EN
Back to the archive

The archive · Hardware & Devices · Product decision · 2025-2026

Heygo bets an AI snowboard coach built on boot sensors becomes consumer hardware

Ex-ByteDance, DJI and Tencent staff built boot-mounted sensors plus AI that coach snowboarders in real time; Thick Snow led a multi-million-dollar angel.

Heygo (Heygo.ai)

The betThat skiing's physical 'dark data' under the boot can be measured, taught to an AI, and sold as affordable consumer hardware with a coaching subscription.Building

What the business is

Heygo makes two IMU sensors that clip to a snowboard boot, plus an AI coaching app that gives real-time voice feedback during a run, lift-line reviews, and personalized progress plans, with a paid tier for deeper analysis.

Starting capitalAngel round of several million US dollars, exclusively from Thick Snow / Houxue Institution (36Kr and iHeima, September 2026).

How it started

Wu Zhenhua, a former ByteDance Feishu regional and user-growth lead who had also been CMO at Vika, founded Heygo in 2025 with co-founders from Tencent, ByteDance, and DJI. The trigger was his own years of skiing frustration: private lessons cost thousands of yuan and coaches could not remember or measure what a rider's body actually did. The team spent two ski seasons in Japan, Xinjiang, and New Zealand collecting millions of labeled runs in temperatures down to -30 degrees Celsius.

What happened

The team built a data-labeling pipeline that breaks each run into turns and indicators, and says current measurement accuracy is within three degrees. In September 2026, 36Kr reported the multi-million-dollar angel round led by Thick Snow Institution, and Heygo scheduled its first hardware presale for September with October delivery across 38 countries and regions; basic software comes with the hardware, while advanced analysis and plans sit behind a subscription.

No ending yet — it is still running.

Background

Heygo is a Beijing-founded hardware and AI startup trying to turn ski coaching into a consumer product. Two sensors clipped to the side of a snowboard boot measure ankle and board motion, and an AI model trained on real runs gives voice feedback through headphones at critical moments, reviews a run on the lift, and builds a personalized progress plan over time. The basic tier ships with the hardware; deeper analysis and individual plans are a subscription.

The company was founded in 2025 by Wu Zhenhua, a former Feishu regional and user-growth lead at ByteDance and ex-CMO of Vika, with co-founders from ByteDance, Tencent, and DJI. The founders' claim was that skiing runs on physical 'dark data' - edge, carve, and weight transfer under the boot - that cameras and even coaches cannot capture. The team spent two ski seasons across Japan, Xinjiang, and New Zealand collecting millions of labeled runs in extreme cold to train the mapping from foot sensors to actual ski behavior.

The bet attracted early backing: 36Kr reported in September 2026 that Heygo had closed a multi-million-dollar angel round led by Thick Snow Institution, and iHeima independently covered the same round on 2026-09-03. Heygo's first snowboarding hardware was scheduled for presale in September 2026 with delivery in October across 38 countries and regions, supported by ski associations and top instructors in New Zealand and Canada.

The open question is commercial verification: skiing hardware plus subscription adds cost to an already expensive hobby, and Heygo has not yet shipped units or disclosed user numbers. The team argues that a 1% penetration of regular skiers can support a strong business, and that the sensor-fusion architecture and personal motion model can later extend to surfing and skateboarding.

What has to be true

  • The founders attacked a real measurement gap: ski boots deform, bindings transmit force, and the board's edge state is invisible to video, so they instrumented the boot instead of the mountain.
  • They validated the data problem firsthand over two hemisphere-spanning ski seasons rather than in a lab, collecting millions of labeled runs in extreme temperatures.
  • Real-time voice feedback gives an immediate reason to buy, instead of asking consumers to watch post-session analytics.
  • The team is deliberately counter-consensus on market size: skiing is niche, but its users have high willingness to pay, and the same motion model could extend to surf and skate.
  • The main risk is that a thousand-yuan hardware plus subscription raises the bar for casual skiers, and no user or revenue figures exist yet.

What can be applied

A niche sport can be a wedge when sensors and data turn expert coaching into affordable feedback; the real asset is a motion model that may transfer.

Aftermath

As of 2026-09-04, Heygo remains a pre-delivery startup: its English site advertises real-time AI snowboard coaching, Discord community, and a newsletter, and Chinese coverage dated September 1-3, 2026 says presale is due that month with first deliveries in October across 38 countries and regions. The company reports several million US dollars in angel funding from Thick Snow Institution and claims millions of real ski-run data points and sub-three-degree measurement accuracy, but has not published customer numbers, shipments, or revenue.

Sources

spotted an error? The archive wants to know.

Your turn

You just read one. Describe what you are building, and see who is betting on the same thing.

Free account · 3 free questions · no card

Related cases