on their own? Because while Roku can theoretically insert commerce-conversion opportunities for Walmart or other partners in ads on other streaming channels, the streaming company can go much deeper with commerce integrations on its channel where it has the full rights to the content and owns all the ad inventory.

Some of those deeper integrations might include in-show shopping moments. For example, imagine watching “Martha Cooks” or “Milk Street’s My Family Recipe” (two of the new shows on Roku TV) and seeing Martha or Christopher Kimball showing off a new holiday recipe. The Roku platform would allow a call to action overlay within the show itself where viewers could click via their remote to view the food items in the recipe, add them to a cart, and a transaction to be completed there on the spot.

While the Walmart and the new show slate are technically different deals, I would be shocked if Roku wasn’t pushing the possibilities around in-show commerce when negotiating with Martha, Emeril and Kimball. Conversely, the timing of the Walmart shoppable ad partnership is giving Roku a premium grocery partner just as the streaming company is beefing up its food and cooking content.

Longer-term, I would imagine other streaming partners like Hulu or even Netflix (which is considering ad content) would be open to enabling in-show commerce on shows on their channel via Roku’s platform. Roku has over 60 million viewers using its platform, giving it one of the largest addressable audiences in the world of streaming. If they can tie all the pieces together and bring more streamer partners on board, the company could position itself as a true TV commerce power player.

The Case for 15-Minute Grocery Delivery is Questionable. So Why Did It Raise So Much Capital?

Delivery & Commerce

For about as long as I’ve been seriously watching the Internet industry, companies have been trying to make a business of home grocery delivery.

It started back in the late nineties when companies like Webvan and Homegrocer raised massive amounts of capital after convincing investors that food shopping would be largely done online in the future.

Webvan would raise almost $400 million in venture investing and another $375 million through an IPO. HomeGrocer raised $440 million in venture capital and almost $288 million going public.

None of it was enough. The two companies would eventually merge and went bankrupt less than a year later.

Of course, some online grocers survived, including some originating in the early days of the Internet. Ocado, conceived in the year 2000, continues to this day and is one of the biggest online grocers (and grocery automation technology companies).

But despite the occasional success story like Ocado, the reality is online grocery shopping is a tough business, one that seems to possibly work as part of a broader omnichannel market approach where grocers like Walmart, Kroger and now, yes, Amazon offer both in-person and online shopping experiences for the consumer. And even Ocado.com is essentially an omnichannel model, partnering in the early days with Waitrose.

Which brings us to the 15-minute grocery category, a model built around hyper-local delivery with distributed micro-fulfillment centers placed in dense urban markets like NYC, Philadelphia, and other locations. Startups in this space focus on convenience, offering a limited set of items, not unlike you might find in a convenience store like 7-Eleven (but usually with a little more fresh food sprinkled into the mix).

The market, which in some ways kicked off with GoPuff’s founding a decade ago, witnessed a whole bunch of new entrants enter the market over the past couple of years, including companies with similarly weird names like Gorillas, JOKR, Fridge No More, Weezy to name a few. These companies feasted on a downright frothy venture capital market, raising a breathtaking $4 billion last year alone:

However, with the worldwide economic climate facing significant uncertainty in the face of decades-high inflation, rising interest rates, and a war in eastern Europe, the easy money spigot has been shut off. As a result, some of these companies are either falling into the deadpool, getting scooped up by other competitors, or like JOKR and Gorillas, attempting to cut costs through layoffs and market pullouts to preserve capital runway as they try to survive what looks to be a long economic winter.

Of course, all of this begs the question: Why did all these startups get so much funding in the first place? As the early online grocers demonstrated, building out a network of stores and warehouses and a delivery infrastructure to get a basket of goods to consumers is an extremely expensive business.

Don’t believe me? This chart from a recent McKinsey report on online grocery shows just how tough the margins are for a standard online grocery business before we even consider the extra costs of accelerated delivery.

For a typical e-grocery business, COGS (cost of goods – i.e. groceries – sold) are the biggest expense, around 70% of a total order. The leftover 30% is eaten up by in-store and pick-and-pack labor, last-mile delivery expenses, and associated e-commerce fees. When it’s all said and done, a typical online grocery order has a negative 13% margin.

Of course, fast-grocery startups might offer slight markups in pricing and also make money through delivery fees (which range from $1.80 to $5 per order) and membership subscriptions, but what’s somewhat surprising in retrospect is that fast-grocery companies don’t have drastically different pricing or fee structures compared to that of traditional e-grocery prices.

Beyond the negative marginal profit of each order, the biggest expense driver for these companies and what likely ate the lion’s share of the billions of dollars in the collective capital runways is the buildout of their fulfillment centers and dark store networks. Being fast requires lots of points of presence to be within a 15-minute delivery window (or shorter, since fulfillment and delivery driver load-in takes at least a few minutes once the order comes through), which means lots of construction, equipment and technology costs.

Indeed, the venture community must have seen something here in a business – online grocery delivery – that has shown itself to be historically unprofitable. My guess is the rationalizations for writing these large checks fell in the following categories:

Customers will pay for convenience: We’re living busy lives and sometimes we just want what we want. If someone can get me a six-pack of beer, a steak, and a bag of chips to my house in 15 minutes, I’ll choose that option.

The pandemic changed the game and converted us into an e-grocery nation: In the early days of the pandemic and throughout 2020, we saw unprecedented conversion rates to online grocery as many consumers were forced to use it for the first time. Surely once they went e-grocery, customers wouldn’t return to the old way of doing things.

The siren song of the giant TAM: Food is a huge industry. I’m sure pitch-deck-making founders convinced investors they could convert a large enough percentage of food shopping customers to their business to take home a healthy percentage of the total available market (TAM) in the long run.

Long-term, technology & automation would drive costs down: I am sure many fast-grocery startup founders thought if they could just amass a large and loyal user-base, they could apply technology and automation to bring down the costs and increase margins as they moved past the large-scale infrastructure buildout of the early years.

These rationales for the fast-grocery business may make sense in a vacuum, and I am sure the impressive growth of early-growth pioneers like GoPuff helped convince many startup founders and eager investors there was some long-term gold to be found in those fast-grocery hills. But therein lies the problem: a closer look at these prospective businesses and anticipation of changing environmental factors – both in the form of the global macro-economic situation and the rise of competitors with built-in cost advantages – should have been enough to turn away some of the investors who jumped into this space.

Consider the e-grocery boom of 2020. While many of us thought that the rapid adoption of e-grocery would likely have some staying power even as the pandemic faded, it was never clear how just how much an average e-grocery shopping consumer would buy online once they had the opportunity to head down to their corner grocery store or load-up on staples at their warehouse store. From the looks of it, many consumers are returning to their local stores.

As for the promise of convenience, even if we assume sub-hour delivery time does offer some value to consumers, that value is reduced if there is a convenience store on the corner where one could just go pick up the goods instead.

And, say, a customer did occasionally use these services, was there any reason to assume they would continue to be that impatient? Amazon and Walmart often can usually deliver within an hour or two. Customers who want something quicker can always use a DoorDash or another food delivery app to get something to you quicker.

In reality, these adjacent competitors should have been the most significant reason investors stayed away from this space. These companies are all logistics-optimized, well-capitalized businesses that are eyeing the same TAM of the newer entrants. They also have legacy businesses with which they’ve built customer lists in the tens of millions in some cases.

We’ll see more consolidation of this market, and my guess is one or two of these startups have a chance to emerge on the other end as long-term survivors. With its early start and a warehouse network that’s largely built out, GoPuff looks like it could have enough of a customer base and capital in the bank to weather the storm. Gorillas, having just raised $1 billion late last year, may have sufficient runway if they can manage their burn rate through the downturn.

But no matter how this market shakes out, investors will be much more hesitant to sink capital in this market, particularly for companies with no discernible differentiation. Long-term, my guess is we might be talking about the fast-grocery boom of the early 20s for the next decade or more as a cautionary tale of a venture-capital fever investing, at least until the next boom cycle causes us to forget the lessons of the past once again.