The performance
layer
between humanoids
and the ground.

The first sole engineered for
humanoid biomechanics.
Sunflower compliant sole — hexagonal lattice geometry
Humanoid robot legs in sneakers
The Problem

Flat feet are
draining humanoid
battery.

Rigid flat foot today
Flat feet→ Torque spikes→ ↑ Current→ I²R loss→ Battery drained
Our compliant sole
Adaptive geometry→ Smooth torque→ ↓ Current→ Less I²R loss→ Longer runtime
The Gap

The math for humanoids
doesn't add up yet.

Target · Roland Berger 2026

Transformative humanoid

Operating rate $2 / hr
Unit price w/ training $20k–30k
Service lifespan 20,000 hrs
Reality · Shipping today

Where humanoids actually are

Runtime per charge 2–8 hrs
Lifespan before service < 1 year
Software vs hardware gap 3–5 years
Source: Roland Berger, Humanoid Robots 2026: The Convergence Moment. Public OEM disclosures.
The Hidden Math

$2/hr assumes 20,000 hours.
Today's robots fail at 2,500.

The Promise
$25,000 robot ÷ 20,000 hrs
$2/hr
Roland Berger 2026 target
Today's Reality
$25,000 robot ÷ 2,500 hrs
8×more
Same robot, one-eighth the operating life
We are not selling feet. We are protecting the most expensive part of the robot — the actuators — from early death. More cycles per robot means closer to the economics that make humanoid deployment viable.
Root Cause

A bad foot cascades.
Humanoids don't know this yet.

Human

The medical precedent

Poor foot biomechanics causes knee pain, hip pain, and back pain. The foot is the load-distribution layer. Get it wrong and every joint above it compensates, strains, and fails early. Orthotics at the foot fix problems three joints away.

Foot→ Ankle→ Knee→ Hip→ Back
Humanoid

The same failure mode

Rigid flat feet force the ankle actuator to absorb every heel-strike impact. Torque spikes propagate up the chain: knee motor load, hip motor load, higher current draw, faster actuator wear, shorter battery life — and premature end of service.

Rigid Foot→ Torque Spike→ Chain Wear→ Early Failure
The Solution

We sit between the robot and the ground.

Patented compliant technology absorbs heel-strike before it reaches the ankle and returns energy through toe-off. No sensors. No batteries in the sole. Tuned per robot, mission, and terrain.

Battery
+24 min
More runtime per charge on Unitree G1 baseline
Peak Ankle Torque
−12.9%
Reduced mechanical load on ankle actuator
Durability
1.17M
Cycles simulated with no functional loss
Energy
−6.4%
Energy consumption per gait cycle
MuJoCo simulation, Unitree G1 baseline vs rigid foot. Hardware validation underway.
Competitive Position

We are the first sole
engineered for humanoid biomechanics.

01

Patent on the method

We didn't patent a sole. We patented the torque-reduction method at ground contact. The performance gain is what's protected — not the product form. Broad defensive coverage.

USPTO Non-Provisional Patent Filed
02

Real hardware data

Every pilot we ship feeds our parametric model. Competitors simulate. We accumulate cycles on actual robots, building a proprietary performance dataset no one else can replicate.

MuJoCo · Pilots Expanding
03

Design maturity

Years iterating compliant flexure mechanisms. New entrants don't start from zero — they start behind. They hit our learning curve on day one and can't catch up without our accumulated dataset.

1.17M Cycles Simulated
Market Opportunity

Michelin didn't make cars.
We don't make robots.

Every humanoid platform needs a sole. We make the only one built for the job.

$38B
Humanoid market by 2035
8+
Commercial OEMs shipping today, all on rigid flat feet
10k+
Units shipping in 2026
×2
Soles per robot, replaced regularly — recurring revenue by design
Public OEM disclosures and third-party market analysis.
Closed Service

Design. Manufacture. Ship.
Three weeks end to end.

Intake

OEM shares specs

Robot model, mass, gait, mission, terrain, fleet size. NDA signed. We validate fit.

Engineer

We design and simulate

Parametric compliant geometry, tuned in MuJoCo, signed off by our team before a single part is cut.

Deliver

We manufacture and ship

Validated supply chain. Certified parts. OEM installs and measures. We iterate on the data returned.

Closed. No open source. No design files leave our hands. The OEM gets the product. We keep the IP. The patent stays valuable.
Revenue Path

Three years to category default.

Year 1 · Pilot

Land first OEMs

Three to five paid pilots. 50–200 soles each. Prove performance on real hardware.

Hardware-validated data. First real-world performance dataset. Patent prosecution underway.

$25k–50k
Year 2 · Fleet

First supply contracts

Two to three OEMs on recurring supply. 500 to 5,000 soles each at $150–300/unit.

Recurring supply relationship. Parametric model proves value across robot platforms and terrains.

$300k–1.5M
Year 3 · Scale

Category default

Five-plus OEMs locked in. Recurring revenue. Referenced across the industry.

Industry standard for humanoid ground-contact optimization. Moat compounds with every robot in the field.

$3M–10M
The Ask

Join us at the ground floor
of humanoid infrastructure.

Pre-Seed · Open Now
$1M
$100k already committed · Closing Q2 2026

Use of funds: three OEM pilots, patent prosecution, first five hires. 18 months of runway to the Series A milestone.

contact@sunflowerrobotics.com
Milestone to Series A
18 mo
Runway to first OEM contract

One signed OEM supply contract.

Hardware-validated performance data.

Category leadership established.

We are raising to lock up the compliant sole category before OEMs try to build it themselves.