Rivian is betting that custom silicon will be its shortcut to safer, more capable automated driving, unveiling a new in-house AI chip and a broader autonomy platform that reaches from sensors to cloud training. The move shifts the electric truck maker away from off-the-shelf processors and toward a vertically integrated stack that it hopes can support both personally owned self-driving vehicles and future robotaxis.
By designing its own processor, Rivian is trying to control everything from how raw camera and lidar data is crunched in real time to how new driving behaviors are learned and pushed back to its fleet. It is a high-cost, high-stakes strategy that puts the company in more direct competition with Tesla, Nvidia and traditional suppliers, even as Rivian continues to wrestle with the economics of building and selling EVs at scale.
RAP1 and ACM3: Rivian’s custom brain for autonomy
The centerpiece of Rivian’s new strategy is a proprietary processor called the Rivian Autonomy Processor, or RAP1, paired with an autonomy compute module known as ACM3. At the core of Rivian’s platform, the ACM3 is specified at 1600 sparse INT8 TOPS (Trillion Operations Per Second), a level of performance that is designed to handle the edge cases of driving rather than just ideal highway conditions. That processing headroom is meant to let the system fuse data from multiple sensors, run complex neural networks and still leave margin for future software upgrades.
Rivian has been explicit that RAP1 is its first in-house processor built specifically for its autonomy roadmap, a shift that moves the company away from generic compute toward silicon tuned for its own models and vehicles. Earlier coverage of the launch described how the company introduced the chip as part of a broader next generation autonomy platform, with the ACM3 module serving as the main computer that will sit inside future vehicles and process torrents of sensor data in real time. In that context, the ACM3’s 1600 sparse INT8 TOPS figure is not just a bragging right, it is the technical foundation for the company’s claim that it can scale from driver assistance to higher levels of automated driving without swapping out the hardware.
Breaking from Nvidia to control its own destiny
Rivian’s decision to build RAP1 is also a decision to walk away from Nvidia as the primary supplier of its autonomy compute. Reporting on the launch noted that the company is effectively ditching Nvidia for its next generation automated driving systems, a bold move given how entrenched Nvidia has become in the automotive AI market. Two RAP1 chips are expected to power the new autonomy computer, underscoring how central Rivian believes its own silicon will be to the driving experience.
That break is not just about performance, it is about strategic control. By owning the chip design, Rivian can tune the hardware to its own neural networks, manage its supply chain more tightly and potentially reduce long term costs, even if custom silicon is challenging and expensive to develop. The company has framed this as a way to better align its autonomy roadmap with its vehicle platforms, rather than waiting on a third party’s product cycle. In practice, that means Rivian can decide when to ramp new features, how to allocate compute between perception and planning, and how aggressively to push over-the-air updates without being constrained by a vendor’s firmware and tooling.
From runner-up to direct rival: positioning against Tesla and Nvidia
Rivian’s move into custom AI chips is also a statement about where it sees itself in the autonomy race. One analysis pointed out that Most folks would think of Rivian as a runner-up to Tesla, especially given how Tesla and Elon Musk himself have previously derided lidar and leaned heavily on camera-only systems. By embracing lidar and building its own chip, Rivian is signaling that it is not content to follow Tesla’s playbook and instead wants to differentiate on both hardware and software.
Another report framed the announcement even more directly as an attempt by Rivian to replace Nvidia and challenge Tesla in the autonomy race, describing how EV startup Rivian Auto is trying to compete with big shots including Tesla by unveiling its own proprietary silicon chip. That framing captures the dual front Rivian is opening: it is no longer just a customer of Nvidia’s automotive platforms, and it is no longer content to let Tesla define what cutting edge driver assistance looks like.
Sensor strategy: lidar, cameras and the R2 roadmap
Rivian’s autonomy push is not just about compute, it is also about the sensors feeding that compute. The company has said it will add lidar sensors to its upcoming R2 SUV, a decision that reflects the research methods and sensor mix it believes are necessary for robust automated driving. Coverage of its autonomy event described how Rivian developed its own artificial intelligence computer chip and said it will add lidar sensors to upcoming R2 SUV models, tying the sensor upgrade directly to the new AI hardware that will process the data.
Additional reporting on the roadmap highlighted how the company is building a next generation autonomy platform that explicitly includes a lidar roadmap for the R2 line. One detailed breakdown introduced readers to the new compute as “Meet the Rivian Autonomy Processor” and emphasized that One of the biggest announcements was RAP1, which is being built specifically for its autonomy roadmap and paired with a universal hands free system and R2 lidar plans. Taken together, those details show that Rivian is designing its sensor suite and compute as a single system, rather than bolting lidar onto an architecture that was never meant to handle its data rates.
From driver assistance to robotaxis and subscriptions
Rivian is not shy about where it wants this technology to lead: toward both advanced driver assistance in personally owned vehicles and a future robotaxi service. One report on the company’s AI and EV strategy noted that Rivian’s announcements included a proprietary chip, RAP1, designed for “self-driving personally owned vehicles”, underscoring that the first wave of deployment is meant to enhance the experience of individual drivers rather than fleets. That same coverage made clear that Wall Street was impressed by the AI and autonomy story, even if concerns about EV demand and capital needs remained.
At the same time, the company has been explicit that Powering Rivian self-driving aspirations will be a new in-house chip set to launch in 2026, with executives outlining how that hardware will underpin both consumer features and a potential robotaxi business. A separate account of the event described how Powering Rivian self-driving aspirations will be a new in-house chip that is integrated into a fresh electrical architecture for future vehicles, a necessary step if the company wants to support the redundancy and safety requirements of commercial autonomous services.
Business model: pricing autonomy as a service
Rivian is also sketching out how it plans to make money from its autonomy stack, and the early signals point toward a mix of one time purchases and subscriptions. One detailed report on the company’s plans stated that the service will offer autonomous driving capabilities and will be available either as a one-time purchase of $2,500 or as a recurring subscription, giving Rivian flexibility to appeal to both upfront buyers and those who prefer lower monthly costs. That pricing structure mirrors how other automakers have approached advanced driver assistance, but Rivian’s custom silicon gives it more control over the feature set it can bundle at each tier.
Behind that pricing is a broader strategy to turn autonomy and software into recurring revenue streams. Multiple accounts of the event emphasized that Rivian Automotive Inc (NASDAQ, RIVN) has developed its own chip and artificial intelligence models to power future self driving and software driven services, positioning autonomy not just as a safety feature but as a core part of its financial story. By controlling both the hardware and the models, Rivian can iterate on features, gate them behind different price points and potentially expand into fleet offerings without renegotiating with a third party chip vendor.
Deep AI integration: training, cloud and the Autonomy Platform
Underneath the hardware and pricing, Rivian is building what it calls an Autonomy Platform that ties together in-vehicle compute, cloud training and data from its growing fleet. The company has described how it is preparing to expand its Rivian Autonomy Platform, with Training for the models using AWS so that the system can learn from vast datasets and then push improvements back to vehicles. That cloud loop is essential if Rivian wants its AI to keep improving long after a truck or SUV leaves the factory.
The company has also emphasized that its autonomy stack is designed to learn from each vehicle, rather than relying solely on synthetic data or limited test fleets. Official materials on the launch stressed that the next generation platform is built for deep AI integration, with the ACM3 and RAP1 hardware acting as the on board execution engine for models that are constantly refined in the cloud. By anchoring training on AWS and deploying through its own Autonomy Platform, Rivian is trying to create a virtuous cycle where every mile driven by a customer feeds back into better perception, prediction and planning for the entire fleet.
AI assistant and in-cabin intelligence
Rivian’s AI ambitions are not limited to what happens outside the vehicle. The company is also rolling out an AI assistant that will live inside its EVs, handling everything from navigation queries to vehicle settings and potentially even autonomy related explanations. Reporting on the feature noted that Rivian’s AI assistant is coming to its EVs in early 2026, giving the company a relatively near term way to showcase its AI capabilities to existing owners even before full autonomy features roll out.
That assistant will sit alongside the autonomy stack, potentially serving as the interface that explains what the vehicle is doing and why, or that lets drivers customize how assertive the system should be. By tying the assistant into the same cloud infrastructure that trains its driving models, Rivian can keep the in cabin experience evolving as quickly as the on road behavior. It also gives the company another software surface where it can differentiate from rivals, especially if the assistant can tap into the same perception data that RAP1 and ACM3 are using to understand the world outside the windshield.
Investor reaction and the Volkswagen connection
Financial markets have responded to Rivian’s autonomy push with a mix of enthusiasm and caution. One account of the event noted that Rivian‘s AI and autonomy announcements impressed Wall Street analysts, but that EV and capital concerns remain, a reminder that custom silicon and robotaxi dreams do not erase the hard math of building factories and selling trucks. Shares of Rivian Automotive Inc have been volatile around the news, reflecting both excitement about the technology and questions about how quickly it can translate into revenue.
Rivian is not pursuing this strategy alone. The company is a key partner of Volkswagen, and coverage of the autonomy launch highlighted how Rivian and Volkswagen are aligned around the new RAP1 silicon chip, which is designed to process large volumes of data in real time. That partnership could give Rivian additional scale for its autonomy platform, while giving Volkswagen access to a modern EV and AI stack without having to build every component from scratch.
Inside the Autonomy AI Day and Rivian’s pitch
The company used its first dedicated autonomy event in Palo Alto, California to knit these threads together into a single narrative. One account of the gathering described how We are experts in medium event in Palo Alto, California captured the company’s attempt to present itself as a serious player in both hardware and AI, not just an EV startup with stylish trucks. Executives walked through the custom chip, the ACM3 module, the lidar heavy sensor suite and the subscription business model, tying them back to a vision of safer, more convenient driving.
Another detailed breakdown of the day emphasized how Rivian framed RAP1 as the heart of a universal hands free system that would eventually span its lineup, starting with the R2 crossover. By anchoring the story in concrete hardware and near term features, rather than distant Level 4 robotaxis alone, Rivian tried to reassure investors that its autonomy bet can deliver incremental value along the way, even as it chases a longer term vision of self driving fleets.
More from MorningOverview