Rivian Autonomy & AI Day! Lidar, Hands-Free Driving, Rivian Assistant & more announced

[Eric9610]

I'd rather they offer it as an add-on instead of baking it into the cost of the vehicle. I'll never use it, so why would I want to pay for it?

I think you missed the point, on a Tesla I am not talking about FSD. This is the basic autopilot that G2 will have to pay $50 for. LOL

Autonomy + on a G2 car will work like regular autopilot not FSD. That will be restricted to G3 models.. Clearly it went over your heads.

Sponsored

 

[TexasBob]

No bounce from this yet:

No value. Table stakes. Build vs buy

 

[krb1183]

is this Rivian Agent only going to be on R2?

 

[captainjp]

is this Rivian Agent only going to be on R2?

Negative. All three Gens
Gen 1 owners, rejoice!

 

[Dark-Fx]

is this Rivian Agent only going to be on R2?

All Rivians.

 

[Administrator]

Added to first post -- official information press release:


Rivian Unveils Custom Silicon, Next-Gen Autonomy Platform, and Deep AI Integration at Inaugural Autonomy & AI Day

PALO ALTO, Calif. – December 11, 2025 – Rivian (NASDAQ: RIVN) today announced significant breakthroughs in vertically integrated automotive technology at its first Autonomy & AI Day. During the event at its Palo Alto offices, the company unveiled its proprietary, purpose-built silicon, outlined its roadmap for next-generation vehicle autonomy, and introduced an evolved software architecture underpinned by artificial intelligence (AI).

“I couldn’t be more excited for the work our teams are driving in autonomy and AI. Our updated hardware platform, which includes our in-house 1600 sparse TOPS inference chip, will enable us to achieve dramatic progress in self-driving to ultimately deliver on our goal of delivering L4. This represents an inflection point for the ownership experience – ultimately being able to give customers their time back when in the car,” said Rivian Founder and CEO RJ Scaringe.

Rivian Autonomy Processor and Gen 3 Autonomy Computer

At the core of Rivian’s technology roadmap is the transition to in-house silicon, designed specifically for the vision-centric physical AI. The first generation Rivian Autonomy Processor (RAP1) is a custom 5nm processor that integrates processing and memory onto a single multi-chip module. This architecture delivers advanced levels of efficiency, performance, and Automotive Safety Integrity Level compliance.

RAP1 powers the company’s third-generation Autonomy computer, the Autonomy Compute Module 3 (ACM3). Key specifications of the ACM3 include:

• 1600 sparse INT8 TOPS (Trillion Operations Per Second).
• The processing power of 5 billion pixels per second.
• RAP1 features RivLink, a low latency interconnect technology allowing chips to be connected to multiply processing power, making it inherently extensible.
• RAP1 is enabled by an in-house developed AI compiler and platform software.

In addition to ACM3, Rivian plans to integrate LiDAR into future R2 models. LiDAR will augment the company’s multi-modal sensor strategy, providing detailed, three-dimensional spatial data and redundant sensing, and improving real-time detection for the edge cases of driving.

Our Gen 3 Autonomy hardware including ACM3 and LiDAR is currently undergoing validation and we expect it to ship on R2 models starting at the end of 2026.

Rivian Autonomy Platform

Rivian also detailed its software-first approach to autonomy, powered by the Rivian Autonomy Platform and an end-to-end data loop used for training. The company introduced its Large Driving Model (LDM), a foundational autonomous model trained like a Large Language Model (LLM). Utilizing Group-Relative Policy Optimization (GRPO), the LDM will distill superior driving strategies from massive datasets into the vehicle.

Software advancements are coming to the company’s second generation R1 vehicles in the near term with the addition of Universal Hands-Free (UHF), bringing hands-free assisted driving for extended periods to significantly more places.

• Available on over 3.5 million miles of roads across the USA and Canada.
• Capable of operating off-highway on roads with clearly painted lines.

Rivian announced its autonomy subscription, Autonomy+, with continuously expanding capabilities, launching in early 2026 and priced at $2,500 (one-time) or $49.99 (per month). These features have the potential to make the roads safer, address customer demand and become veritable drivers for the business.

Rivian also detailed plans to continuously improve the autonomy capabilities of its Gen 2 R1 and future R2 vehicles, with a clear trajectory including point-to-point, eyes off and personal L4.

Rivian Unified Intelligence and Rivian Assistant

Beyond vehicle autonomy, Rivian is harnessing AI across the business with the Rivian Unified Intelligence (RUI). The shared, multi-modal and multi-LLM data foundation is helping the company develop powerful new features, improve service infrastructure, and even power predictive maintenance. The highlight of this new architecture is the Rivian Assistant, a next-generation voice interface launching in early 2026 on Gen 1 and Gen 2 R1 vehicles.

• Built on Rivian’s edge models to understand your vehicle, your digital life and the world around you.
• Connects vehicle systems with third-party apps using an in-house agentic framework, with Google Calendar named as the first integration.
• Augmented by frontier large language models for grounded data, natural conversation and powerful reasoning.

RUI will fundamentally change service by embedding AI into diagnostics. It’s an expert assistant for technicians, scanning telemetry and history to pinpointing complex issues. This same advanced intelligence will soon power Rivian’s mobile app, improving self-service diagnostics.

Rivian’s unique, vertically integrated strategy allows the company to rapidly evolve the entire vehicle experience—from user-facing features to foundational technology. In the company’s next stage of growth, integration meets acceleration as this next-generation hardware and software come to market.

 

[ukyank]

No bounce from this yet:

Of course not, the only thing that continues to matter for the company’s viability is the sales of the R2, everything else is noise. Good sales = likely company succeeds; poor sales = likely the company folds

 

[mkhuffman]

No value. Table stakes. Build vs buy

Maybe, but they are demonstrating a new way for the company to make money, and improve their chances of success in the automobile market.

Not only will they be one of the leaders in the industry, they will be able to do it inexpensively (relatively). The legacy car companies are not able to match this cost structure. I suspect only Tesla can. (And maybe the Chinese start-ups, but they cheat and lie about everything so who knows what they really can do at scale.)

$50 a month is a good revenue stream.

 

[Dave Cundiff]

I want the in-vehicle interface they showed, a whole lot more than I want CarPlay!

In my experience, CarPlay gets in the way of the car's in-house features. This will use them instead.

Good moves, Rivian!

 

[captainjp]

Was refreshing to watch a CEO that can actually speak in full sentences…

 

[TexasBob]

Very solid tech stack. Here is a summary from Gemini of its competitive positioning fwiw.

Here is the summary of the Rivian Autonomy Stack and where it stands competitively.

1. The Rivian Autonomy Stack (The "Peregrine" Architecture)
A. The Hardware: "Thor-Class" Custom Silicon Rivian isn't buying chips; they built a monster.

• The Chip: RAP1 (Rivian Autonomy Processor 1).
• Specs: 1,600 TOPS (Sparse INT8) per Gen 3 computer (Dual Chip). This is massive.
• The "Killer Feature": The Neural Engine has a dedicated hardware block for "Unstructured Point Cloud Processing".
  • Significance: Unlike Tesla (which forces everything into 2D video), Rivian’s chip eats raw Lidar/Radar data natively. This is hardware-level validation of the "Multimodal" thesis.
• Process: TSMC 5nm Automotive.

B. The Software: "Vision-Language-Action" (VLA) They are explicitly moving to the reasoning architecture

• Architecture: Sensors → Transformer Encoders → World Model → Trajectory.
• Training Tech: They explicitly mention "Tokenized Data" and "GRPO Learning" (Group Relative Policy Optimization).
  • Significance: GRPO is a cutting-edge Reinforcement Learning technique used to train LLMs to "reason." Rivian is applying this to driving, confirming they are treating the car like an agent, not a reflex loop.
• Validation: "Apprentice Mode". This is their "Shadow Mode"—the agent runs in the background, predicting moves and learning from the human driver without taking control.

2. Competitive Analysis

Feature	Rivian (Gen 3)	Tesla (HW4 / AI4)	China (Zeekr/XPeng)	Verdict
Compute	1,600 TOPS	~300–500 TOPS	2,000+ TOPS(Thor/Turing)	Rivian ≈ China > Tesla. Rivian matches the high-end "Thor" standard 18 months early.
Sensors	Lidar + Radar + Cam	Camera Only	Lidar + Radar + Cam	Rivian Wins on Robustness. The slides explicitly show Lidar detecting obstacles in fog where cameras fail.
Architecture	World Model / VLA	End-to-End NN	End-to-End NN	Parity. Everyone is converging on World Models, but Rivian has the compute to run it now.
Scalability	Chiplet (RivLink)	Single SoC	Single/Dual SoC	Rivian Wins. The 128 Gbps interconnect allows them to stack chips easily.

Export to Sheets
3. The "Secret Sauce" Approach
Rivian’s presentation reveals three strategic pivots that separate them from the pack:

1. Hardware-Native Fusion: They didn't just add Lidar; they built a chip that speaks Lidar. This eliminates the processing penalty that usually makes sensor fusion slow.
2. "Agentic" Cabin: They are moving to Edge AI Agents. The slides show a "Small Language Model" (SLM) running on the chip (not in the cloud) to handle voice and vehicle commands. This means your car is smart even when offline.
3. The "Data Recorder" Edge: Their "Autonomy Data Recorder" specifically identifies edge cases (animals, tunnels, weird merges) on the car and uploads only that interesting data. This solves the "Data Deluge" problem—they don't upload boring highway miles, just the confusing stuff to train the World Model.

Summary: The "Prophet" Validation
Rivian has effectively positioned itself as the "Western Waymo"—a consumer brand that owns the full "Chip-to-Wheel" stack with the sensor robustness of a Robotaxi and the business model of a luxury automaker.

 

[renderpaz]

Was refreshing to watch a CEO that can actually speak in full sentences…

And start ON TIME!

 

[Eric9610]

Maybe, but they are demonstrating a new way for the company to make money, and improve their chances of success in the automobile market.

Not only will they be one of the leaders in the industry, they will be able to do it inexpensively (relatively). The legacy car companies are not able to match this cost structure. I suspect only Tesla can. (And maybe the Chinese start-ups, but they cheat and lie about everything so who knows what they really can do at scale.)

$50 a month is a good revenue stream.

Some legacy brands have much better level 3 and 4 tech already and just have not implemented. BMW and MB have had Full hands-off driving for over 10 years, capable of even being driven on race tracks. And many brands have LiDar already implemented. This technology is not new to the industry, just new to Rivian.

Liability is going to be the biggest hurdle.

 

[mkhuffman]

Does the live stream keep dropping for others? My feed has dropped several times.

Mine never dropped. Maybe the network in your area was congested.

 

[R1S88]

I understand being focused on whether your current vehicle will get these capabilities, but let's not bury the lede here. Getting rid of Nvidia and using their own chip with lidar etc. seems like a huge leap. Fully executed, this has to greatly enhance Rivian's long-term viability. Am I wrong to be super excited about this for Rivian?

Can we shift the conversation to discuss what was revealed today and away from sadness about your current vehicle?

Sponsored