A behind the scenes look at Russian search giant Yandex's efforts to bring self-driving technology to Moscow's chaotic, snow filled roads
Context & Ripple Effects
This ZDNet feature lands between two milestones in Yandex's autonomy arc: the closed-circuit prototype unveiled in May 2017 and the free autonomous ride-hailing trial launched months later in Innopolis. Its contribution is the why: Moscow's snow and traffic chaos are framed not as an obstacle but as the training ground that forces the stack to generalize.
That framing matters because Yandex's roadmap quickly outgrew the search company's balance sheet — a hardware-and-software partnership with Hyundai followed in 2019, then a $150M spinout of the self-driving unit away from the Uber joint venture in 2020.
First-order effects
- Yandex's engineering team shifts validation from closed circuits to live Moscow streets, where snow, erratic drivers, and road conditions become the core test cases its competitors in milder climates don't face.
Second-order effects
- Proving the system on Moscow's roads gives Yandex the credibility to run the Innopolis public ride-hailing trial later that year, and makes the unit attractive enough for Hyundai to co-develop hardware and software rather than build alone.
Third-order effects
- If the pattern holds, autonomy at Yandex separates from the parent's ad-search economics entirely — culminating in the 2021 deal buying out Uber's stakes and making self-driving a standalone, domestically controlled Russian asset.
The trend: Russia's largest internet company is building a sovereign autonomous-driving stack, tested on conditions Western rivals don't train for and progressively walled off from foreign joint-venture partners.