Horizon Robotics (09660) Report Interpretation

Overview

Bernstein reviews Horizon Robotics' weaker-than-expected 1H26 results and reduces its forecasts and valuation. The report nevertheless maintains Outperform, arguing that the miss was already disclosed while HSD's expansion across the five largest domestic OEMs and its first signs of overseas progress strengthen the longer-term commercial case.

Core views

Horizon reported 1H26 revenue of RMB 2.055 billion, up 32.9% year on year, but the result was 8.3% below Bernstein's forecast and 6.4% below consensus. Gross profit was RMB 1.356 billion, also up 32.9%, while the 66.0% blended gross margin exceeded Bernstein's 61.0% forecast and the 62.0% consensus estimate. Bernstein cautions that the blended margin benefited from product mix rather than uniformly stronger underlying economics. Products-segment gross margin fell to 36.2% from 44.2% a year earlier because of the low-margin domain-controller delivery model. Management said newer design wins no longer require those deliveries, leading Bernstein to expect Products margin to begin improving from late 2026. Profitability was also weaker than Bernstein expected. The operating loss was RMB 1.672 billion and operating margin was negative 81.3%, compared with negative 97.3% in 1H25 but well below Bernstein's negative 57.7% forecast and consensus of negative 72.3%. Adjusted recurring net loss widened to RMB 1.671 billion from RMB 1.333 billion, with an adjusted net margin of negative 81.3%. Reported profit attributable to shareholders was RMB 3.784 billion, but Bernstein attributes the positive figure to non-recurring adjustments rather than the underlying operation. The institution's central point is that these weaknesses had already appeared in the preliminary disclosure, so the final results contained no material incremental negative surprise. The operational debate has shifted to HSD design-win timing and content per vehicle. Design-win progress and HSD-equipped model launches were slower than Bernstein had expected in 1H26, while initial sales of the relevant models were relatively soft amid weakness in China's passenger-vehicle market. Horizon shipped 2.2 million chips during the half and management reduced its FY26 shipment guidance to 5 million from 5.4 million at the beginning of the year. Intensifying OEM price competition is moving more L2++ programs toward single-J6M or dual-J6M configurations instead of the J6P configuration Bernstein had assumed. This reduces semiconductor content per vehicle, but the report argues that it also improves HSD's price-performance proposition for cost-conscious automakers. Bernstein sees customer breadth as the larger positive development for 2H26 and 2027. Management confirmed for the first time that HSD has secured design wins across all five leading domestic OEMs. Based on management's description, Bernstein infers that Geely is the incremental customer. Overseas traction is also emerging: J6B-led ADAS solutions have received orders from several overseas-market OEMs, and HSD has begun engaging with an overseas Tier 1 supplier for the first time. The latter remains an early signal, but it suggests that Horizon's overseas opportunity could eventually expand from entry-level ADAS into the higher-content HSD stack. Bernstein believes these developments can create upside against the softer 1H26 base. The weaker near-term results lead to substantial model changes. Bernstein forecasts FY26 revenue of RMB 5.345 billion, representing 42% growth. The narrative says the FY26 forecast was cut by 8.2% versus its prior estimate; the accompanying model table shows the new forecast 8.8% below the old RMB 5.858 billion estimate and 8.2% below consensus of RMB 5.824 billion. Forecast revenue is RMB 9.541 billion in 2027, RMB 15.342 billion in 2028, RMB 22.122 billion in 2029 and RMB 31.388 billion in 2030, implying growth of 79%, 61%, 44% and 42%, respectively. These 2027-2030 forecasts are approximately 2% lower than before, with the table showing reductions of 2.3%, 2.3%, 2.1% and 2.1%, mainly because of greater uncertainty around changes in the business model. Bernstein also modestly trims gross-margin assumptions as competition increases, partly offset by a greater contribution from the higher-margin Licensing & Services segment. Forecast gross margin moves from 65.9% in 2026 to 60.6% in 2027, 58.0% in 2028, 56.6% in 2029 and 55.6% in 2030. The model retains elevated but disciplined R&D spending and lowers steady-state 2029-2030 operating-margin forecasts by approximately four percentage points. Forecast operating margin is negative 57.4% in 2026 and negative 14.7% in 2027 before turning positive at 9.6% in 2028, 22.1% in 2029 and 31.4% in 2030. Management retains its 2028 breakeven target, while Bernstein moves its own breakeven assumption to the end of 2028 or 2029. Bernstein maintains Outperform but cuts its price target to HKD 8 from HKD 10. The main model-update discussion applies an 8.1 times 2BF EV/Sales multiple, down from 10.3 times, to RMB 11.2 billion of 2BF sales, down from RMB 11.5 billion; 2BF refers to 2H27 through 1H28. A separate valuation-methodology section also cites the 8.1 times multiple but states RMB 11.5 billion of 2BF sales. At the 31 August 2026 closing price of HKD 4.72, the report presents 69% upside to the revised target. Bernstein's Outperform definition for Asian ex-Japan stocks denotes expected outperformance of the relevant market index by more than 15 percentage points over the next 12 months.

Analysis framework

Bernstein first compares the final 1H26 results with its own estimates, consensus and the earlier preliminary disclosure to separate new information from already-known weakness. It then links shipment volumes, chip configuration and domain-controller mix to revenue and margins; assesses domestic and overseas design-win breadth; updates its 2026-2030 operating model; and applies a forward EV/Sales multiple to 2BF sales to derive the revised price target.

Methodology notes

• Event-Driven and Behavioral FinanceExpectation Gap and Expectation Management

  Actual-versus-expectations and preliminary-disclosure analysis

  The report compares 1H26 results with Bernstein's forecast and consensus, then asks whether the final release contained new negative information beyond the preliminary disclosure. This supports its conclusion that the earnings miss was already de-risked.

• Industry AnalysisVolume-price decomposition

  Shipment volume and content-per-vehicle decomposition

  Bernstein separates the number of chips shipped from the value of semiconductor content in each vehicle. The shift from J6P toward single- or dual-J6M configurations can reduce content per vehicle even if program coverage expands.

• Corporate Fundamentals and FinanceThree-statement linkage

  Integrated 2026-2030 financial forecasting

  The report provides linked income-statement, balance-sheet and cash-flow forecasts and uses revenue growth, gross margin, R&D spending and operating leverage to estimate the timing of operating breakeven.

• Valuation methods

  Forward EV/Sales multiple valuation

  Bernstein values Horizon by applying an 8.1 times enterprise-value-to-sales multiple to sales for 2BF, defined as 2H27 through 1H28, and uses the result to set the HKD 8 price target.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Horizon Robotics (09660.HK)

  The report-wide primary covered company and provider of automotive intelligent-driving chips and solutions.

  Strengths

  HSD design wins across all Top-5 domestic OEMs, an improving price-performance proposition, J6B-led overseas orders and initial overseas Tier 1 engagement.

  Weaknesses

  1H26 revenue and adjusted profitability missed expectations, HSD launches progressed slowly, FY26 shipment guidance was reduced and lower-spec configurations reduce content per vehicle.

  Comparison

  Revenue was 8.3% below Bernstein's forecast and 6.4% below consensus, while blended gross margin exceeded both forecasts; the stock had underperformed ASIAX by 82.0% over 12 months.

  Risks

  Slower smart-driving penetration, greater OEM in-house development and shipment or deployment delays at major customers.

Key data

• 1H26 revenueRMB 2,055.0 millionUp 32.9% year on year; 8.3% below Bernstein's forecast and 6.4% below consensus.
• 1H26 gross profit and marginRMB 1,356.1 million; 66.0%Gross profit rose 32.9%; margin was 499 basis points above Bernstein's forecast and 395 basis points above consensus.
• Products-segment gross margin36.2%Down from 44.2% in 1H25 because of the low-margin domain-controller delivery model.
• 1H26 operating resultRMB -1,671.5 million; -81.3% marginMargin improved from -97.3% a year earlier but was below Bernstein's -57.7% forecast and consensus of -72.3%.
• 1H26 adjusted recurring net lossRMB -1,671.2 millionCompared with RMB -1,332.5 million in 1H25; adjusted net margin was -81.3%.
• Reported profit attributable to shareholdersRMB 3,783.9 millionThe positive result was supported by non-recurring adjustments and did not represent underlying profitability.
• Chip shipments and FY26 guidance2.2 million in 1H26; 5 million FY26 guidanceFull-year guidance was lowered from 5.4 million at the start of the year.
• Revenue forecastsRMB 5.345bn/9.541bn/15.342bn/22.122bn/31.388bnForecasts for 2026E through 2030E, respectively, with growth of 42%/79%/61%/44%/42%.
• Forecast gross margins65.9%/60.6%/58.0%/56.6%/55.6%Bernstein estimates for 2026E through 2030E, reflecting competition and segment-mix changes.
• Forecast operating margins-57.4%/-14.7%/9.6%/22.1%/31.4%Bernstein estimates for 2026E through 2030E; the model turns positive in 2028.
• EV/Sales valuation metrics11.9x/8.4x/4.7xReported valuation metrics for 2025A, 2026E and 2027E.
• Price target valuation8.1x 2BF EV/SalesReduced from 10.3x; the main model update applies it to RMB 11.2 billion of 2BF sales, while a separate methodology section states RMB 11.5 billion.
• Market price and targetHKD 4.72 close; HKD 8.00 target; 69% upsideClosing price as of 31 Aug 2026; target reduced from HKD 10.00.
• Share-price performance-45.5% YTD, -8.7% 1M, -42.5% 6M, -50.0% 12MRelative performance versus ASIAX was -66.1%, -11.7%, -49.8% and -82.0%, respectively.

Impact & implications

Bernstein views the weaker result, lower shipment guidance and reduced chip content per vehicle as reasons to cut near-term estimates and valuation. It nevertheless argues that full coverage of the five leading domestic OEMs, improving delivery economics and early overseas engagement broaden Horizon's future revenue opportunity, leaving meaningful upside if programs convert into shipments during 2H26 and 2027.

Risks

• Smart-driving penetration may increase more slowly than expected.
• OEM in-house development may take more market share than expected.
• Shipment or deployment delays at major customers could reduce revenue growth.

What to watch

• Whether HSD programs across the Top-5 domestic OEMs convert into launches and shipments during 2H26 and 2027.
• Progress toward management's revised FY26 shipment guidance of 5 million chips.
• Whether Products-segment gross margin begins improving from late 2026 as new wins avoid domain-controller deliveries.
• The mix between J6P and lower-content single- or dual-J6M configurations as OEM price competition continues.
• Conversion of overseas J6B orders and the first overseas Tier 1 HSD engagement into deployments.
• Whether operating breakeven is achieved in 2028 or slips into 2029.