--- title: "LI: Auto biz 'hurting' — counting on robots as the 'lifeline'?---" type: "Topics" locale: "en" url: "https://longbridge.com/en/topics/39223321.md" description: "Li Auto has effectively entered a loss-making phase. The company appears to be on a loss trajectory." datetime: "2026-03-12T13:02:47.000Z" locales: - [en](https://longbridge.com/en/topics/39223321.md) - [zh-CN](https://longbridge.com/zh-CN/topics/39223321.md) - [zh-HK](https://longbridge.com/zh-HK/topics/39223321.md) author: "[Dolphin Research](https://longbridge.com/en/news/dolphin.md)" --- > Supported Languages: [简体中文](https://longbridge.com/zh-CN/topics/39223321.md) | [繁體中文](https://longbridge.com/zh-HK/topics/39223321.md) # LI: Auto biz 'hurting' — counting on robots as the 'lifeline'?--- $Li Auto(LI.US) released its 2025 Q4 results after Hong Kong trading and before U.S. market open on Mar 12 Beijing time. Li Auto had already guided for pressure on Q4 vehicle GPM, but even against a cautious setup, the print was mediocre. Details below: **① Revenue slightly missed, driven by a larger drop in ASP:** Vehicle revenue was RMB 27.3bn (-36% YoY), below the street at RMB 28.2bn, as ASP fell QoQ by RMB 27k to RMB 250k (vs. cons. RMB 258k). Mix continued to drift downmarket (higher i6 mix, lower Mega mix), and bigger discounts on legacy L-series further pulled ASP lower. **② Vehicle GPM broadly in line with guidance and expectations:** Q4 vehicle GPM was 16.8%, down 300bps from Q3’s underlying 19.8% ex-Mega recall, and within management’s 15.8%-16.8% guide. The decline was mainly ASP-driven. Despite strong i6 demand, supply-chain constraints (notably batteries) limited scale benefits, with total deliveries up only 17% QoQ to 109k units, which was not enough to offset lower pricing. **③ OP turned negative; net profit propped up by financial income:** Q4 OP swung to a factual loss of ~RMB 400mn, vs. near breakeven in Q3 on an underlying basis (-RMB 70mn ex-Mega). With Q4 deliveries -31% YoY and vehicle GPM below the 20% ‘healthy’ line (at 16.8%) while R&D did not ease, net profit fell sharply. This marks the most precarious OP position since 2023, underscoring heavy operating pressure. **Headline net profit stayed marginally positive, but not due to core biz improvement; it relied heavily on ~RMB 430mn of interest income from the company’s large cash balance, masking underlying weakness. Ex this non-operating financial income, Li Auto would have been in the red.** Overall, despite prior signposting, both ASP and vehicle GPM declined QoQ, and net profit is hovering near breakeven. Operating pressure remains significant. The Q4 ramp from the new all-electric i6 could not offset the steep decline in the L-series base (L-series deliveries -60% YoY; total Q4 deliveries -31% YoY). The low-price strategy for EVs dragged overall vehicle GPM (i6 GPM <15% when monthly sales are below 20k) and is also cannibalizing the L-series base. **On the outlook that matters most to the market:** **① Volume guide broadly in line:** Q1 deliveries are guided at 85k-90k, roughly in line with the street at 88k. This implies Mar deliveries of 31k-36k (vs. Jan/Feb at 28k/26k), driven by gradual i6 capacity release in Mar (management previously expected i6 to reach 20k/month after CNY). But legacy L-series volume pressure remains heavy: L-series sold only 18k units in Jan-Feb 2026 (9k/month), down 67% YoY from 55k. Aging models, intensified competition (rivals launching larger and cheaper L-series competitors), and cannibalization from Li’s i6+i8 persist; thus Q1 guided deliveries are still 3%-9% below 2025 Q1’s 93k. **② Revenue guide well below expectations, implying further ASP erosion:** With volume roughly in line, revenue guidance is far short of consensus, as Q1 2026 ASP is set to fall further to ~RMB 222k from RMB 250k in Q4 2025. Dolphin Research believes the sharp QoQ ASP drop is mainly due to: a) i6 mix rising as capacity ramps, with the i6 share moving from 26% in Q4 2025 to ~60% in Q1 2026; b) deeper discounts to clear legacy models ahead of the new product cycle starting in Q2. Given the lower ASP and still tepid volumes in Q1 (deliveries -17% to -22% QoQ vs. Q4 2025, lifting per-unit fixed-cost burden and diluting scale benefits), plus recent upstream price increases in lithium batteries and memory chips pushing BOM costs higher, Dolphin Research expects vehicle GPM to trend down QoQ in Q1 and reach an operating trough (vs. the market looking for flat GPM vs. Q4’s 16.8%). **Looking past the Q1 trough to a full-year 2026 view:** **① Major L-series refresh plus a new all-electric i9** Starting in Q2 2026, Li Auto will roll out three major L-series refreshes, beginning with the L9 in Q2. The L9 price band is expected to shift down to the former L8’s RMB 320k-380k range (vs. the flagship L9 LIVIS currently at RMB 559.8k), followed by L7 and L6. From the announced L9 updates, Li is delivering a spec bump at competitive pricing: a) larger dimensions; b) much bigger battery and EV range, with a new range-extender system to cut fuel use; c) in-house Mach chips with up to 2,560 TOPS; d) upgraded by-wire chassis (incl. 4WS) and mechanical braking. The 2026 playbook appears to be: a) more features for less, moving L9 pricing into the L8 band; b) streamlined SKUs with near full-spec at entry, to defend the vulnerable L-series base as the market pivots aggressively to next-gen range-extended platforms. **② Continued commitment to AI investment:** a) In-house compute platform nearing rollout: the 5nm in-house Mach 100 (M100) chip is close to mass production, delivering up to 2,560 TOPS with a dual-chip per vehicle. It will debut on the new flagship L9 Livis and is planned to enable VLA foundation models as standard across the lineup in 2026. b) Embodied intelligence: extending R&D into ‘spatial robots’, integrating foundation models, in-house chips, and a robotics OS. Rumor has it the first two-wheeled embodied robot could debut in H1 this year. Based on the current share price, the market implies 2026 deliveries of 500k units, up 23% from 406k in 2024, mainly driven by refreshed L-series and a full-year i6 contribution. On that volume, Dolphin Research estimates Li Auto is trading at ~1x 2026E P/S. Despite a lackluster print and a weak revenue guide, supportive factors include: a) a very strong cash buffer, with cash of RMB 101.2bn (net cash RMB 91.7bn); b) a new model cycle starting in Q2 (L9 refresh first); c) a potential debut in embodied intelligence. These could underpin the stock near term and offer some upside optionality at a reasonable valuation. Whether the shares can stage a stronger reversal still hinges on the Q2 ‘more-for-less’ major refresh and its ability to stabilize the fragile L-series base amid an increasingly competitive range-extended market. **Detailed analysis follows:** With deliveries already disclosed, the key incremental items are: **1) Q4 vehicle GPM; 2) Q1 2026 outlook.** **I. Underlying vehicle GPM falls below the 20% ‘healthy line’** On the core auto business, management had guided Q4 vehicle GPM at 15.8%-16.8% (down 3-4ppt QoQ from Q3’s underlying 19.8% ex-Mega recall). The low guide reflected a lower Mega mix and heavier discounts on the L-series, so the street expected ~16.6%. Actual Q4 vehicle GPM was 16.8%, in line with guidance and expectations, but ASP still declined QoQ, dragging margins. (Note: 2022 Q2 and 2023 Q2 vehicle GPM figures exclude the RMB 800mn+ contract loss and RMB 400mn warranty accrual impact, respectively.) **From a per-vehicle economics perspective:** **1\. ASP fell sharply QoQ on downmarket mix and heavier promo** Q4 ASP was RMB 250k, down RMB 27k QoQ and below the street’s RMB 258k, leading to the revenue miss. Drivers were downmarket mix and bigger discounts on legacy L-series models: **① Higher i6 mix, lower Mega mix** Model mix shifted downmarket as the low-priced i6 surged, with its share up ~25.7ppt QoQ to 26.2%, while the high-priced Mega fell ~5ppt QoQ to 3.2%, weighing on ASP. **② Deeper discounts on the L-series base** The L-series continued to face internal and external pressure: internal cannibalization from lower-priced EVs and intensifying competition from large, cheaper range-extended SUVs from peers. Li Auto therefore increased discounts across the L-series base. Discounts were larger than in Q3 (L6: -RMB 33k, L7/L8: -RMB 40k, L9: -RMB 45k, Mega: -RMB 12k), alongside continued insurance subsidies and 0%/low-rate financing. Still, L-series sales fell 60% YoY to ~62k in Q4, and even with i8+i6 EV contributions, total deliveries dropped 31% YoY to 109k. **2\. Per-vehicle cost fell RMB 15k QoQ, but scale could not offset ASP pressure** Q4 per-vehicle cost was RMB 208k, down RMB 15k QoQ (ex-Mega recall), helped by the lower-cost i6 and a slight volume uptick. But this could not offset the ASP decline, leaving GPM lower QoQ: **① Slight QoQ volume uptick lowered per-unit fixed costs:** Q4 deliveries were 109k, up 17% QoQ, near the top end of guidance, driven largely by the low-priced i6, allowing some scale benefits. However, overall volume still contracted YoY on L-series weakness. **② i6 GPM is structurally lower and weighed on mix** Management indicated i6 GPM would be below 15% before dual-supplier battery transition is complete and monthly volume exceeds 20k; only then could GPM exceed 15%. In Q4, i6 ramp was constrained by battery supply, averaging ~9.5k/month (Dec at 16k), weighing on overall vehicle GPM. **3\. Q4 per-vehicle GP was RMB 42k, slightly below the street’s RMB 43k** Per-car GP fell to RMB 42k from RMB 55k in Q3 (ex-Mega recall), a multi-year low since 2021, with vehicle GPM down 300bps QoQ to 16.8%. This breaks the long-held ~20% ‘healthy margin’ threshold, highlighting ongoing competitive and operating pressure. **II. Q1 revenue guide disappoints, with further ASP decline ahead** **a) Volume guide in line; i6 capacity gradually releasing** Q1 2026 deliveries are guided at 85k-90k, in line with the 88k consensus. Mar deliveries implied at 31k-36k (Jan/Feb at 28k/26k), as i6 supply starts to loosen in Mar; a second battery supplier is being introduced to ease CATL shortages. Note the i6 CATL-battery variant still has a lengthy 19-22 week delivery lead, so full i6 capacity release likely awaits Q2. Even with i6 contribution, Q1 is still 3%-9% below 2025 Q1’s 93k, given the sharp L-series decline that i6 has not fully filled. L-series Jan-Feb 2026 sales were just 18k (9k/month), -67% YoY from 55k. Dolphin Research sees the drop driven by: a) aging models and customer wait for Q2 refreshes (e.g., L9); b) intensified competition with larger, cheaper L-series rivals; c) cannibalization by i6+i8; d) sector-wide impact from the purchase tax rollback. **b) Revenue guide disappoints; ASP to decline sharply QoQ:** Q1 2026 revenue guide is RMB 20.4bn-21.6bn, well below the street at RMB 23.9bn, and -16% to -21% YoY. With volumes roughly in line, the shortfall is due to a sharp ASP drop to ~RMB 222k from RMB 250k in Q4 2025. Dolphin Research attributes the further QoQ ASP drop to: ① i6 mix rising as capacity releases, from 26% in Q4 2025 to ~60% in Q1 2026; ② deep discounts to clear legacy models ahead of the Q2 major cycle (e.g., new L9), plus growing discounting pressure in EVs (notably i8), with price cuts across both range-extended and EV lines further dragging blended ASP. Given lower ASP, weaker scale (Q1 deliveries -17% to -22% QoQ vs. Q4 2025), and upstream price hikes in lithium and memory inflating BOM, Dolphin Research expects Q1 vehicle GPM to decline QoQ, below the market’s 16.8% expectation (Q4 2025 level). Specific guidance will depend on management commentary at the results call. **III. R&D still ramping, while S&M remains under control** **1) R&D: modest QoQ uptick** Q4 R&D was RMB 3.02bn, up RMB 50mn QoQ and slightly above the street at RMB 2.95bn. Dolphin Research believes this reflects: a) front-loaded R&D for an intensive new model cycle in 2026, covering the full L-series refresh (L6/L7/L9) and a new all-electric full-size SUV i9; b) sustained ‘All in AI’ investment. AI is Li’s largest R&D line. In 2025, AI spend exceeded RMB 6bn of the total RMB 11.3bn, split roughly into: ~RMB 2.7bn (45%) for AI infrastructure, including foundation models, in-house inference chips, OS, and cloud/edge compute clusters; ~RMB 3.3bn (55%) for productization, including Li Assistant, multimodal agents, smart cockpit, intelligent manufacturing, and business analytics, helping its in-house AD models roll out via OTA to all AD Max models in H2 2025. **For 2026, AI investment strategy remains unchanged:** ① In-house compute platform rollout: the 5nm ‘Mach 100 (M100)’ chip is nearing mass production, delivering up to 2,560 TOPS per vehicle with dual chips, debuting on the new L9 Livis and expanding VLA foundational models across the lineup in 2026. ② Embodied intelligence: extending into ‘spatial robots’ by integrating foundation models, in-house chips, and a robotics OS, with a two-wheeled robot rumored to debut in H1. ③ Pushing to L4: Li targets achieving L4 autonomous driving by 2028 at the latest. Management has reflected that overemphasis on R&D ROI led to underinvestment when revenue softened, slowing model cadence and AI innovation. Going forward, R&D is unlikely to be cut further even if sales dip; instead, spend may increase. **2) S&M: down QoQ, well controlled** Q4 S&M was RMB 2.65bn, down RMB 120mn QoQ and below the street at RMB 2.74bn, likely due to a) headcount cuts in marketing (post-merger, 15% frontline sales reduction), lowering payroll; b) lower launch marketing spend (mostly in Q3); c) a shift from aggressive expansion to cautious operations. In Q4, only six retail and 15 service centers were added nationwide, and in Q1 2026, Li even closed or optimized underperforming stores to cut costs. **IV. Revenue contracted YoY again** With volumes known, total Q4 revenue was RMB 28.8bn, slightly below the street at RMB 29.6bn, marking another -35% YoY decline. Auto revenue was RMB 27.3bn (vs. cons. RMB 28.2bn), pressured by i6-driven ASP dilution and heavier promo on legacy L models; other revenue was RMB 1.52bn, up RMB 20mn QoQ, likely supported by higher attach rates from a larger installed base. Overall GPM was 17.8%, down 260bps from Q3’s underlying 20.4% (ex-Mega recall), mainly on heavier promos and downmarket mix. Other biz GPM rose 600bps QoQ to 36%, likely on higher-mix, higher-margin accessories (e.g., home chargers). **V. FCF turned positive; liquidity remains ample** Q4 operating cash flow was RMB 23.5bn, up RMB 11.0bn QoQ, while capex fell RMB 450mn QoQ to RMB 1.05bn. FCF was RMB 2.5bn, up RMB 11.4bn QoQ, turning positive, lifting cash and cash equivalents to RMB 101.2bn (net cash RMB 91.7bn), providing a solid liquidity buffer. Q4 capex of RMB 1.1bn fell ~RMB 450mn QoQ, likely as supercharger buildout slowed (487 stations added to 3,907, vs. 569 added in Q3) and store expansion moderated. For Dolphin Research’s past notes, please see: **Earnings reviews:** Sep 28, 2025 i6 launch review: [Li Auto (with event transcript): i6 Orders Soared, Why Did the Stock Swoon on the News?](https://longportapp.cn/en/topics/34579320?app_id=longbridge&utm_source=longbridge_app_share&channel=t34579320&invite-code=4NOXYT&locale=zh-CN&community_badge=1&profile_following_followers_activities=1) Aug 28, 2025 earnings review: [Li Auto: Range-Extender is Tiring, Pure EVs Are Tough — Can i6 Defy Gravity?](https://longportapp.cn/en/topics/33431267?app_id=longbridge&utm_source=longbridge_app_share&channel=t33431267&invite-code=4NOXYT&locale=zh-CN&community_badge=1&profile_following_followers_activities=1) Aug 28, 2025 call transcript: [Li Auto (2Q25 Trans): Back to a Single-SKU Blockbuster Model](https://longportapp.cn/en/topics/33444867?app_id=longbridge&utm_source=longbridge_app_share&channel=t33444867&invite-code=4NOXYT&locale=zh-CN&community_badge=1&profile_following_followers_activities=1) Aug 9, 2025 i8 launch review: [Li Auto: i8 Blinked Fast, Yet Couldn’t Save the Day?](https://longportapp.cn/en/topics/32789622?app_id=longbridge&utm_source=longbridge_app_share&channel=t32789622&invite-code=4NOXYT&locale=zh-CN&community_badge=1&profile_following_followers_activities=1) May 29, 2025 earnings review: [Range-Extender Is Aging, Pure EVs Uncertain — Reality Bites](https://longportapp.cn/en/topics/30164891?app_id=longbridge&utm_source=longbridge_app_share&channel=t30164891&invite-code=4NOXYT&locale=zh-CN&community_badge=1&profile_following_followers_activities=1) May 29, 2025 call transcript: [Li Auto (1Q25 Trans): Confident New L-Series Can Quickly Reach 50k/Month](https://longportapp.cn/en/topics/30170983?app_id=longbridge&utm_source=longbridge_app_share&channel=t30170983&invite-code=4NOXYT&locale=zh-CN&community_badge=1&profile_following_followers_activities=1) **Deep dives:** **Jun 24, 2025:** [**Aito vs. Li Auto: Two Titans — Who Wins?**](https://longportapp.cn/en/topics/31060822?app_id=longbridge&utm_source=longbridge_app_share&channel=t31060822&invite-code=4NOXYT&locale=zh-CN&community_badge=1&profile_following_followers_activities=1) **Jun 25, 2025:** [**Seres vs. Li Auto: Who Is the True ‘Chinese BBA’?**](https://longportapp.cn/en/topics/31114489?app_id=longbridge&utm_source=longbridge_app_share&channel=t31114489&invite-code=4NOXYT&locale=zh-CN&community_badge=1&profile_following_followers_activities=1) Risk disclosure and disclaimer: [Dolphin Research Disclaimer and General Disclosures](https://support.longbridge.global/topics/misc/dolphin-disclaimer) ### Related Stocks - [LI AUTO-W (02015.HK)](https://longbridge.com/en/quote/02015.HK.md) - [Li Auto Inc. (LI.US)](https://longbridge.com/en/quote/LI.US.md) - [Mega Matrix Inc. (MPU.US)](https://longbridge.com/en/quote/MPU.US.md) ## Comments (5) - **包身工 · 2026-03-12T14:05:12.000Z**: Well said - **春风不语,勾栏听曲! · 2026-03-12T13:12:55.000Z**: It lacks uniqueness and cost-effectiveness, and user reputation isn't good either; it's only going downhill. - **Distributed · 2026-03-12T13:11:11.000Z**: Where did you see that the L9 needs to reach the price level of the L8? - **Distributed** (2026-03-12T13:12:20.000Z): ...doesn't seem reliable, let's see next month - **Dolphin Research** (2026-03-13T09:10:21.000Z): From research minutes and major bank reports