The Hard Thing About Hard Tech

Or How "Welcome to the Roboverse" Became Real

With a nod to Ben Horowitz' book The Hard Thing about Hard Things, I wanted to explore why innovating in deep tech/hard tech is so challenging and how founders can cultivate the necessary mental models and resilience to succeed in this space. I also get to brag a little.

I meet regularly with founders in robotics, sensing and applied AI. Their companies exist at the intersection of hardware and software. Critically, their mission is to impact the physical world we live in, not just move pixels on a screen. And the real world is chaotic, slow and resistant to change.

Many of the conversations I have with deep tech founders go something like this:

We've developed this awesome tech after years of R&D. Our CTO came from [prestigious university] and has key patents on [some specific tech]. We have proven that the technology works, our pilot customers saw great results, but they still won't commit to deploying it at scale until the product is mature. We secured some [government grants/angel investment] to get here, but now we need some real money to develop the product further. However, investors want to see growing revenue in the millions. How can we get revenue if we don't have a product to sell? How can we have a product to sell if we don't have the funds to develop it?

If this resonates with any of you reading this article, just know that you are heard. I understand your pain but, unfortunately, there's no easy fix -- regardless of what any consultants might tell you. You picked the hard path of hard tech. The main question is: Are you willing to see it through?

The challenge that deep tech founders run into is that true innovation, particularly in the physical world, engenders persistent skepticism -- from customers, investors, candidates, press and even their parents. That's because hard tech is about much more than a new "experience"; it's actually creating something that, while still operating within the laws of physics, is really pushing the limit of what is actually possible. And the hallmark if innovation in this space is that people haven't seen it work before, so they're right to question whether it will work.

Founders and their teams must steel themselves for a longer journey during which the feasibility and business viability of the work they're doing will be constantly called into question. Moreover, the team itself will not know the answer in advance: they're embarking on an adventure without a clear path to their destination. Without the confidence and resilience to work past all the obstacles and the nay-sayers to see it through, they might as well not get started.

Working in deep tech requires a different mental model than, say, creating a cheaper/regional copy of a proven business, developing a new mobile front end to an existing process for increased convenience, or convincing people to create content for free so others can scroll endlessly through it. It's not that those are easy to do, just that success or failure follows a more well-defined trajectory.

To avoid just talking about generalities or throwing any founder I've spoken to under the bus, I'll use my own journey at InOrbit.AI below. My goal is sharing some lessons learned along the way that may help others.

InOrbit is a platform for robot operations or RobOps at scale. It's a software platform to enable companies operating autonomous or "smart" robots to maximize the value of every robot. It currently supports robots across four continents carrying tasks as diverse as material handling, restocking, delivery, inventory management or cleaning, and is trusted by some of the top brands in the world. However, that wasn't always the case.

TBH, since forming the company we have faced relentless skepticism. Initially, the typical question was "Why would any robotics company use your software instead of creating their own?". After we showed that our platform was 10X better than any internally developed tools, the question became "Will you be able to scale technically? Can you find more customers?" When we proved that this is indeed a recurring and growing need, we decided to focus on multi-vendor orchestration to enable the largest users of robots to mix-and-match. Until recently, people would ask us "Why wouldn't customer get what you offer from their robot vendors?" and "Are there any customers so advanced as to have multiple robots vendors?"

Along the way, we discovered that we needed a combination of show AND tell. First, we developed a compelling narrative, drawing parallels with other industries that became more specialized as they matured. In order to dispel skepticism, we had to show the tech at work, but we couldn't do it at our customers' facilities until they gave us access, usually in the form of a proof of concept engagement. PoCs are fine, but they take time. In order to go faster, we decided to, in a way, become our own customer.

We did something very unusual. (Some even called it crazy.) We decided to create the InOrbit Robot Space in the heart of Silicon Valley as a space representing the environment of many of our customers, and also opened it up for our ecosystem partners, the robotics community at large and even the general public. This resulted in over 250 demos in just 9 months -- not all of them with customers, but each one an opportunity to tell our story. Over 1,300 people had an experience interacting with our robots, our software or our brand.

What I want to highlight here is that this went against the prevailing current. We were told it wouldn't work or, worse, that it was a stupid move, but we were willing to make that bet and also sustain the effort over time to get the outcomes we were seeking. Differentiation comes from doing something different, and many people will reject different until they've seen the results.

In deep tech, it often takes a while for reality to catch up to the vision. Founders can often visualize the future before others do, by projecting forward current technologies. But often, progress on many interdependent areas is required, including some that are outside the startup's control.

Two years ago, I published an article on Forbes entitled "Forget The Metaverse — The Roboverse Is Already Here". It painted a vision for how AI-powered robots loaded with sensors that allow them to navigate the real world would converge with 3D simulated worlds through spatial technology. The article included the phrase "Welcome to the roboverse". From there, it evolved into this newsletter, which aims to cover the latest advances in robotics, automation and future of work, with a personal view on entrepreneurship and startups.

We are finally making good on this vision, through a collaboration with Nvidia and other partners such as OTTO Motors (now part of Rockwell Automation) and Ekumen. We are able to showcase how to seamlessly blend simulated and real-world robot operations through a single pane of glass. Using realistic simulations opens up significant opportunities to define robot missions, execute them in simulated space, then deploy the same missions in the physical world, get consistent analytics about both and continue iterating to drive optimization.

So, if you're a founder with a big vision for the future, your main job is to tell your story, so others can share your vision and join your effort, then make it a reality. Your confidence will be tested at every turn; you will be told it can't be done by (mostly well-intentioned) experts; you will be turned down by customers and investors alike. But if you are right and you persevere (plus a little bit of luck doesn't hurt), you'll eventually have the satisfaction to prove the haters wrong.