← Back
Untu AI · Founding Round · 2026
Every category has a company that got there first

India doesn't have its
legged-robotics company.
We're building it.

Not a pitch for a robot. A pitch for the decision to become the company an entire category gets built around.

Founding round · ₹5 Cr
01
The scale of the thing

Legs are the layer of autonomy
nobody has finished building yet.

Wheels needed a century to cover the world. Rotors took thirty years. Legs — the only form factor that goes wherever a person goes — are still, in 2026, an open category with no settled global winner and no Indian company in the race at all.

~$3B
Global quadruped market today — early enough that no one owns it yet
~$17B
Where it goes by 2035, on current trajectory
Zero
Indian companies that have shipped a legged robot that walks on its own
Untu AI · Ground Zero
02
The category is still open

Every legged-robotics leader on Earth
was one small team that got there first.

Boston Dynamics. Unitree. Ghost Robotics. None of them were the biggest company in robotics when they started — they were the first team to make the walk work, in their country, before anyone else tried seriously.

Untu AI · Ground Zero
03
The hard part is already done

It walks. Untethered, on its own compute, ten months from a blank file.

Jul 2025
Empty CAD file — the starting line
May 2026
Untethered walk on real hardware, own compute
2
Full redesign-and-reprint cycles already completed
3
Founders. No bought chassis, no licensed controller, no black box.

Most attempts at this never clear this line. We are past the part that kills most legged-robotics companies — the part where nobody knows yet if the team can make the machine walk at all.

Untu AI · Ground Zero
04
What building a category company looks like

Years of unglamorous work,
before anyone outside the room believes it.

Every category-defining hardware company spent years being a small team in a room nobody was watching, iterating on a machine that mostly didn't work yet. That stage is not a detour before the real company starts. It is the company.

Untu AI · Ground Zero
05
The ten-year shape

A quadruped is the first product.
Not the whole company.

Now → Year 2

One platform, proven

Rumi, field-hardened — the loop validated end to end on real terrain.

Year 3 → 5

Multiple form factors

The same design→train→transfer loop applied beyond one chassis — the loop is the asset, the robot is one expression of it.

Year 5 → 10

A category company

Manufacturing at national scale, sovereign-relevant supply chain, the default Indian answer to legged autonomy.

This round doesn't fund a product roadmap. It funds the first ten years of a company someone was always going to build — the only open question is which country's team builds it.

Untu AI · Ground Zero
06
Why India, why now

The talent, the timing and the money
are aligned for maybe a five-year window.

Talent

A generation of Indian robotics researchers with sim-to-real experience, trained on problems the world didn't have five years ago.

Timing

GPU-parallel simulation just made legged control a small-team problem instead of a national-lab problem. That door doesn't stay this open.

Capital

India is writing ₹1.39 lakh crore a year of mandated-domestic defence cheques with no legged-robotics company yet positioned to cash one.

Untu AI · Ground Zero
07
The ask, reframed

This isn't seed capital for a robot startup.
It's founding capital for a category.

₹5 Cr doesn't buy a product milestone. It buys the next two years of the only team in the country that has already cleared the hardest binary in legged robotics — before someone else decides to be first instead.

₹5 Cr · 10% · founding round Untu AI — the ground layer, built from India, for the next decade
Untu AI · Ground Zero
08