The vision of autonomous driving is tested by real demand, and Rivian (RIVN.US)'s self-developed chips and software have failed to make waves

The Zhitong Finance App learned that Rivian Automotive (RIVN.US), an American electric pickup leader with the title of “Electric Vehicle Upstart”, is being re-evaluated by Wall Street analysts after showing a new self-developed autonomous driving chip and third-generation car supercomputer, the “Large Driving Model (Big Driving Model)” AI technology stack, and Rivian's exclusive in-vehicle AI “Rivian Assistant” on its “AI and Autonomous Driving Day”. However, the logic of the bullish camp and the bearish camp is very firm, highlighting Wall Street's huge differences over Rivian's future prospects, and Wall Street analysts are not disapproving of the company's autonomous driving and self-developed chip strategies, but rather tend to question the prospects for demand.

Rivian said the upcoming R2 SUV will be equipped with a chip called Rivian Autonomy Processor 1 (RAP1) and a new lidar sensor. The chip, combined with new sensors and a new AI model, will push Rivian to finally achieve full autonomous driving capabilities. Vidya Rajagopalan, vice president of hardware at Rivian Electric, said that the chip uses “multi-chip package” chiplet advanced packaging technology and has “high memory bandwidth”, which is the key to in-vehicle AI applications and fully autonomous driving. The Rivian chip has a bandwidth of 205 GB per second.

Two RAP1 chips will drive Rivian's next-generation in-vehicle computer, called Autonomy Compute Module 3. The system can process 5 billion pixels per second, and the performance is four times that of the Nvidia systems currently used in Rivian models. Rivian's SUVs and pickups currently offer assisted driving features that require the driver's supervision at all times.

The company also detailed Universal Hands-Free Driving (Universal Hands-Free Driving) achieved on more than 3.5 million miles of roads in North America through the Autonomy+ paid subscription, and plans to add LiDAR (LiDAR) and radar technology to future models. The ultimate goal is to achieve L4 autonomous driving capability, and finally have a potential Robotaxi capability foundation for fully autonomous vehicles that target Atlas and Waymo.

Wall Street investment agency Needham reiterated its “buy” rating for the electric vehicle manufacturer after the AI and Autonomous Driving Day event and raised the target price to $23. This move boosted its confidence in the Rivian market's unique positioning, as the automotive software system (which now also includes AI big models) defines vehicles as the basic threshold for the industry.

The Needham analyst team pointed out that Rivian's vertical integration enables it to more fully control the operating system, thereby achieving faster learning and functional iterations on driver interfaces and autonomous driving technology. This is thought to support long-term sustainable competitive advantage and strengthen the company's competitive confidence in the longer-term autonomous driving roadmap and differentiating points compared to traditional automakers (legacy OEMs).

Chris Pierce, an analyst at the agency, emphasized, “Rivian presented a complete technology stack, covering self-developed silicon, an autonomous driving computing architecture, a basic Large Driving Model, and a vertically integrated exclusive software platform; this separates it from traditional automakers that still rely on third-party suppliers. “In the future, Rivian will be able to drive faster software platform iteration and cost optimization, thereby creating an expanding competitive advantage and supporting stronger confidence in long-term performance prospects.”

Morgan Stanley remained cautious about Rivian after the event. Andrew Percoco, a senior analyst at the agency, stated, “Rivian is developing a complete set of hardware and software products to stay competitive in the 'Electric Vehicle 2.0' world. However, there are still a number of risks on the demand side that could limit the enormous data collection needed to reach higher levels of fully autonomous driving, and thus also limit its path to long-term profitability.” The agency gave Rivian's stock the most negative rating of “Underweight (Underweight),” and the target price was only $12.

Wells Fargo rated Rivian's stock with a neutral rating position of “equal weight (equal weight)”. The bank's analyst Colin Langan said that the enthusiastic and bullish atmosphere in the stock price reflects the optimism of consumers and investors about the company's disruptive automotive software technology, product strength and brand strategy, and advanced partnerships. “However, the company has very little room for fault tolerance in all aspects of the business. There are still many details to be seen about its limited production and commercialization history. “Rivian must prove that it can acquire an increasingly large customer base while keeping advertising costs low,” the analyst warned.

Rivian shares rose 1.3% in pre-market trading on Friday, after falling 6.1% at the close of Thursday's trading session following its “AI and Autonomous Driving Day.” The electric vehicle stock has risen more than 23% year to date, and the stock price hovered around $16.6 before the US stock market.

Rivian went public in 2021 and is one of the largest IPOs in US stock history. At the time, the company could be called the “pearl in the pocket of the American automobile industry” and was regarded by Wall Street as Tesla's strongest competitor. Prior to major traditional car manufacturers, Rivian pioneered full-size electric pickups and SUVs. But since then, the company has continued to face operational difficulties — its only plant in Illinois is expected to produce less than 50,000 vehicles this year, far below the capacity limit.

Because of continuing to burn money, Rivian continues to lay off employees, and its stock price has dropped more than 80% from its high point after the IPO. Even so, Volkswagen, which has signed a long-term cooperation agreement with Rivian, has promised to invest nearly $6 billion in joint ventures between the two sides to utilize Rivian's expertise in software platforms and automated automobile manufacturing lines. Meanwhile, Rivian continues to recruit top talent from Tesla, Apple, and Silicon Valley.

The core of Rivian's current vehicle and upcoming R2 SUV is a basic autonomous driving system called the Large Driving Model. Similar to Tesla's FSD, this system is capable of deep learning from past and future driving behavior, enabling Rivian to improve autonomous driving capabilities for older R1 models lacking lidar and equipped with Nvidia Orin chips before launching a more complete R2 autonomous driving platform in 2027. The first generation upgrade function of the Rivian software platform will be relatively limited, far less than the “full self-driving” (FSD) function currently provided by Tesla.