No one is debating the reduction of dimensionality for L4 after the bankruptcy of ARGO AI

ARGO AI Has Recently Closed Down

Although this is external news, we are able to see some of the difficulties that L4 autonomous driving companies are facing from it. Whether it’s an autonomous driving company or a car company like Tesla, everyone is moving towards a common goal, which is to gradually popularize autonomous driving.

Who is ARGO AI?

ARGO AI is an autonomous driving company. Its founder, Bryan Salesky, was the head of hardware development for Google’s autonomous driving project, and Peter Rander was the chief engineer for Uber’s autonomous driving department, ATG.

In October 2016, ARGO AI was officially established and quickly recruited a team of 200 people with the founders’ qualifications and their position in the North American autonomous driving industry by the end of the year.

Coincidentally, the same year, Ford’s arch-rival acquired Cruise, a start-up in autonomous driving, which made Ford a bit anxious. Fortunately, ARGO AI, with its relatively strong background and qualifications, received 1 billion dollars from Ford just over three months after its establishment.

As a pioneer in the pure electric transformation, Volkswagen is also very eager in the European stronghold where there is a relative shortage of autonomous driving talents. At the same time, Volkswagen’s relationship with Ford is relatively close, so the focus has naturally shifted to ARGO AI.

In June 2020, Volkswagen invested $2.6 billion in ARGO AI, and even included its own autonomous driving department in Munich as a branch of ARGO AI in Europe, with both Volkswagen and its dear friend Ford holding 42% of the shares.

Last year, Volkswagen announced its development strategies in software, especially in autonomous driving. Volkswagen has two solutions: one is for private use and is managed by CARIAD, and the other is for shared mobility and is managed by ARGO AI.

ARGO AI’s Collapse

ARGO AI announced layoffs as early as July this year, which actually hinted at the unhealthy state of the entire company.

The core reason for ARGO AI’s collapse is actually the predicament that most L4 companies currently face. The seemingly bright commercial prospects not only fail to make money in practice, but also become a continuous money-burning pit.

As the main investors of ARGO AI, Volkswagen and Ford have been waiting for L4 technology feedback. For example, when Ford invested in ARGO AI initially, it planned to provide autonomous driving solutions that were already road-worthy for use by 2021.### No Reduction without Progress

When the autonomous driving race first started, a large number of people in the market were focused on the higher cost and higher ability of L4 solutions. They hoped to achieve the reduction of difficulty and cost of achieving autonomous driving for individual cars through high-dimensional capability.

But after a few years, most L4 companies in the world, especially those in China, have almost all launched their own automated valet parking solutions to help automakers promote and popularize the L2 level of driver assistance.

For these self-driving companies that burn through cash like water on the Robtaxi project, they finally got their second leg. Although this cash flow race not only has competition from similar L4-category valet parking products, it also has competition from Tier 1 companies like Horizon and Bosch.

But whether it is from a risk assessment perspective or development prospects in the future, this cash flow will be the decisive factor for their entire valuation and development. Because the financing scale of L4’s fleet is almost capped, commercial realization is unlikely.

Data, Data, Data!

Whether L4 companies invest billions to establish self-driving fleets or enterprises like Tesla push individual cars through shadow algorithms, the core of achieving autonomous driving at the current stage is perception. On the one hand, it is to enhance perception ability through hardware, on the other hand, it is to continuously collect corner case scenes. Whoever has enough data and can use it effectively will grow faster.

This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.