Rivian Unveils Custom Silicon, R2 Lidar Roadmap, and Universal Hands Free

Original Article ↗ Hacker News Discussion ↗

Custom Silicon & Autonomy Strategy

  • Many see Rivian’s custom chip as a bold but risky “build vs buy” move given the cost, lead time, and Rivian’s ongoing cash burn.
  • Supporters argue:
    • Automotive-grade, power‑efficient compute with long lifetimes is a niche COTS doesn’t fully satisfy.
    • Owning the stack (chips + software) could become a lucrative B2B platform, especially after the VW deal.
  • Skeptics counter:
    • Nvidia, Qualcomm, etc. already sell strong automotive silicon (Orin, Thor).
    • Volume, iteration speed, and tailoring from those vendors may beat a bespoke ASIC economically.
    • This looks to some like an “AI hype” side quest instead of focusing on getting $40k R2s out profitably.

Rivian Software Quality

  • Owners report sharply mixed experiences: some say their trucks are rock-solid; others describe severe bugs (doors not opening, UI misfires, unusable mobile app).
  • The fact that a major legacy OEM paid billions for this stack is viewed as either a huge validation or an indictment of how bad incumbent software must be.

Lidar: Capability & Safety Debates

  • Thread dives deep into lidar types:
    • 905 nm vs 1550 nm wavelengths, camera damage vs eye safety, and differences in how eyes vs lenses interact with IR.
  • Consensus: automotive lidars are designed as Class 1 (eye‑safe in normal use), but edge cases (e.g., very close exposure, many concurrent lidars, device failures) are not fully understood and certification transparency is limited.
  • Some worry about cumulative exposure (humans, animals, insects); others think it’s minor compared to sunlight and existing risks.

Market Reaction & Business Model

  • Commenters note Rivian’s stock dropped on the announcement; proposed reasons:
    • Market distrust that Rivian can execute custom silicon and Gen3 autonomy while still ramping R2.
    • Fear that current and near‑term vehicles are now implicitly “obsolete.”
  • Many expect autonomy to be sold as a subscription, likely bundled with insurance, citing:
    • Ongoing software/ops costs and liability.
    • Waymo data suggesting lower injury rates, creating room to capture insurance savings.
  • Others push back, preferring “you get what you buy” with no ongoing fees and warning about “subscription to life” dynamics.

CarPlay / Android Auto & Affordability

  • A large contingent says the main things they want from Rivian are:
    • CarPlay/Android Auto.
    • Lower prices.
  • Several would have bought a Rivian but instead chose other EVs largely due to CarPlay and better lease economics.
  • Rivian’s stated rationale for rejecting CarPlay (a fully integrated, consistent in‑house UX) is widely seen as control/lock‑in; some accept it if the native UX is good, others say it’s a deal breaker.

Autonomy Tech Landscape: Tesla, Waymo, Rivian

  • Strong disagreement over whether lidar‑heavy approaches (Waymo, Rivian roadmap) or camera‑only (Tesla) is the right bet.
  • Pro‑Waymo/lidar side:
    • Waymo already runs fully driverless paid rides in multiple cities; Tesla FSD still requires supervision.
    • Lidar simplifies depth, object detection, and robustness (night, fog, long tail scenarios).
  • Pro‑Tesla/camera side:
    • Remaining failures are mostly planning, not perception; if cameras solve depth well enough, lidar just adds cost/complexity.
    • Tesla’s vertically integrated, software‑defined architecture and scale give it better economics.
  • Some suggest Rivian’s best‑case niche is as a licensable autonomy platform for other OEMs if camera‑only stumbles and Waymo is viewed as too dominant or too “Google.”

Ownership vs Robotaxis vs Transit

  • One large subthread argues that many “reasons for autonomy” (skip driving, sleep, work en route, safer roads) are better solved by robust public transit, rail, biking, and fewer cars overall.
  • Others strongly prefer private vehicles as:
    • Mobile storage, private space, pet‑ and kid‑friendly, road‑trip and off‑road capable.
    • More convenient than Waymo/Uber, especially outside dense cores.
  • There’s concern that widespread autonomy could increase VMT, sprawl, energy use, and tire pollution, even as crash rates fall.

Insurance, Liability & Safety Engineering

  • Many expect autonomy and insurance to converge: OEM‑provided coverage tied to use of the self‑driving stack.
  • Questions raised:
    • How to handle catastrophic software failures affecting many vehicles at once?
    • Who is liable when autonomy fails (OEM vs driver), especially at L3+?
  • Some criticize over‑the‑air updates for safety‑critical systems (Tesla cited) and note that traditional automotive functional safety standards (ASIL, etc.) and regulatory evidence are still sparse in public.

Competition & Geopolitics

  • Multiple comments argue Rivian (and other US EV makers) survive partly due to US protectionism; Chinese OEMs are said to offer better‑specced, cheaper EVs with strong ADAS already.
  • Debate on whether Rivian’s R2/R3 can compete on price and features in Europe once BYD/Xiaomi and others expand further.