---
title: "RIVN (Trans): R2 conversion above expectations; Q4 is the real test"
type: "Topics"
locale: "en"
url: "https://longbridge.com/en/topics/43102111.md"
description: "Below is Dolphin Research's Trans of $Rivian Automotive(RIVN.US) FY26 Q2 earnings call. This section summarizes the core highlights.1) Full-year delivery guide raised: 65k–70k units for 2026 (covering R1, R2 and commercial vans), up 3k vs. prior. It implies 42.4k–47.4k deliveries in H2, with the R2 ramp skewing volume to Q4..."
datetime: "2026-07-31T01:52:02.000Z"
locales:
  - [en](https://longbridge.com/en/topics/43102111.md)
  - [zh-CN](https://longbridge.com/zh-CN/topics/43102111.md)
  - [zh-HK](https://longbridge.com/zh-HK/topics/43102111.md)
author: "[Dolphin Research](https://longbridge.com/en/news/dolphin.md)"
---

# RIVN (Trans): R2 conversion above expectations; Q4 is the real test

**Below is Dolphin Research's takeaways from** $Rivian Automotive(RIVN.US) **FY26 Q2 earnings call**

**I. Key takeaways**

1\. **Full-year deliveries raised**: FY26 deliveries guided to 65k–70k units (R1, R2 and commercial vans), up 3k vs. prior. This implies H2 deliveries of 42.4k–47.4k, with R2 ramp skewed to Q4.

2\. **FY P&L and capex guide**: FY26 Adj. EBITDA loss guided to $1.8bn–$2.0bn, with the low end improved by $0.1bn and the midpoint by $50mn, driven by better-than-expected regulatory credit revenue in Q2 and higher deliveries, partly offset by higher raw materials, memory and logistics. **FY capex cut by $250mn to $1.7bn–$1.8bn, benefiting from project efficiency and spend pacing, with spend focused on Normal plant R2 completion and tooling, sales & service and charging build-out, and the Georgia greenfield plant.**

3\. **Gross profit cadence and full-year trend**: Full-year GP still up YoY, but new model launch complexity will continue to weigh on auto GP in Q3 (similar to Q2). From Q4, as production and deliveries ramp, contribution turns positive; management reiterated R2 is on track for positive exit-rate GPM in 2026, making 2026 a transition year toward long-term auto profitability.

4\. **Q2 key financials**: Consolidated revenue $1.66bn (+27% YoY). GP $179mn with GPM 11% (includes $138mn D&A and $31mn SBC); Adj. EBITDA loss $379mn. Auto revenue $1.14bn (+23% YoY) on +14% delivery growth and +$103mn regulatory credits, partially offset by mix/ASP pressure from higher EDV and R2 mix; auto GP loss $36mn vs. $335mn loss a year ago, improving $299mn YoY. Around $100mn of incremental OpEx/COGS recognized in the quarter due to the R2 ramp (vs. a more normalized production level); unit COGS was $96.7k, or approx. $88.4k ex that $100mn (about $8.0k–$8.2k per unit), an improvement of about $5k QoQ. Software & services revenue $515mn (+37% YoY), of which $308mn (~60%) came from the JV with Volkswagen Group; segment GP $215mn with 42% GPM. Q2 production 12,613 units and deliveries 12,194 units, exceeding the prior 9,000–11,000 guide.

5\. **Capital structure and liquidity**: Quarter-end cash, cash equivalents and ST investments were about $5.3bn. In Jul., the company sold 86.25mn Class A shares via a follow-on, raising about $1.3bn for general corporate purposes, including equity commitments and reserves tied to the $4.5bn U.S. DOE loan for the Georgia plant; expects an additional $1.0bn non-recourse debt from Volkswagen Group and a $250mn incremental equity investment from Uber this year (both subject to conditions). Current liquidity plus targeted future capital exceeds $14.0bn.

**II. Call details**

**2.1 Management remarks**

1\. **R2 launch and ramp**

a. R2 started customer deliveries in Q2 and is viewed as the key growth and profitability driver. In a U.S. market with limited high-quality EV choices, R2 pricing is compelling and media reviews are strong.

b. The Launch Edition is a performance trim priced at $58k, the most expensive R2; Premium and Standard are slated for early 2027. Reservation-to-order conversion is meaningfully above internal expectations.

c. Over 57k test drives were conducted in Q2, a company record at roughly 2x prior levels.

d. R2 began with a single shift, with a second shift targeted by end-Q3; the added shift will not materially contribute volume in Q3, with impact starting in Q4. The biggest focus in the ramp is the supply chain, with Rivian teams embedded on-site at select suppliers.

e. Versus R1, R2 adopted structured Design Validation Builds (DVB) and Manufacturing Validation Builds (MVB), with MVB starting in early 2026; ramp surprises are far fewer than R1. Configurations are also greatly simplified (all within the Launch Edition with limited color options), whereas R1 launched R1T, R1S and EDV in the same three-month window with thousands of possible configurations.

2\. **Manufacturing footprint**

a. The Normal plant is focused on the R2 ramp; the Georgia plant is progressing well.

b. Combined, Normal and Georgia can support 515k units of annual capacity, with ample room to expand.

3\. **Commercial vans and Amazon (EDV)**

a. Over 40k Rivian EDVs are in operation at Amazon, covering thousands of cities in North America; larger packs and AWD variants are in development to support Amazon demand.

b. The commercial van platform surpassed 1bn miles driven in Q2.

c. Management **believes Amazon's accelerated purchases validate EDV's TCO advantage, though broader fleet adoption of electrified, highly connected vehicles is slower than expected.**

4\. **Autonomy and in-house silicon**

a. **Autonomous driving is one of Rivian's top investment areas, with advanced driver-assistance seen as a key differentiator and share driver.**

b. Roadmap: **point-to-point capability by year-end (Level 2++, hands-off but eyes-on); Level 3 (hands-off, eyes-off) in 2027; Level 4 in 2028 for both consumer and Robotaxi R2 variants.**

c. Autonomy+ has been monetized since Apr., with improving attach trends; another Autonomy & AI Day is planned this year.

d. **Targeting late-2026 rollout of Gen-3 autonomy hardware, including the first in-house chip RAP1 and added LiDAR; RAP1 is on plan and in final mass-production testing.**

e. The technical path is an end-to-end trained neural network, the LDM (large driving model), using one model and workflow from point-to-point to Level 4, with only sensor topology and inference capacity differing.

f. **Robotaxi collaboration with Uber is on track, with year-end as the next milestone, and a plan to deploy in select cities in 2028 and scale rapidly thereafter.**

5\. **Software and AI experience**

a. In May, Rivian launched Rivian Assistant (AI voice assistant integrated with in-car ecosystem and apps like Spotify and Google) across all R1s; R2 will get it via OTA later this year.

b. The software & services segment delivered strong growth, driven by E/E architecture and software development services for the JV, maintenance & repairs, and Autonomy+, partially offset by lower used-vehicle resale revenue.

**2.2 Q&A**

**Q: How is the R2 ramp progressing and what quantifiable metrics can you share? Over the next year, will R2 delivery cadence be supply- or demand-constrained?**

A: The first production ramp of any new model is extremely complex and essentially a synchronized effort across hundreds of suppliers, with total output constrained by the slowest ramp. Ensuring all suppliers increase output in a coordinated manner is the core focus and the reason to start with one shift, stabilize, then add the second. Prior experience with R1 and its iterations gave us capabilities we did not have then.

**We are progressing week by week, and the raised guide reflects that. The main watch item remains the supply chain, making sure suppliers are ready for higher H2 output, especially full two-shift operations.**

Demand is very strong, with encouraging reservation-to-order conversions. Even for the $58k Launch Edition, the most expensive R2, conversion is well above expectations, which is a good signal; Premium and Standard still target early 2027. Even at the higher price point, market interest is very high.

**Q: What does the R2 ramp mean for the P&L? With the current spend environment, do you still expect positive GPM exit rate in 2026?**

A: We still **expect R2 to achieve a positive GPM exit rate in 2026, driven by ongoing efficiency gains at higher volumes and fixed-cost absorption.**

**Q: How is EDV commercialization outside Amazon? For non-Amazon pilots, what constrains deployment most — TCO, upfront capex, or charging density?**

A: First, Amazon is scaling internally, as seen in the first two quarters, with purchases well above prior years purely reflecting the EDV platform's TCO advantage. As Amazon demonstrates cost benefits at scale, we believe this will carry over to other fleet operators that need to stay competitive and manage their cost structures, which should ultimately drive EDV opportunities.

That said, broader fleets are adopting electrification and highly connected vehicles more slowly than we expected or hoped. But Amazon's step-up is a powerful signal, essentially a billboard validating EDV's capabilities and commercial value.

**Q: Benchmarking global EV platforms, especially Chinese OEMs, what structural or operational learnings will inform R3 and future models?**

A: Every automaker understands what peers are doing. Rivian vehicles have been torn down by automakers globally, and Chinese models are similarly benchmarked by OEMs and third parties. Manufacturing techniques like joining, high-pressure die casting, and stamped part consolidation are not magically different in China versus the West; they are common to top-tier architectures, and R2 reflects this with heavy use of HPDC and an intense focus on part consolidation and elimination.

**The biggest U.S.–China cost gap is input costs.** Labor is much cheaper in China, and the cost of capital is lower or often zero via local Gov. support, compounding through the supply chain to the OEM and ending in much lower costs than Western production. This raises trade policy questions and complicated supply-chain strategies.

We address this complexity not only at Tier-1s but down to raw materials, considering origin and adhering to existing trade frameworks. In a fully free-trade world, supply chains would purely optimize for cost, concentrating in countries with the lowest labor, energy, land and capital costs; but reality involves deliberate industrial policy that changes how we make decisions, and for certain components, domestic U.S. sourcing is advantageous.

**Q: How to break down the $50mn improvement in the FY Adj. EBITDA guide? How much is from better regulatory credit sales, and how should we think about input-cost inflation?**

A: The EBITDA guide low end improved by $100mn, yielding a $50mn midpoint improvement. H1 regulatory credits were above expectations; looking ahead, higher deliveries (+3k units) add further contribution. Offsets come from increases in raw materials, memory and logistics, resulting in the current Adj. EBITDA range and outlook.

For H2 vs. H1, on GP we noted **Q3 auto GP will still be impacted by R2 ramp inefficiencies and complexity, with tailwinds from scale and volume starting in Q4.** In addition, we will keep investing in R&D, especially autonomy, while sales growth outpaces network expansion, driving ongoing fixed-cost leverage.

**Q: What is Autonomy+ attach and usage today? How will point-to-point this year drive demand and attach, and thus subscription revenue?**

A: Autonomy+ attach so far is encouraging. Notably, customer engagement and feature usage, and the related paid conversion, are strong; with a clear roadmap, we expect attach to rise as capabilities compound.

Think of this as a staircase of features culminating in high-level autonomy. A near-term step is point-to-point: enter an address in-car, and the vehicle drives itself to the destination. Initially this is Level 2++ — hands-off but eyes-on — then in 2027 hands-off and eyes-off; that is critical, as users regain their time without constant attention prompts, which is the true definition of Level 3. In 2028 we introduce Level 4, enabling Robotaxi with no driver in the front seat.

Beyond Robotaxi, 'personal Level 4' on owned vehicles is compelling: your car can take you to the airport, pick up kids, or grab groceries. This roadmap should structurally lift Autonomy+ penetration.

We also think consumer behavior is at an inflection point. Historically, autonomy was not a primary purchase driver for most buyers except early adopters; now it is becoming a key criterion. That is why we are investing heavily in Autonomy+: market share will be decided by vehicles' autonomy capabilities and, relatedly, by their software and AI-defined feature sets — our two strategic pillars.

**Q: What is the early R2 buyer profile — existing EV owners, first-time EV buyers, or cross-shoppers?**

A: R2 was designed to attract a highly diverse customer base across body styles, brands and demographics, and we are seeing that in the data. Customers come from a wide mix of prior vehicles, which is unsurprising given that the mid-size two-row SUV is the largest U.S. segment and often the largest globally.

Our execution on space, storage, performance, range, efficiency and on-/off-road dynamics casts a wide net. One of the most exciting metrics is the sizable number of first-time EV buyers, which is powerful for brand connection; it shows strong product-market fit and that we are activating previously latent demand. We also see switchers from Tesla and R1, but the share of customers who never owned an EV is particularly notable.

**Q: How big was the IEEPA tariff refund benefit in Q2 GP? Also, guidance implies larger H2 Adj. EBITDA losses vs. H1; how does that square with the Q4 positive GP exit-rate commentary?**

A: As noted last quarter, **we sized IEEPA tariffs in the tens of millions, and recognized roughly half of the total as receivables this quarter — the portion not affected by the DOJ appeal.**

On higher H2 Adj. EBITDA losses, a key difference is the $164mn regulatory credit revenue that benefited H1 GP. In H2, Q3 still sees R2 ramp complexity in auto GP, with Q4 improving; more importantly, H2 lacks that regulatory credit tailwind and includes continued R&D and SG&A investment.

**Q: With 57k test drives, how are test-drive-to-order conversions trending? How many experienced Autonomy+? Will deliveries this year skew to higher trims or mainly performance (Launch Edition)?**

A: The 57k test drives set a Rivian record and required intensive preparation to support roughly 2x prior volume. We anticipated R2 would sharply increase brand exposure and customer touchpoints, primarily via test drives.

We do not disclose conversion rates externally, though it is a key internal metric and closely linked to availability of other trims. Today we are offering the performance Launch Edition; Premium and Standard arrive in early 2027. Launch Edition conversion is clearly above expectations, reflecting strong product pull, while we recognize and look forward to offering the other trims as planned.

**Q: R&D was roughly flat QoQ in Q2. Given heavy autonomy investment, how should we think about H2 R&D?**

A: **R&D is expected to step up in H2, primarily driven by GPU purchases for autonomy training, with spend ramping through H2.** In Q2, engineering, design and development moderated as we entered production, creating a sequential dip, but autonomy spend within R&D is still rising significantly.

**Q: The ~$100mn incremental operating cost tied to R2 ramp vs. a 'normalized' level — what does normalized mean? One shift or two? Does it assume a more normal mix?**

A: The $100mn reflects a bundle of ramp inefficiencies such as expedited freight and supplier premiums that fade as volume scales over coming quarters. It also includes costs that would ordinarily be absorbed into inventory but were expensed this quarter due to not yet reaching the Q4 run-rate level. That is how we arrive at the ~$100mn estimate.

**Q: So a meaningful portion is lack of scale and asset utilization; once a second shift comes online and GP exits positive in Q4, we can back into a near breakeven volume — is that fair?**

A: Directionally, that is a reasonable way to think about it. To clarify, 'Q4 normalized level' does not equal fully ramped R2 capacity, but rather a nearer-term, more normal production environment suitable for comparing to standard costs.

**Q: Update on RAP1 and Gen-3 autonomy hardware — still on track to SOP this year?**

A: RAP1 is progressing to plan as the backbone of our Gen-3 autonomy hardware platform. RAP1 and the rest of the system upgrades are on track to land this year.

**Q: Despite the $100mn incremental ramp cost, unit COGS only rose by several thousand dollars — does that imply other cost improvements in EDV or R1? Why didn't raw-material pressure show up more in per-unit cost?**

A: Q2 unit COGS was $96.7k. Ex the $100mn, at roughly $8.0k–$8.2k per unit, unit COGS would be about $88.4k, an improvement of about $5k QoQ; a key driver was higher volume. As the year progresses, this remains a core theme with unit COGS continuing to trend down, and we also benefited from IEEPA receivables in the quarter.

**Q: Was the IEEPA benefit booked in COGS?**

A: Yes, in COGS.

**Q: Looking back at the early R1 ramp, the company acknowledged it tried to do too much. Now with RAP, autonomy, and network architecture all in motion, how is R2 execution different, and how are resources allocated?**

A: R1 feels both recent and long ago, but the company launching R2 is very different in process maturity, team depth and development sequencing. What we are experiencing in the ramp and supplier scale-up is the product of the past 2–3 years of work.

A major difference is our structured DVB and MVB approach — building to production spec on pilot lines, then to production spec in the factory — with MVB starting in early 2026, giving time to resolve ramp, design and supplier issues ahead of SOP. As a result, R2 ramp surprises are far fewer than R1.

That does not mean zero surprises; launching a new model is highly complex for any automaker. We have prepared extensively and will remain cautious on supplier pacing, ensuring synchronized ramp and meticulous planning before adding the second shift.

Another enabler is far simpler product complexity. R1 launched R1T, R1S and a commercial van within the same three months with thousands of configurations, whereas R2's buildable combinations are very limited within the Launch Edition, even with restricted color choices. This was deliberate to enable a smoother, faster ramp, and the feature set included in Launch Edition reflects considerable thought.

We had internal expectations for Launch Edition conversion and the actual rates are clearly above them, which is good news. That does not mean we are not launching other trims; customers waiting for Standard or Premium should not worry — those arrive in early 2027.

**Q: With point-to-point and eyes-off in 2027, how do you think about pricing and ARPU uplift for autonomy, and how will pricing evolve?**

A: We do see pricing upside as capability expands, but we believe it must be earned by delivering a broader feature set and performance.

Another factor is market dynamics — pricing for these features may be fluid. As noted, the industry is at an inflection: advanced autonomy is shifting from a nice-to-have to a must-have, and understanding is improving. Transitioning to Level 3 makes capability easier to explain and grasp for mass-market customers, reducing explanation costs and enabling adoption beyond early adopters, which should in turn support higher pricing than today.

We continue to evaluate the ultimate pricing headroom, but we see opportunity to move above current levels over time.

**Q: California's new EV subsidy program requires OEM co-funding. Will Rivian participate and have you seen incremental demand?**

A: We welcome California's EV incentives, which align with our mission, and we look forward to bringing more first-time EV buyers into the Rivian community. The program provides $3,500 for new vehicles and $1,750 for used, co-funded by OEMs and the state; there are OEM caps, so customer eligibility will be limited.

**Q: On the flip side of conversion, what are the main reasons some customers did not convert — wait times, or waiting for other trims?**

A: There is a long list, but the most common reason is waiting for a configuration different from the Launch Edition package, by far the top driver.

Because conversion is above plan, we track it dynamically and closely to provide delivery windows to Launch Edition customers, which requires forecasting conversion when the backlog is ultimately digested. Our guidance incorporates these dynamics, and deliveries of other trims begin in 2027.

**Q: You may host another Autonomy & AI Day this year. It has been a while since IPO-era long-term targets — will you update LT financial goals for investors?**

A: We are evaluating the right timing over the next few quarters to host a broader investor day focused on long-term financials.

For Autonomy & AI Day itself, last year's event disclosed long-running internal development including our in-house chip and the Level 4 roadmap; since then we announced the Uber Robotaxi partnership. We want to provide more detail on the path to Level 4 and **show technical progress beyond what current customers see in production vehicles**, improving visibility and confidence around our targets.

**Q: Considering frictions between Uber and Waymo and uncertainty around Lucid's financial stability and its work with Nuro, how much headroom remains for Rivian-Uber?**

A: A core element of our approach is vertical integration of vehicle and tech stack to deliver cost advantages and technical simplification in vehicle architecture and autonomy hardware — including sensor integration, redundancy in control systems, and compute/inference architecture. With in-house silicon, we can deliver high-performance inference at very low in-vehicle cost, providing a strong cost position for Level 4.

We also use an end-to-end, neural network-based LDM (large driving model). Progress on the consumer platform — point-to-point, hands-off, eyes-off, Level 3 to Level 4 — shares the same LDM workflow. It is a single workflow rather than parallel L2 and L4 stacks; one LDM, one end-to-end model.

Vehicles may have different sensor topologies to handle edge cases (e.g., L3 vs. L4) and inference capacity around RAP1 may vary, but the platform and model are the same. This is a pivotal shift for us and the industry. Historically (pre-2022/2023), L4 and L2 tech topologies were very different; the move to Transformer-based encoders and end-to-end models is driving convergence, benefiting our architecture and the data flywheel from the deployed fleet.

**Q: Specifically on Uber, at announcement Rivian was one of multiple partners. Can Rivian become a larger autonomy partner for Uber?**

A: We have a very strong working relationship with Uber, from top leadership to execution teams. Senior-level engagement and execution alignment have been very positive.

**We are working toward year-end milestones and a 2028 deployment in select cities, followed by rapid expansion.** We are encouraged by the partnership, but our focus remains on ensuring the tech is ready, vehicles can deliver L4 performance, and doing so with clean vehicle-level execution and compelling unit economics.

**Q: How many R2s were delivered in Q2? The ramp headwinds you warned about did not materially show up in margins (helped by IEEPA and ZEV credits), yet you flag Q3 pressure — was Q2 ramp smaller, or smoother than expected? Could Q3 margins surprise to the upside?**

A: In Q2, we transitioned from pre-production to SOP, with the cost structure impacting OpEx/COGS for about two-thirds of the quarter, not a full-quarter effect. In Q3, we will absorb a full quarter of ramp effects and add the second shift, bringing additional labor costs. In Q4, we should start to see scaled-production benefits and fixed-cost absorption at the R2 and plant level.

On manufacturing, the Normal team is laser-focused on ramping the first shift, work that started in Q2 and continues in Q3. We do not expect the second shift to add material volume in Q3, but we should see it in Q4. Compared with R1 launch, the team's capabilities have improved markedly; the winning teams are those that can solve issues quickly. **We are focused on training line operators, commissioning automation, and raising output through the ramp. We are also focused on the supply chain**, historically a critical area; we have embedded at a handful of suppliers with good collaboration, and suppliers are receptive to support.

<End of text\>

**Risk disclosure and disclaimer:**[**Dolphin Research Disclaimer and General Disclosure**](https://support.longbridge.global/topics/misc/dolphin-disclaimer)

### Related Stocks

- [UBER.US](https://longbridge.com/en/quote/UBER.US.md)
- [RIVN.US](https://longbridge.com/en/quote/RIVN.US.md)
- [VOW.DE](https://longbridge.com/en/quote/VOW.DE.md)
- [AMZN.US](https://longbridge.com/en/quote/AMZN.US.md)

## Comments (1)

- **過客 · 2026-07-31T04:06:02.000Z**: Robotics in Hong Kong