Why I'm Building an App Platform for Robot Vacuums Instead of Humanoids
November 26, 2025
I'm building Remake.ai to
support all types of navigation-capable robots, but I'm
starting with SLAM-capable home robot vacuum cleaners equipped
with 2D LiDAR sensors. Not humanoids. Not because I'm
risk-averse, but because I've learned that the hardest part
of building a platform isn't the technology—it's
getting people to actually use it.
Here's why
robot vacuums are the right starting point.
The Installed Base Problem
Home vacuum cleaners have become ubiquitous. US
household penetration for robot vacuum cleaners exceeds 10%,
with some estimates suggesting it could reach 20% in the near
future. The global market shipped 22.1 million units in 2025,
generating $4.98 billion in revenue. This isn't a niche
market anymore—it's mainstream.
For my
startup's business model—apps for robots—this means I can
explain what we do to almost anyone. "It's like an App
Store, but for your Roomba" gets instant understanding. Try
explaining an app platform for a humanoid robot that
doesn't exist in anyone's home yet.
The De-Risked Form Factor
Consumers have spent over a decade getting
comfortable with robots moving around their homes. Yes, people
trip over their Roombas occasionally. Yes, they run into
furniture. But these safety concerns have been largely
normalized. The form factor is accepted.
Compare this
to humanoids. I asked my wife if she'd be comfortable with
a humanoid robot at home—one that's stronger than her,
powered by AI that could potentially malfunction or be hacked.
Her answer: absolutely not. Mine too, frankly.
Even
smaller humanoids make me uncomfortable. Child-sized robots
doing household tasks? That crosses into disturbing territory
for a lot of people, myself included.
The Manufacturing Ecosystem
Thanks to companies like iRobot and the explosion
of Chinese manufacturers (Roborock, Ecovacs, Xiaomi),
there's now a thriving global supply chain for robot vacuum
components. Costs have plummeted. A capable SLAM-equipped vacuum
with LiDAR can be manufactured affordably, which means the
potential market for third-party apps is economically viable.
Humanoids?
Tesla is targeting 5,000 Optimus units in 2025. BYD aims for
1,500. Even with aggressive scaling plans, we're talking
about tens of thousands of units in the near term, not millions.
The manufacturing ecosystem is still being built.
The Uncanny Valley Problem
There's substantial robotics research on the
"uncanny valley"—the phenomenon where human-like
robots that are almost but not quite
human trigger discomfort and aversion. First described by
roboticist Masahiro Mori in 1970, the effect has been validated
across numerous studies using fMRI, behavioral research, and
real-world robot interactions.
Recent research has
even identified two uncanny valleys:
one for highly human-like robots, and another for moderately
human-like robots. The brain's parietal cortex lights up
when it detects mismatches between human-like appearance and
robotic motion—our visual processing systems literally reject
the contradiction.
Robot vacuums don't have this
problem. They're clearly robots. No one expects them to
look human, so there's no uncanny valley to navigate.
The Privacy Advantage
SLAM-enabled home vacuums use 2D LiDAR sensors that
measure distances to surrounding objects—and even that only at
floor level. They can detect walls, furniture, people, and pets
without actually capturing images. No cameras recording your
daily life. No facial recognition. No visual data being uploaded
to the cloud.
Even if you're comfortable with
humanoid robots in principle, would you want one watching you
with cameras and advanced sensors in your private residence
every day? The major humanoid companies—Tesla, Figure AI,
Agility Robotics, Boston Dynamics, Apptronik, UBTECH, Sanctuary
AI, Unitree, Engineered Arts—are all building robots with
extensive visual and sensor arrays. That's a fundamentally
different privacy proposition.
What This Means for Remake.ai
A SLAM-enabled home vacuum cleaner robot isn't
perfect, but it's:
- Affordable
- Already in
millions of homes
- Capable of running apps for
entertainment, security, companionship, education, and more
-
Privacy-preserving by design
- Accepted by consumers as
safe
It's a platform that exists
today, not one I'm betting will
exist in five years.
We're building the first
App Store for consumer robots. Starting with vacuums means we
can ship to real customers, generate real revenue, and prove the
platform model works—before the humanoid wave arrives. And when
it does, we'll already have the infrastructure, developer
ecosystem, and consumer trust to expand.
Sometimes
the best strategy isn't to build for the future. It's
to build for the largest deployable base you can reach right
now.