Momenta IPO: How Chinese Startup Chose L2 Instead of Waymo Path

Momenta held its IPO on the Hong Kong Stock Exchange on July 8, 2026. The first day of trading valued the company at 700 billion Hong Kong dollars (approximately $90 billion). This is a venture startup, founded in 2016, that chose a path opposite to the rest of the autonomous driving industry.

Strategy: L2 Instead of L4

In 2018–2019, the entire AI industry world bet on L4 (full autonomy without a driver). Waymo received a valuation of $175 billion, PonyAI and Yushu Zhixing raised hundreds of millions. Capital reported heroic driverless journeys.

CEO Cao Xudong chose L2 — driving assistance (adaptive cruise control, lane-keeping, partial steering). This was considered niche, boring. Financing in 2019 was slow. Inside the company, doubts arose.

But Cao believed in the "data flywheel" — the idea that if a system can collect data from millions of cars, it will quickly converge toward perfection. Waymo's L4 fleet is limited (expensive cars, few vehicles). L2 can be deployed immediately across all production vehicles.

Results: 3000 Times More Data

By 2026, the strategy had won.

• Market share: 60% of all L2 systems deployed on production vehicles in Europe
• Specific models: 50%+ for each of 210+ customized automotive platforms
• Deployed in the real world: 100+ million vehicles with Momenta software
• Team: just over 1000 employees
• Margin: grew over three years from 17.5% → 49% → 71.6%

Cao told investors about a visit to Tesla to test Musk's FSD (Full Self-Driving). Tesla does this with 200–300 engineers, Momenta — with 1000+. He interpreted this as inefficiency. Internally he promised: "We need to be even faster, more efficient."

Engineering Efficiency as Strategy

All of Momenta's successes are tied to engineering efficiency. The company invested in automation tools: CI/CD pipeline, road-testing robots, data versioning. Cao personally tracks the top-10 tools.

One example: the "rt spa" tool for automating the complete road-testing cycle. Cao himself named the tool, led the first 10 architecture meetings, attended every iteration. Result: an algorithm researcher presses a button, and in 4 hours — a ready report on version quality.

Second: a system for preparing versions for release. How to determine if a new version is good? You need to look at 1000+ metrics. Momenta automated this through a dashboard.

Competition and Challenges

The L2 market stopped being a niche. Competitors: Huawei, Horizon Robotics, Qingzhou, Waymo, Baidu Apollo. Prices for L2 systems have fallen. Qualcomm and NVIDIA compete on chips. New competitors (Wenmiao, Elementai) try to undercut Momenta on price.

Cao sees the future in software-hardware integration and world models for L4 — but that's still in the future (2027–2028).

What This Means

Momenta demonstrates that in the AI industry, the same principles apply as in software/hardware engineering: data > model size, iterativeness > single breakthrough, engineering culture > genius. The company won not because it had the best scientists, but because it chose the right data scale and invested in production engineering. This is a lesson for the entire AI industry: not everything needs to be won in laboratories.