Hey people! Welcome back to Forests Over Trees, your weekly tech strategy newsletter. It’s time to zoom-out, connect dots, and (try to) predict the future.

Robot Roll-up

Warning: this week’s topic has me extremely excited. If your mind is racing at the end…you’re not alone.

Robots have been back in the news!

I promise I’ll throw you a lifeline via a framework in a minute. But first, let’s get you as excited as I am.

• Re-factored tractors – Bluewhite is a company retrofitting autonomous robotics to any tractor. Your entire current fleet can self-drive,  and they’ll help manage the tech for you. For my acronym lovers, they’re calling it RaaS (robotics as a service). They raised $39M in January.

• Apple is getting involved – According to Bloomberg, they’re exploring personal robotics as a “next big thing”. If anyone can deliver a beautiful, simple, powerful robot that consumers actually trust, it’s the Cupertino company.

• Starship is nailing it – Their delivery robots have traveled 11M miles! That’s short of Waymo’s 20M, but Waymo had a 5 year headstart. They’re officially the biggest autonomous delivery service, even though they’re branded as focusing primarily on last mile. They raised $90M in February.

• Cruise’s co-founder left self-driving for robots – Granted that Cruise is having a really tough time since a crash led GM to pull back (as I’ve dissected before), this is still a cool move. Kyle Vogt, the former CEO, launched The Bot Company in May, and he’s already raised $150M! Here’s the pitch from his LinkedIn for reference “We’re building bots that do chores so you don’t have to.” To which I must say, respectfully, “shut up and take all my money”.

Are you sufficiently jazzed!?

Good.

Alright, so if we zoom-out, why is all of this happening now?

Everything is getting better

Basically, a bunch of good things are happening at once.

Faster

Computing power (on device and in the cloud) has improved dramatically over the last 5 years, for robots that can “think” faster.

Smarter

And, they’re thinking of better ideas thanks to advances in AI.

One robot-specific AI breakthrough came from Google’s Deepmind teaming up with 30+ other researchers globally for what was called the RT-X project (imagine Michael Jordan and the USA Dream Team, but for robotics!).

They proved that by pooling training data, they could improve the robot’s ability to generalize by 3x. Generalizing is hugely important. It’s the difference between spelling-out every single instruction needed vs. just telling the robot what the goal is.

Cheaper

And, hardware costs are coming down! According to a Stanford report, the cost of a robotic arm dropped 46% from 2017 to 2021.

Because of all that, VCs are salivating at the chance to be early to another AI-like wave of investment.

So what’s the landscape looking like, and where are the opportunities?

One Robotics Framework to Rule them All

OK, obviously this is a complex and evolving space, but here’s my 2 cents on a helpful (albeit dumbed-down) industry overview.

For me, the most distinguishing characteristics of today’s robots are:

1. Are they generalists or are they specialists?

2. Are they highly capable/impactful, or just full of potential?

Today, I think the answers to those questions are inversely correlated, so visually it looks like this:

At the bottom-right of the curve, these are purpose-built, fully commercialized, all-star robots, like Kiva (Amazon’s warehouse robot), Starship (which we talked about earlier), or even dishwashers. Their impact has been (mostly) realized, and future potential is capped.

As you move up the curve, you get robots with lots of potential, but which haven’t grown into it yet. And they are generalists – highly adaptable and good at many different things. In this camp, I would put all the “humanoid” robots.

One that’s caught a lot of attention is from Figure AI. In February, they raised $675M from the likes of Jeff Bezos, OpenAI, and Nvidia. They claim to only need 2 years to get robots to commercial buyers… I’ll believe that when I see it. This 2-minute demo is pretty impressive though.  

Another one that’s a sign of the times is Tesla’s “Optimus”. Just as with his other bold predictions, Elon is promising to start selling this one in 2025. I’ll believe that when hell freezes over.

Wrapping Up

As we put a bow on this first foray into robotics, I want to leave you with one other thought. This kept coming up for me in this week’s research. And unfortunately, it doesn’t fit neatly in the framework…

Robotics will impact every part of tech. It’s not just the use cases that matter (the things the robots can do). The “players” (companies, countries) creating them matter just as much.

All the things we cover here together will get impacted. For example:

• Japan is including robotics in their $26B in investments aimed at fighting population loss with tech (which we covered in May)…

• According to TechCrunch, Nvidia is jumping in to design the chips and some of the AI middleware to make robots work…

• China is including robotics in their plan to fight an ongoing trade war with the US (which we covered way too much! Here and here)…

So grab some popcorn and strap in folks! Should be a fun show.

Quote of the Week

— Reid Hoffman

Quick News Reactions

• Anduril’s founder did what? — He made a retro Gameboy, he’s selling them for $200, and he’s shipping them during the 2024 holiday season. Palmer Luckey is clearly incredibly creative, but I would not have expected Nintendo clones from the founder of a defense tech company.

• FTC still jabbing and hooking — They started a probe of Microsoft and Inflection AI’s partnership, after Microsoft acqui-hired the company for a ~$650 “licensing fee” (potentially doing it this way to avoid anti-trust scrutiny). The chaotic partnering up in AI will absolutely continue despite the scrutiny. It’s not worth missing a goldrush just because you might get your wrist slapped.

•  SAP is acquiring a SaaS tool — They’re paying $1.5B for WalkMe, which helps users get the most out of other software by explaining/prompting them to explore new parts of the UI (almost like a tutorial). I doubt they’ll get any anti-trust scrutiny, especially since users will rejoice at the opportunity for even a marginally more user-friendly SAP UI (kidding!…sort of)