A robotic future for retail grocery, finally. Why we invested in Finally
Gil Dibner & David Peterson

We’re thrilled to announce our investment in Finally, an Israeli company building an end-to-end platform to manage store-level robotic fulfillment centers for grocery e-commerce.

Anatomy of an unlikely venture investment

We first met the Finally team over a year ago, but under a different name. They contacted us as they were considering expanding their nearly fully automated grocery store (called SuperDuper…great name for a grocery store, right?) from the suburbs of Tel Aviv to the suburbs of Boston and beyond. Their business plan (which they emailed us in advance) called for raising a massive round and rapidly scaling out a consumer-facing retail operation across New England. This plan — obviously — fell far outside of our mandate as a tech-oriented B2B fund. On the other hand, the team was so impressive and their insights on the grocery retail space were so deep that we took several calls. As we dug deeper and deeper into the business, it became apparent that the founders were truly extraordinary. Ultimately, however, we couldn’t convince ourselves that the plan to build a consumer-facing grocery brand made enough sense or could fall within our mandate. We parted ways and planned to keep in touch.

Six months later, they reached out to us again. “We took your feedback to heart,” they said and proceeded to outline how their thinking had evolved since we last met and what they had done about it. What we encountered was the same incredible people with a completely different — and far better plan. SuperDuper now had pivoted to selling their robotic fulfillment software platform to other retailers directly and, after just a few months, already had some of the largest players in the world knocking down their door. They provided us with a list of more than half a dozen top grocery brands across the US, UK, and Europe — all of which were in advanced discussions with them. On top of that, they had identified and recruited as advisors two of the leading retail grocery and robotics operators in the US — offering them significant equity in the business to help them navigate the new market. We asked about their existing grocery store. “We’ve already found a buyer and have agreed to sell it.” Their pivot was not just in their slides. They were already deeply engaged in executing their new vertical enterprise software strategy.

That got our attention.

So over the next few weeks, we dug in. We spent a lot more time with the team, spoke with grocery experts, operators, and prospective customers, and slowly became convinced that (1) making e-commerce work for grocery retailers is a multi-billion dollar problem, (2) it’s unsolved for a whole variety of reasons, and (3) SuperDuper may have finally figured out how to do it.

Post investment, we’ve worked with the team on their early GTM efforts, connected them with relevant sales advice, and encouraged them to undertake a rebranding effort, which led to the new name (Finally) and website. They have, indeed, sold their existing robotic supermarket business to a local chain, but you can see a video of it in operation (fulfilling real orders) on their website. They also kept their original office space, a platform perched right over the robots which whiz around on the floor below as the Finally team codes.

Our investment thesis.

Finally’s opportunity emerges from two dynamics dominating the grocery industry:

1. E-commerce is taking over grocery (a trend only accelerated by the COVID-19 pandemic)

2. Grocery stores can’t do e-commerce profitably

Grocery is a massive industry ($700B+ per year in the US alone). Every year, e-commerce eats a larger share of grocery purchase volume. And every year, grocery stores continue to lose money on those orders. It’s a slow-moving disaster, and grocery players have invested billions of dollars trying to solve this problem to no avail.

A history of failure. Generally, retailers have taken two approaches to solving the e-commerce problem, neither of which works particularly well.

• On one end of the spectrum, there’s the “centralized” approach, best exemplified by Ocado. Ocado is an online grocery with just seven fulfillment centers to serve the entirety of the UK’s 67 million people. These fulfillment centers are wonders of robotics technology. The problem is that these fulfillment centers require massive upfront investment ($100M+ per facility) and are in the middle of nowhere (how do you deliver orders to the far-reaches of the UK profitably?). It’s a model that can work, but the scale required is huge. And it doesn’t make much sense for retailers that have an existing real estate portfolio. In addition, Ocado-style facilities are rigidly designed and centralized, meaning that any breakdown in the facility shuts down the entire operation. Consequently, Ocado has yet to make a profit and most observers suspect that it never will.

• On the other end of the spectrum, there’s the “hands-off” approach, typified by a company like Instacart. In this model, the retail chains aren’t involved in capturing the value or driving efficiency. Instacart shoppers are deployed to shop (referred to as “picking”) in stores manually. But this isn’t ideal. One of the main issues is stockouts. Instacart doesn’t know what’s on the shelves, so customers, inevitably, order things that are out of stock. This leads to orders with missing items or imperfect replacements, both of which reflect poorly on the retailer, even though there’s nothing they could have done about it. In addition, the presence of increasing numbers of “Instacart pickers” within the retail environment degrades the shopping experience for traditional customers, which ends up driving more people online, perpetuating the cycle and driving more pain for retailers. In the Instacart model, grocery stores are effectively providing a free distribution platform that another company is leveraging — something that is clearly unsustainable.

Enter Finally. Finally has figured out a third path — a robotic “store fulfillment center” or SFC — which has been designed specifically to serve the needs of existing grocery stores entering the omnichannel era.

Finally’s disruptive approach is based on a few key insights:

• First, if the fulfillment center is close enough to the end customer, last mile delivery is a solved problem. Thankfully, a grocery retailer’s existing real estate portfolio has been carefully optimized for proximity to customers. Finally’s key insight is that grocery retailers have already invested billions of dollars building out completely optimized networks of store locations (i.e. distribution centers), not to mention strong brands with consumers. If only they could figure out some way to fit a super efficient fulfillment center into an existing store footprint in a way that allowed the store to continue to serve walk-in customers…

• Second, the answer isn’t hardware, it’s software. Since Kiva was bought by Amazon in 2012, 150 vendors have popped up selling essentially identical AGV and AMR robotic systems. Costs have plummeted, as a result. And yet retailers still can’t make a profit. It’s not about the robots, it’s about how you use them. Crucially, Finally does not manufacture any hardware at all and is completely agnostic to which robotic systems are used to enable its SFC to operate.

• Third, for Finally to work in any store, no matter the footprint or staff skill set, the solution must be comprehensive. Order management, warehouse management, warehouse control and execution, inventory management, product/pricing catalog, transport management. Finally has built it all. And their system can work with any off-the-shelf robotics system, as well as seamlessly integrate with the retailer’s existing ERP, EDI and GS1.

• Fourth, Finally understands that this is ultimately a real estate arbitrage. Grocers only need to devote a small fraction of a retail locations footprint to enable a Finally SFC. The result is (1) far more efficient use of store footprint, (2) ability to repurpose some of the store footprint to other uses (like a sushi shop or a pharmacy), (3) the ability to continue to serve walk-in customers with a better experience, and (4) new online grocery capabilities enabled by Finally.

The result is a complete end-to-end software and hardware solution that can fit on a small footprint, requires only a few junior employees to operate, and can fulfill orders profitably each and every time.

Lastly, as with any investment, a big part of our thesis with Finally is the team. We’re big fans of vertical software because of the leverage founders can get on their domain expertise. But it’s not every day you meet a team as committed to understanding an industry before trying to sell into it as the Finally team. (It’s no small feat to start your own grocery store, let alone one that was growing and shockingly profitable!). The result is a team with unique, hard-won insights about an old, slow-moving, and tech-averse industry that is in desperate need of technological innovation…just the type of tough market we love to invest in!

We believe Finally has a shot at becoming the software layer for the omnichannel era of grocery and we’re incredibly excited to partner with Avi, Lenny, Or and the entire Finally team on this journey!

A Digital Fabric for Maritime Trade
Why we invested in Portchain.

Three Keys to the Kingdom
The sometimes-competing and sometimes-aligned goals that early-stage founders must manage.

Customer-driven Entrepreneurship
Reframing the critical unlock in early-stage venture.

It’s 3am. Do You Know What Your Third-party API Calls are Doing?
Announcing our investment in Lunar.dev.

Netherlands / SaaS. TheyDo raised $34M, led by Blossom Capital, to continue investing in its customer journey mapping platform after recent acceleration of the enterprise business.

Sweden / Mobility. Voi raised $25M from existing investors to finance expansion of its electric scooter platform across Europe.

UK / Fintech. Griffin raised $24M, led by MassMutual Ventures, NordicNinja and Breega, for its “banking-as-a-service” platform, following its successful application for a banking license in the UK.

UK / Gaming. Kaedim raised $15M, led by a16z, to build its AI-powered software platform that converts 2D images into 3D models, and double down on its new marketplace which features 10,000 pre-generated 3D assets, as well as user-created 3D assets.

UK / Logistics. Packfleet raised $10M, co-led by General Catalyst and existing investor Voyager Ventures, to support the expansion of Packfleet’s carbon neutral last-mile delivery network.

Germany / DevTools. Steadybit raised $6M, led by Paladin (with participation by Angular Ventures), to continue investing in the growth of its chaos engineering-based testing platform.

ENTERPRISE/TECH NEWS

From copilot to coworker. This past week, Cognition Labs released Devin, the first “AI-powered software engineer” capable of autonomously completing relatively complex coding tasks. Watch the demo. It’s worth a few minutes of your time. And you’ll see how different this is than the coding “copilots” that have come before. This comment from Andrej Karpathy (founding team of OpenAI) on how to think about the process of automating software engineering is well worth a read, as is this profile from Ashlee Vance on the Cognition team.

To Mars. SpaceX’s Starship rocket completed its third test flight earlier this week, going “farther than ever” before. While radio/visual communication was lost towards the end of the flight, this was a significant improvement over the previous two test flights and a massive milestone for SpaceX. For some background on what went into the launch, check out this fantastic interview with Gwynne Shotwell, COO SpaceX.

Automate vulnerabilities away. Fully automated vulnerability research is going to change the cybersecurity landscape, according to Jason Clinton, CISO Anthropic. One interesting impact is that, according to Clinton, patches to fix vulnerabilities will be coming out daily. As a result, “defenders should prepare to patch their environments on a rolling basis.”

Going small. Apple seems to be “going small” on AI with its latest acquisition of Darwin AI, opines Michael Parekh in his excellent newsletter. Unlike other players, Apple is squarely focused on making models that run on devices. From the Bloomberg article on the Darwin AI acquisition: “DarwinAI has developed AI technology for visually inspecting components during the manufacturing process and serves customers in a range of industries. But one of its core technologies is making artificial intelligence systems smaller and faster. That work could be helpful to Apple, which is focused on running AI on devices rather than entirely in the cloud.”

HOW TO STARTUP

Resilience matters. In a much-discussed interview, Jensen Huang, President of Nvidia, spoke candidly about the importance of pain and suffering in the startup journey, because, as he says, “resilience matters in success.” From Huang:

“Unfortunately, resilience matters in success. I don’t know how to teach it to you except I hope suffering happens to you.

To this day, I use the word, the phrase “with pain and suffering with great glee”…and I mean that. “Boy this is going to cause lots of pain and suffering and I mean that in a happy way.

Because you want to train and refine the character of the people in your company. You want greatness out of them. Greatness is not intelligence. Greatness comes from character.

And character is not formed out of smart people but from people who have suffered. So I wish upon you ample doses of pain and suffering.”

Cyber consolidation. Last year there were 13 acquisitions in the cybersecurity space. Just this past week two more were announced. Wiz is buying Gem Security for $350M. Zscaler is buying Avalor for $310M. The age of consolidation in cybersecurity is upon us. This was to be expected. There were many cybersecurity companies out there with features posing as platforms. And those are the first vendors to get cut when CFOs are tightening belts. Expect more acquisitions in the future as the large platforms seek to fill feature gaps with up-and-comers.

HOW TO VENTURE

Mortality rates. Some interesting data from Peter Walker, head of insights at Carta, on the progression of startups through their financing journey. Of the 3,067 startups founded in 2018: 49% have shuttered, 5% were acquired, and 0.2% went public. The rest are in various stages of funding.

Pre-seed syndication. Some reflections from Charles Hudson here on the changing nature of pre-seed syndication. He looked at all his coinvestors from Fund I and II, and tried to figure out what they’re doing now (given that they aren’t coinvesting with Precursor as much anymore). Some have disappeared, some have moved into more capital intensive categories where he doesn’t play (deep tech, defense), but the vast majority have just moved up into seed. His takeaway: it’s harder to raise a pre-seed these days as the usual suspects just aren’t around the table any more.

Januar was selected to participate in Mastercard Lighthouse FINITIV’s program.

Datos Health’s remote care automation platform was selected by Northern Health to enable a healthier community by improving patient health outcomes.

Sisense targets hardcore developers with new toolkit which enables developers to compose applications using prebuilt APIs, libraries and visualizations.