August 2, 2026 · Sunday2026 年 8 月 2 日 · No. 29 · updated更新于 09:41 () SIGNALS信号EVENTS日历ARCHIVE归档LOG日志ACCESS接入ABOUT关于SAVED收藏

AUTOSIGNAL

Human-curated. Expert-annotated. Every signal traced to source. 人工精选 · 专家点评 · 每条信号可溯源

← Events← 返回日历

EVENT BRIEF · 事件摘要与解读

BMW Group Q2 2026 Results宝马集团 2026 年二季度业绩

2026-07-31 Earnings · 财报 Brief v1

OFFICIAL MATERIALS官方材料直达

KEY NUMBERS · SOURCE: THE THREE IR PDFS ABOVE, NOT MEDIA核心数字 · 取自上述三份 IR 备案而非媒体转述

Metric指标Q2 2026本季Read读法
Group revenue集团营收€31,259M (−7.9%)H1 €62,266M (−8.0%); FX-adjusted −7.8% in Q2, so this is volume, not currency上半年 €622.66 亿(−8.0%);Q2 剔除汇率后仍 −7.8%——是销量问题,不是汇率问题
Group EBT · margin集团税前利润 · 利润率€1,697M (−35.1%) · 5.4%H1 €4,045M (−29.4%), margin 6.5%. Net profit €1,200M (−34.9%), EPS €2.05上半年 €40.45 亿(−29.4%),利润率 6.5%。净利 €12.00 亿(−34.9%),EPS €2.05
Automotive EBIT · margin汽车板块 EBIT · 利润率€629M (−60.7%) · 2.3%Reported, no adjusted alternative offered — see Read ①. H1 €1,974M · 3.6%报表口径,宝马未提供任何调整后口径——见解读①。上半年 €19.74 亿 · 3.6%
What sits inside that 2.3%2.3% 里面装了什么Tariffs (US, EU) 1.25pp · BBA purchase-price-allocation D&A 1.2pp进口关税(美、欧)1.25pp · 华晨宝马购买价格分摊摊销 1.2pp△ The two together (2.45pp) exceed the reported margin itself△ 两项合计 2.45pp,已超过报告利润率本身
Financial Services vs Automotive金融服务 vs 汽车FS EBIT €647M vs Auto €629M; FS EBT €627M vs Auto €559M金融 EBIT €6.47 亿 vs 汽车 €6.29 亿;税前 €6.27 亿 vs €5.59 亿△ The lending book out-earned the car business this quarter — see Read ③△ 本季放贷比造车更赚钱——见解读③
Deliveries交付量590,947 (−4.9%)BMW 508,663 (−7.7%) · MINI 81,032 (+17.1%) · Rolls-Royce 1,252 (−11.5%)BMW 品牌 50.87 万(−7.7%) · MINI 8.10 万(+17.1%) · 劳斯莱斯 1,252(−11.5%)
China deliveries中国交付量117,815 (−30.2%)H1 261,773 (−20.4%) vs market −20.2%. △ China fell from 27.2% to 19.9% of Q2 group deliveries — see Read ②上半年 26.18 万(−20.4%),市场 −20.2%。△ 中国占集团 Q2 交付比重由 27.2% 降至 19.9%——见解读②
Europe · Americas · APAC/EE/ME/Africa欧洲 · 美洲 · 亚太/东欧/中东/非洲+7.6% · +9.4% (US +11.9%) · −10.9%+7.6% · +9.4%(美国 +11.9%) · −10.9%US growth was ICE-led as market BEV demand fell; APAC lost ~9,000 units to China export spillover美国增长由燃油车驱动(当地纯电需求下滑);亚太少交付约 9,000 辆,主因中国出口外溢
BEV · electrified纯电 · 电动化合计116,807 (+5.2%) · ~163,00019.8% BEV share, 27.6% electrified. But H1 BEV 204,295 is −7.4% — Q2 growth did not repair the half纯电占比 19.8%,电动化 27.6%。但上半年纯电 20.43 万仍为 −7.4%——单季回升未修复半年
Europe BEV欧洲纯电81,500 (+37.9%) · 31.3% share△ 69.8% of all group BEVs were sold in one region. Management says this supports meeting EU CO₂ targets in 2026△ 集团纯电销量的 69.8% 集中在欧洲一个区域。管理层称据此可再次达成 2026 年欧盟 CO₂ 目标
EBIT bridge vs Q2 2025EBIT 同比桥Volume/mix/pricing −€1.8B · FX −€0.4B · Other cost changes +€1.2B · SG&A +€0.2B · R&D in P&L −€0.1B销量/结构/定价 −€18 亿 · 汇率 −€4 亿 · 其他成本变动 +€12 亿 · 销管费 +€2 亿 · 研发费用化 −€1 亿Management attributes a large share of the −€1.8B to China. Note “Other cost changes” (+€1.2B, incl. warranty, manufacturing and tariff effects) is the single largest positive and the least itemised管理层称 −€18 亿中很大一部分来自中国。注意「其他成本变动」+€12 亿(含质保、制造与关税相关)是最大正项,也是拆解最少的一项
Automotive free cash flow汽车板块自由现金流€513M (−73.4%)H1 €1,290M (−45.0%); FY target >€2.5B. Working capital −€1.3B on inventory build (production > sales); capex-below-depreciation contributed +€0.6B上半年 €12.90 亿(−45.0%),全年目标 >€25 亿。营运资金 −€13 亿(产量高于销量致库存堆积);资本开支低于折旧贡献 +€6 亿
R&D · capex研发 · 资本开支R&D €1,959M (−3.8%) · capex €1,277M (−16.7%)研发 €19.59 亿(−3.8%) · 资本开支 €12.77 亿(−16.7%)H1 capex −30.5%; capex ratio 3.1% (H1) vs 4.0%. Depreciation to stay above capex all year — see Read ④上半年资本开支 −30.5%,资本开支率 3.1%(去年 4.0%)。全年折旧将持续高于资本开支——见解读④
Financial Services金融服务New contracts H1 866,088 (+5.0%) · penetration 52.9% (2025: 43.7%)上半年新增合同 86.61 万(+5.0%) · 渗透率 52.9%(去年 43.7%)Driven partly by Chinese local banks cutting F&I brokerage commissions since mid-2025. H1 PBT €1,008M (−15.4%) on a UK compensation provision; credit loss ratio 0.27%部分源于 2025 年年中起中国本地银行大幅削减金融保险产品佣金。上半年税前 €10.08 亿(−15.4%),因英国消费者补偿计提;信贷损失率 0.27%
Motorcycles摩托车EBIT €141M · margin 15.2%EBIT €1.41 亿 · 利润率 15.2%Deliveries 60,112 (−1.9%). The smallest segment ran 6.6× the car margin交付 60,112(−1.9%)。最小的板块跑出汽车板块 6.6 倍的利润率
FY26 guidance (confirmed)2026 全年指引(维持)Auto EBIT margin 1–3% · deliveries slightly down · Group EBT significantly down · Auto RoCE 1–5% · FS RoE 13–16%汽车 EBIT 利润率 1–3% · 交付略降 · 集团税前利润显著下降 · 汽车 RoCE 1–5% · 金融 ROE 13–16%Includes up to 1.25pp of workforce-restructuring burden — coincidentally the same figure as the Q2 tariff hit. Two different 1.25s; do not conflate them其中含最多 1.25pp 的裁员重组负担——与 Q2 关税冲击恰好同为 1.25pp。两个 1.25 含义不同,勿混用
Capital return股东回报Buyback tranche 2 done 26 Jun (2 months early); tranche 3 to finish by 30 Nov (5 months early)回购第二批 6 月 26 日完成(提前 2 个月);第三批 11 月 30 日前完成(提前 5 个月)€2B total programme; €1.25B returned in 2026回购计划总额 €20 亿;2026 年内返还 €12.5 亿
Neue Klasse (product only)新世代(仅产品口径)iX3 “on track to reach” 100,000 orders · 50,000 built at Debrecen · i3 Launch Edition early ordering from JuneiX3 「有望达到」 10 万订单 · 德布勒森已产 5 万台 · i3 首发版 6 月起提前开订Note the wording: 100,000 is a target being approached, not an achieved number. 40 new/updated models by end-2027注意措辞:10 万是「即将达到」而非已达成。2027 年底前推出 40 款新车与改款
L3 / ADAS / software monetisationL3 / ADAS / 软件变现Absent from all three documents三份文件中完全缺席No take rate, price, subscriber, deferred-revenue or OTA figure. The desk flagged this pre-call; the filing does not answer it — see Read ⑤没有任何附着率、定价、订阅数、递延收入或 OTA 数字。这是本台会前列出的观察点,备案未作回答——见解读⑤
◆ Market context (not filing data)◆ 市场背景(非备案数据)EBT €1,697M vs ~€1.6B consensus; margin guidance was cut from 4–6% to 1–3% on 16 June; shares near six-year lows税前 €16.97 亿对共识约 €16 亿;6 月 16 日利润率指引由 4–6% 下调至 1–3%;股价接近六年低点Press reporting also puts the severance programme at up to ~8,000 roles with savings from 2028 — neither figure appears in the three PDFs, which say only “voluntary severance in indirect functions in Germany”媒体报道称裁员规模最多约 8,000 人、节支自 2028 年起——这两个数字均未出现在三份 PDF 中,备案只写「德国间接部门自愿离职计划」

MARKET CONSENSUS IN ONE PARAGRAPH市场共识一段话

This quarter was pre-announced: BMW cut guidance by ad-hoc release on 16 June, so the €1,697M EBT — marginally ahead of a roughly €1.6B consensus, per market reporting rather than the filing — settled nothing. Sell-side attention has moved past the print to two questions the documents only partly answer: how deep the restructuring actually goes (JPMorgan reportedly called the June cut “radical”; the shares have been trading near six-year lows), and what a 1–3% full-year margin implies for H2 given H1 already ran 3.6%. Arithmetic the desk would put to management: hitting the top of the guided range requires the second half to be materially worse than the first, before the restructuring charge even lands. Treat every number in this paragraph as market context, not filing data — the filing itself gives no consensus, no share-price reaction, and no headcount figure.这是一个「已经预告过」的季度:宝马 6 月 16 日以临时公告下调指引,因此 €16.97 亿的税前利润(据市场报道略高于约 €16 亿的共识,非备案口径)并未解决任何争议。卖方的注意力已越过本期数字,转向文件只回答了一半的两个问题:重组究竟做到多深(据报道摩根大通称 6 月的下调「radical」,股价一直在六年低位附近),以及在上半年已录得 3.6% 的前提下,全年 1–3% 的利润率对下半年意味着什么。本台会向管理层提出的算术是:要落在指引区间上沿,下半年必须显著差于上半年——而这还没算上重组费用落地。本段所有数字均属市场背景而非备案数据:备案本身不提供共识值、不提供股价反应、也不提供裁员人数。

AUTOSIGNAL READ · ANALYSISAUTOSIGNAL 解读

Editorial analysis built on primary data (the three BMW IR statements above, plus our own 2026 Q2 archive — Mercedes-Benz, GM, Mobileye and Tesla briefs). Market-reaction items are labelled ◆. Not investment advice.基于一手数据的编辑部分析(上述三份宝马 IR 文件,以及本台 2026 Q2 自有档案——奔驰、通用、Mobileye、特斯拉简报)。市场反应类信息以 ◆ 标注。不构成投资建议。

The one number BMW refuses to dress up宝马唯一拒绝美化的数字
2.3% is a reported margin with no adjusted twin. Compare the same fortnight: Mercedes-Benz Cars posted €49M of reported EBIT (0.2% RoS) but headlined €909M adjusted (4.0%); GM headlined $3.57 adjusted EPS against $1.41 GAAP; Mobileye headlined $0.19 adjusted against a GAAP loss. BMW headlines the number that hurts. That is a disclosure culture worth crediting — and it changes how you read the two burdens disclosed inside it: 1.25pp of US/EU import duties and 1.2pp of purchase-price-allocation depreciation on BBA. Together 2.45pp, i.e. larger than the margin that survived. But resist the obvious add-back: the BBA amortisation is the accounting price of fully consolidating the Chinese JV, and China is precisely where the operating collapse happened. You cannot strip out the depreciation and keep the volume. Tariffs are the genuinely exogenous line; BBA is the cost of a strategic choice made years ago, now being paid at the worst possible moment. 2.3% 是一个没有「调整后孪生兄弟」的报表数字。对照同期:奔驰乘用车报表 EBIT 只有 €4,900 万(利润率 0.2%),但标题写的是调整后 €9.09 亿(4.0%);通用标题是调整后 EPS $3.57(GAAP 仅 $1.41);Mobileye 标题是调整后 $0.19(GAAP 亏损)。宝马把最难看的那个数字放在标题上——这份披露文化值得记一笔,同时也改变了我们读它内部两项负担的方式:美欧进口关税 1.25pp,华晨宝马购买价格分摊折旧摊销 1.2pp,合计 2.45pp,比幸存下来的利润率还大。但不要顺手做加回:BBA 摊销是全资合并中国合资公司的会计代价,而中国恰恰是经营崩塌发生的地方——你不能把折旧剔掉、却把销量留下。关税是真正的外生项;BBA 是多年前一项战略选择的成本,如今在最糟糕的时点结账。
China: full consolidation means nowhere to hide — and a chosen denominator中国:全并表意味着无处可藏——以及一个被选择的分母
Three German and American premium makers, three structurally different Chinas. GM's China is an equity line ($83M, +17%) that rounds to zero against $3.9B of EBIT — an option, not a pillar. Mercedes runs much of its China exposure through equity accounting and took the pain as a roughly €704M non-cash impairment of Chinese equity-method investments, stripped out of its adjusted EBIT. BMW fully consolidates BBA, so a −30.2% quarter lands directly in revenue, EBIT and free cash flow, with no adjustment layer. That is the honest structure and the brutal one. Now the language: Nedeljković says China performance “was at least in line with the overall market development.” On the H1 basis quoted in the CFO statement that is defensible — −20.4% versus a −20.2% market. On the quarter actually being reported, deliveries fell 30.2%. The filing gives no Q2 market-decline figure to compare against, so the only period where “in line” can be verified is the one that is not the reporting period. Choosing the denominator is not lying; it is the most consequential editorial decision in the whole statement. △ Derived: China fell from 27.2% to 19.9% of group Q2 deliveries in twelve months. Whatever BMW's China strategy becomes, it is now being executed on two-thirds of the base it was designed for. 三家德美豪华车企,三种结构完全不同的「中国」。通用的中国是一行权益法收益($8,300 万,+17%),相对 $39 亿 EBIT 四舍五入等于零——是期权,不是支柱;奔驰的中国敞口大量走权益法,痛苦以约 €7.04 亿中国权益法投资非现金减值体现,并在调整后口径中被剔除;宝马全资合并 BBA,于是 −30.2% 的一个季度直接砸进营收、EBIT 和自由现金流,中间没有任何缓冲层。这是最诚实的结构,也是最残酷的结构。再看措辞:Nedeljković 称中国表现「至少与整体市场发展同步」。按 CFO 讲稿引用的上半年口径,这站得住——−20.4% 对市场 −20.2%;但按正在报告的这个季度,交付下滑 30.2%。备案没有给出 Q2 单季市场跌幅可供对照,因此唯一能验证「同步」的期间,恰好不是报告期。选分母不算说谎,但这是整份讲稿里最有后果的一个编辑决定。△ 本台测算:十二个月内,中国占集团 Q2 交付的比重由 27.2% 降至 19.9%。无论宝马的中国战略最终成为什么,它现在只能在原设计基数的三分之二上执行。
The lender out-earned the carmaker放贷的赚过了造车的
Straight from the table, no interpretation needed: Financial Services delivered €647M of EBIT against Automotive's €629M, and €627M of EBT against €559M. Mercedes showed the same shape in an even starker form — FS EBIT at 9.7× reported Cars EBIT. This is now a structural feature of German premium, not a quirk of one quarter. The mechanism is disclosed and deserves attention: penetration jumped to 52.9% from 43.7%, and the CFO attributes part of that to Chinese local banks cutting F&I brokerage commissions since mid-2025 — BMW is writing more of its own paper in the market where residual values and pricing are least predictable. The credit loss ratio at 0.27% is reassuring but structurally lagging; loans written into a −30% market show up as losses years later, not quarters. A finance arm that carries group earnings while the product business runs at 2.3% is a warning about the product business, not a triumph of the finance arm. 直接来自财务表,无需诠释:金融服务 EBIT €6.47 亿,高于汽车板块的 €6.29 亿;税前 €6.27 亿,高于 €5.59 亿。奔驰呈现的是同一形状的更极端版本——金融 EBIT 是乘用车报表 EBIT 的 9.7 倍。这已是德系豪华的结构性特征,而非单季偶然。机制在备案中写明,值得留意:渗透率由 43.7% 跃升至 52.9%,CFO 将其部分归因于 2025 年年中起中国本地银行削减金融保险经纪佣金——也就是说,宝马在残值与定价最不可预测的市场里,自己承接了更多信贷。0.27% 的信贷损失率看着安心,但它天然滞后:在一个 −30% 的市场里放出去的贷款,不良是以年计、而非以季计浮现的。当金融板块扛起集团利润、而产品业务只跑出 2.3%,这是关于产品业务的警报,不是金融板块的捷报。
Where the €513M of cash actually came from€5.13 亿现金究竟从哪来
Free cash flow of €513M (−73.4%) is not a small number rescued by discipline; it is a small number assembled from unusual parts. Working capital took out €1.3B because production exceeded sales — inventory built, and management says it has already cut production plans to bring stock back toward prior-year levels by December, which mechanically weakens H2 revenue. The largest positive contribution, +€600M, came from depreciation exceeding capital expenditure, a condition the CFO says will hold all year. Read that plainly: the cash contribution comes from yesterday's investments amortising faster than today's are being made. H1 capex is down 30.5%, R&D down 7.6%, and the R&D ratio is 6.0% for the half. Meanwhile the third buyback tranche was pulled forward to finish five months early, returning €1.25B in 2026. Accelerating shareholder returns in a 2.3%-margin quarter, funded partly by underinvesting relative to depreciation, is a defensible signal to equity holders and a genuine risk to a company that is one year into a platform generation. Mercedes made the opposite call on the same problem — PP&E up 33% while cutting expensed R&D through capitalisation. Both are choices; only one of them is disclosed as a choice. €5.13 亿(−73.4%)的自由现金流,不是靠纪律守住的小数字,而是由几个非常规部件拼起来的小数字。营运资金流出 €13 亿,因为产量高于销量——库存堆积,而管理层称已下调生产计划、力争年底把库存拉回上年水平,这在机制上就意味着下半年收入更弱。最大的正贡献 +€6 亿来自折旧高于资本开支,CFO 明确这一状态将贯穿全年。直白读法:现金贡献来自昨天的投资摊销得比今天的投资投得更快。上半年资本开支 −30.5%,研发 −7.6%,半年研发比率 6.0%。与此同时,第三批回购提前五个月完成,2026 年返还 €12.5 亿。在一个 2.3% 利润率的季度加速股东回报,而资金部分来自「相对折旧的投资不足」,对股东是可辩护的信号,对一家刚进入新平台世代第一年的公司则是真实风险。奔驰面对同一问题做了相反选择——固定资产投资 +33%,同时靠提高资本化率压低费用化研发。两者都是选择,但只有一个被当作「选择」披露出来。
A Neue Klasse quarter with no software in it一个没有软件的「新世代」季度
This is the finding that matters most for the SDV thesis, and it is a finding about absence. Across three documents and roughly 10,000 words, the Neue Klasse appears exclusively as orders, plant shifts, model launches and drivetrain variants. There is no L3 statement, no ADAS take rate, no software subscription line, no deferred revenue, no OTA revenue, no architecture or compute disclosure. The desk went into this call with “Neue Klasse software stack” and “L3/ADAS commercialisation” on the watch list; the filing does not engage with either. AI appears three times, every time as an internal cost tool — Mertl on using industrial data plus AI to shorten virtual validation cycles, Nedeljković on automating processes to increase speed. The only product-software sentence in the entire package is that the China-market iX3 and long-wheelbase i3 “incorporate technologies from local tech players,” with no partner, scope or economics named. Set that against the peer set: GM disclosed $6.3B of OnStar-line deferred revenue and a 30–40% Super Cruise renewal rate; Mobileye reports systems shipped and average system price to the dollar ($48.5); Mercedes gave what our brief called “no financial disclosure whatsoever” on MB.OS. BMW is now in the Mercedes column. And note the horizon language: returning to the 8–10% EBIT corridor “by the beginning of the next decade” is a four-plus-year target with no interim milestone attached — structurally the same unfalsifiable form we flagged in GM's “largely meeting milestones” and Tesla's “production commenced.” The four internal workstreams Nedeljković named — customer journey, organisational structure, delivery and purchasing, engineering — are all process. None of them is a product or software commitment. A company can be genuinely ahead on vehicle software and choose not to quantify it; but researchers can only price what is disclosed, and on this evidence BMW is monetising AI as an internal margin lever while leaving the vehicle-software business case entirely unpriced. 这是本季对 SDV 议题最重要的发现,而它是一个「缺席」型发现。三份文件、约一万字里,「新世代」只以订单量、工厂班次、车型上市和动力总成变体的形式出现。没有 L3 表述,没有 ADAS 附着率,没有软件订阅收入行,没有递延收入,没有 OTA 收入,没有电子架构或算力披露。本台带着「新世代软件栈」和「L3/ADAS 商业化」两个观察点进场,备案对二者均未回应。AI 只出现三次,每次都是内部降本工具——Mertl 讲用工业数据结合 AI 缩短虚拟验证周期,Nedeljković 讲用 AI 自动化流程提速。整份材料中唯一一句与产品软件相关的话,是中国市场的 iX3 与 i3 长轴版「整合了本土科技企业的技术」,未点名合作方、范围与经济性。对照同业:通用披露 OnStar 口径递延收入 $63 亿、Super Cruise 试用期满 30–40% 续费;Mobileye 把出货量与平均系统价精确到美元($48.5);奔驰在 MB.OS 上则被我们的简报记为「零财务披露」。宝马现在站在奔驰这一栏。再看时间锚:「在下一个十年之初」回到 8–10% EBIT 区间,是一个四年以上、且不附任何中期里程碑的目标——在结构上与我们此前标记的通用「大部分里程碑达成」、特斯拉「production commenced」属于同一类不可证伪表述。Nedeljković 点名的四条内部战线——客户旅程、组织结构、交付与采购、工程——全是流程,没有一条是产品或软件承诺。一家公司完全可能在整车软件上真正领先却选择不量化;但研究者只能为已披露的东西定价。就现有证据而言,宝马把 AI 变现在内部利润率杠杆上,而把整车软件的商业模式完全留白。

THE BOTTOM LINE一句话结论

BMW reported the quarter without a cosmetic layer — one margin, 2.3%, with the tariff and China-consolidation costs left inside it — and answered it with process: severance, leaner structures, less capex, faster buybacks, and a return to 8–10% pushed out to “the beginning of the next decade”; what a company this honest about its numbers did not disclose is any figure at all for the software, ADAS or autonomy business the Neue Klasse was supposed to be built to sell, which leaves the most important question of BMW's decade the only one with no number attached to it.宝马以零修饰的方式交出了这个季度——只有一个利润率 2.3%,关税与中国并表成本都原样留在里面——并用「流程」来回应:裁员、精简结构、削减资本开支、加速回购,以及把回到 8–10% 的时点推到「下一个十年之初」;但这家在数字上如此诚实的公司,唯独没有披露任何一个关于软件、ADAS 或自动驾驶业务的数字——而新世代原本正是为卖这些东西而生的。于是宝马这个十年最重要的问题,成了唯一没有数字的那一个。

WATCHLIST (POST-CALL)后续观察清单

□ Q3: does any quantified Neue Klasse software/ADAS metric ever appear, or does the silence hold · □ Whether the iX3 “on track to 100,000 orders” converts into a stated orders-achieved number · □ H2 margin arithmetic: 1–3% FY vs 3.6% H1 already banked, before the restructuring charge · □ Size and P&L timing of the severance programme once quantified in the filing (press reports are not filing data) · □ Inventory normalisation: does the production cut show up as weaker H2 revenue · □ China Q2-basis market comparison — will management ever quote the quarter, not the half · □ Financial Services: penetration above 50% plus a −30% China market — watch the 0.27% credit loss ratio and residual assumptions · □ Whether capex stays below depreciation into 2027 while 40 models launch□ Q3 是否首次出现可量化的新世代软件/ADAS 指标,还是继续沉默 · □ iX3「有望达到 10 万订单」何时兑现为「已达成」的明确数字 · □ 下半年利润率算术:全年 1–3% 对上半年已实现的 3.6%,且尚未计入重组费用 · □ 裁员规模与损益计入节奏何时在备案中量化(媒体报道不等于备案数据) · □ 库存正常化:减产是否体现为下半年收入进一步走弱 · □ 中国的 Q2 单季市场对照——管理层是否会改用季度口径而非半年口径 · □ 金融服务:渗透率突破 50% 叠加 −30% 的中国市场,盯住 0.27% 信贷损失率与残值假设 · □ 在 40 款新车上市期内,资本开支是否持续低于折旧至 2027 年