← Events← 返回日历
EVENT BRIEF · 事件摘要与解读
Mobileye Q2 2026 Earnings CallMobileye 2026 年二季度财报电话会
2026-07-23
Earnings · 财报
Brief v1
OFFICIAL MATERIALS官方材料直达
KEY NUMBERS · SOURCE: 8-K FILING, NOT MEDIA核心数字 · 取自 8-K 备案而非媒体转述
MARKET CONSENSUS IN ONE PARAGRAPH市场共识一段话
On paper this was a clean beat-and-raise: $508M revenue against ~$482M expected, $0.19 adjusted EPS against $0.06, guidance up. Shares opened up nearly 10% premarket and closed down about 15% — the sharpest single-day fall since August 2024. Nothing in the numbers caused that. Two announcements did: the founder-CEO of 27 years is leaving, and the company that spent its whole existence insisting it sells systems rather than operates fleets will now own and run robotaxis itself. The sell-side question stopped being "how good is the quarter" and became "who executes a capital-intensive pivot, and with whose money."纸面上这是一份干净的超预期加上调:营收 $5.08 亿对预期约 $4.82 亿,调整后 EPS $0.19 对预期 $0.06,指引上修。股价盘前一度涨近 10%,收盘跌约 15%——2024 年 8 月以来最大单日跌幅。跌的不是数字,是同日发布的两件事:执掌 27 年的创始人 CEO 将卸任;这家一直强调「我们卖系统、不运营车队」的公司,宣布要自持并自营 robotaxi 车队。卖方的问题从「这季有多好」变成了「谁来执行这场资本密集的转型,用谁的钱」。
AUTOSIGNAL READ · ANALYSISAUTOSIGNAL 解读
Editorial analysis built on primary data (SEC 8-K/10-Q, call transcript, our own signal archive incl. the Tesla and GM Q2 briefs published days earlier). Not investment advice.基于一手数据(SEC 8-K/10-Q 备案、电话会转写、自有信号档案——含前几日发布的特斯拉与通用 Q2 简报)的编辑部分析。不构成投资建议。
①A 216% EPS beat written by a tax law216% 的超预期是一部税法写的
Adjusted EPS came in at $0.19 against $0.06 consensus — a 216% beat that reads like an operating breakthrough and isn't one. Israel's newly enacted R&D Law offset $93M (non-GAAP) of Q2 R&D expense, and because it applies retroactively from January, the entire first half landed in this one quarter. Strip it out and H1 adjusted operating margin is 23%, up six points YoY — genuinely good, but a different story from "profit up 46%." The right read is neither "one-time noise" nor "operating leverage": the law has no expiration date and was passed to offset OECD Pillar Two, so it permanently re-bases margins for an Israeli R&D house. It is a structural gift from a jurisdiction, and it should be valued like one — at a lower multiple than an earned margin, and with a new dependency on Israeli tax policy that did not exist last quarter.
调整后 EPS $0.19 对预期 $0.06,超预期 216%——看起来像经营突破,其实不是。以色列本季新颁的研发法案抵扣了 $9300 万(非 GAAP)研发费用,且法案自年初追溯适用,于是整个上半年的补贴一次性落在这一个季度。剔除之后,上半年调整后经营利润率 23%,同比提升 6 个百分点——这确实不错,但和「利润增长 46%」是两个故事。正确读法既不是「一次性噪音」也不是「经营杠杆」:法案没有到期日,是为对冲 OECD 全球最低税(支柱二)而立,因此它把一家以色列研发型公司的利润率基线永久性抬高了。这是一份来自司法管辖区的结构性馈赠,就该按馈赠估值——给的倍数应低于挣来的利润率,并且从这一季起,公司多了一项此前不存在的依赖:以色列税收政策。②From supplier to operator: the pivot that competes with its own customers从供应商到运营商:一次与自己客户竞争的转身
Mobileye will now own and operate robotaxis: at least one U.S. city in 2027, 100–200 vehicles at launch, 10,000–20,000 over five years, funded from $1.2–1.3B of cash plus ~$350M annual operating cash flow. Management's unit math — vehicle under $100k, ~$125k annual revenue per car, opex "a few tens of thousands" — is defensible on a spreadsheet and brutal on a balance sheet: 10,000 vehicles is roughly $1B of capital before a single depot, and it converts a 66%-gross-margin licensing business into a fleet business with depreciation, insurance and remote-assist labor. The deeper problem is channel conflict. Mobileye's self-driving-system customers are MOIA (VW Group), Holon, and the OEMs it has spent a decade convincing that it is safely non-competitive. As of this call, its supplier now runs a rival fleet in the same market. Every Tier-1 that has tried to be both parts vendor and service operator — Bosch, Continental, Aptiv's Motional — found that the two businesses have different investors, and usually had to pick.
Mobileye 将自持并运营 robotaxi:2027 年至少进入一座美国城市,首批 100–200 辆,五年内 1–2 万辆,资金来自 $12–13 亿现金加每年约 $3.5 亿经营现金流。管理层的单车账——整车含传感器与算力低于 $10 万、单车年收入约 $12.5 万、年运营费用「几万美元」——在表格上站得住,在资产负债表上很残酷:1 万辆车约需 $10 亿资本,还没算一个车队场站;并且它把一门 66% 毛利率的授权生意,改造成一门要计折旧、保险和远程接管人力的车队生意。更深的问题是渠道冲突。Mobileye 自动驾驶系统的客户正是 MOIA(大众集团)、Holon,以及被它花十年说服「我不与你竞争」的车企。从这场电话会起,它们的供应商在同一市场里自己下场跑车队。所有试图同时做零部件供应商和出行运营商的 Tier-1——博世、大陆、安波福的 Motional——最终都发现这是两类投资人的生意,并且通常必须二选一。③The founder leaves on the day the strategy changes创始人在战略改变的同一天离场
Two 8-Ks filed the same morning: quarterly results, and Amnon Shashua stepping down as CEO after 27 years, staying on as a director with the chairmanship offered once a successor is found. His stated reason is that autonomous driving is "largely solved" and his energy belongs in humanoid robotics via Mentee, acquired this year. Read the sequencing rather than the sentiment: the board is hiring a search firm to find someone to execute a capital-intensive fleet business the founder just announced and will not run. Markets priced exactly that gap — up 10% premarket on the numbers, down 15% by the close on the governance. Note also that Intel still holds ~77%: the search happens under a majority owner with its own balance-sheet timetable, and every Mobileye strategy question ultimately routes through what Intel needs its stake to be worth.
同一个上午两份 8-K:季度业绩,以及 Amnon Shashua 在执掌 27 年后卸任 CEO——他将留任董事,新 CEO 到任后董事会已邀他出任董事长。他给出的理由是自动驾驶「问题基本已解决」,而他的精力应投向今年收购的 Mentee 所代表的人形机器人。要读的是顺序而不是措辞:董事会要聘猎头,去找一个人来执行一项由创始人刚刚宣布、而他本人不会执行的资本密集车队业务。市场给这道缝隙定了价——盘前因数字涨 10%,收盘因治理跌 15%。另需记住 Intel 仍持股约 77%:这场 CEO 搜寻发生在一位有自己资产负债表时间表的控股股东之下,而 Mobileye 的任何战略问题,最终都要绕回「Intel 需要这份持股值多少钱」。④SuperVision: 40k shipped, 30k demandedSuperVision:发出 4 万台,终端要 3 万台
H1 SuperVision shipments were ~40k against end-market demand of ~30k, and the full-year target stayed "slightly below 60k" — meaning roughly 10k units sit in customer inventory and H2 shipments must absorb it. Management frames this as customers hedging component supply, which is plausible; Q3 revenue is guided down 5–6% YoY, which is what absorption looks like. The signal worth keeping: this is the second consecutive quarter in which the premium-ADAS story runs ahead of end-user pull-through. Set it against the same week's GM brief, where Super Cruise attach hit 30–40% at launch with revenue recognized on subscription rather than shipment. Shipment-based ADAS revenue tells you what OEMs ordered; subscription-based tells you what drivers chose. Only one of those numbers survives a demand air pocket.
上半年 SuperVision 出货约 4 万台,终端需求约 3 万台,而全年目标仍维持「略低于 6 万台」——意味着约 1 万台压在客户库存里,下半年出货必须先消化它。管理层解释为客户对元器件供应的对冲备货,这解释可信;而 Q3 收入指引同比下降 5–6%,正是「消化库存」的样子。值得存档的信号是:这已是连续第二个季度,高端 ADAS 的叙事跑在终端拉动之前。把它对照同一周的通用简报:Super Cruise 新车附着率 30–40%,而收入按订阅确认而非按出货确认。按出货确认的 ADAS 收入告诉你车企订了多少;按订阅确认的收入告诉你用户选了多少。需求出现真空时,只有后一个数字扛得住。⑤Growth now rides on Chinese exports — and so does the risk增长骑在中国车企出海上,风险也是
Volumes rose 3% while revenue stayed flat: EyeQ shipments outpaced top-10 customer production by roughly eight points, but ASP fell because the incremental units came from Geely and Chery exports into emerging and developed markets. Management's confidence line was that "we haven't really seen a successful launch of a Chinese solution in foreign markets." That is true today and it is precisely the exposure: Mobileye's volume growth is now underwritten by the export success of Chinese OEMs whose domestic supply base — Horizon Robotics, Huawei, in-house teams at BYD and XPeng — is the single most likely source of its displacement. Every export unit strengthens the customer that has the strongest structural incentive to switch, at an ASP that is already diluting the mix. That is a growth engine and a countdown clock running on the same shaft.
出货量 +3% 而营收持平:EyeQ 出货增速跑赢前十大客户产量约 8 个百分点,但 ASP 下滑,因为增量来自吉利、奇瑞出口到新兴与发达市场的车型。管理层的信心句是「我们还没真正见到中国方案在海外市场成功落地」。这句话今天是对的,而它恰恰就是敞口所在:Mobileye 的销量增长如今由中国车企的出海成绩承保,而这些车企的本土供应链——地平线、华为,以及比亚迪、小鹏的自研团队——正是最可能替代 Mobileye 的力量。每一台出口车都在增强那个「最有结构性动机换供应商」的客户,而且是以一个正在稀释产品结构的 ASP 成交。增长引擎和倒计时钟,装在同一根轴上。THE BOTTOM LINE一句话结论
A quarter whose profit came from a tax law, a five-year strategy that turns a 66%-margin supplier into a fleet owner competing with its own customers, and the founder announcing both on his way out — the technology question at Mobileye now looks smaller than the ownership question: who runs it, whose capital funds it, and what Intel's 77% wants it to become.一个利润来自税法的季度、一份把 66% 毛利率供应商变成与自己客户竞争的车队所有者的五年战略、以及在离场路上同时宣布这两件事的创始人——Mobileye 的技术问题如今看起来已经小于所有权问题:谁来经营、谁的资本供养、以及 Intel 手中 77% 的股份希望它变成什么。
WATCHLIST (POST-CALL)后续观察清单
□ CEO search: internal operator vs external fleet executive — the choice IS the strategy · □ Q3: does the ~10k SuperVision channel inventory clear, and does the −5–6% revenue guide hold · □ MOIA/Holon reaction to their supplier becoming an operator · □ Named city, vehicle platform vendor and upfit partner for the 2027 launch · □ Gen 3 software shipping by year-end & Orlando commercial start · □ EyeQ ASP trend as China export mix grows · □ Whether the R&D Law grant stays at the $180–200M run-rate in FY27 guidance□ CEO 人选:内部技术派还是外部车队运营高管——这个选择本身就是战略 · □ Q3:约 1 万台 SuperVision 渠道库存能否消化,−5~6% 的收入指引是否守住 · □ MOIA/Holon 对「供应商下场做运营」的反应 · □ 2027 首城名称、整车平台供应商与改装伙伴落地 · □ Gen 3 软件年底交付与奥兰多商业化启动 · □ 中国出口占比上升下的 EyeQ ASP 走势 · □ FY27 指引中研发法案补贴是否维持 $1.8–2.0 亿的年化水平