Mobileye Global Inc. (MBLY)

The only profitable, self-funding name in AV perception — using its EyeQ ADAS cash engine to fund a robotaxi ramp and a $900M humanoid-robotics bet.

Investment Research · Photoncap-style deep dive · v1 of "Mobileye" · May 14, 2026

What Mobileye physically does

Mobileye sells the eyes and, increasingly, the driving brain for vehicles. The core product is the EyeQ system-on-chip — a custom automotive vision processor that takes raw camera (and increasingly radar and lidar) input and runs the computer-vision and neural-network workloads that detect lanes, vehicles, pedestrians, signs and free space. EyeQ ships inside the ADAS (advanced driver-assistance systems) modules of roughly 200 million vehicles cumulatively; it is the highest-volume dedicated automotive perception chip in the world, and the recurring per-unit royalty-plus-silicon revenue from it is the cash engine that funds everything else Mobileye does. EyeQ is the binding constraint in the sense that perception — knowing what is around the vehicle and where it is going — is the hard, safety-critical layer of any driver-assist or autonomous system, and Mobileye has two decades of road-validated data and a fleet-sourced mapping system (Road Experience Management, REM) that competitors cannot replicate quickly.

On top of the EyeQ base, Mobileye sells a tiered ladder of autonomy products. SuperVision is a hands-off, eyes-on "everywhere" point-to-point assisted-driving system; Chauffeur is the next rung, an eyes-off conditional-autonomy system; and Drive is the full robotaxi-grade stack Mobileye supplies to fleet operators. The strategic logic is that the same perception core, the same REM maps and the same compute family scale from a $1,300 SuperVision module up to a robotaxi brain — Mobileye monetizes the entire autonomy curve from one technology base rather than betting the company on robotaxi alone.

The 2026 strategic expansion is into physical AI beyond the car. In January 2026 Mobileye agreed to acquire Mentee Robotics — an AI-first humanoid-robotics company founded by Mobileye's own CEO, Prof. Amnon Shashua — for roughly $900 million, a deal that closed in Q1 2026. The thesis Shashua articulated at CES 2026 is that the perception, world-modeling and sensor-fusion stack Mobileye built for vehicles is substantially transferable to humanoid and general-purpose robots, and that "physical AI" is the next addressable market for the same core competency. That is what puts MBLY squarely in the Robotics theme: not as a components supplier, but as a perception-and-autonomy platform deliberately extending from cars into robots.

Product roadmap

The EyeQ silicon roadmap runs through successive generations — EyeQ6 in volume now, with the high-end EyeQ6 High powering SuperVision and Chauffeur — and the company continues to win ADAS sockets across roughly three dozen OEMs. The dated, named product events that matter for the trade: SuperVision, priced around $1,300 per unit, is scheduled to integrate into Porsche and Audi models by late 2026; Chauffeur, priced between roughly $2,500 and $3,000 per unit, is set to launch with Audi in 2027, initially as an eyes-on system. On the design-win front, Mobileye disclosed at Q1 2026 a win with Mahindra covering SuperVision and Surround ADAS across at least six models from 2027 — a third Surround ADAS customer and a second SuperVision customer.

On robotaxi, the Drive program advances with named partners: Volkswagen's MOIA robotaxi program is moving toward series production on the Mobileye Drive stack, and Volkswagen Autonomous Mobility outlined at CES 2026 a roadmap targeting commercial robotaxi service across six cities by the end of 2027. Mobileye also continues robotaxi work with additional fleet partners. On the Mentee Robotics side, the disclosed roadmap is early: first on-site proof-of-concept humanoid deployments with customers are expected during 2026, intended to operate autonomously without teleoperation, with series production and commercialization targeted for 2028. What Mobileye does not do: it does not build cars, it does not operate robotaxi fleets itself (it supplies the stack to operators), and Mentee aside, it is not yet a robot manufacturer at scale — the humanoid business is a 2028-commercialization option, not a 2026 revenue line.

The financial print

Mobileye reported Q1 2026 results on April 23, 2026. Revenue was $558 million, up 27% year over year, driven by higher EyeQ unit volumes, increased ADAS fitment rates at core Western OEMs, and notably robust Chinese-OEM export demand. The headline GAAP number was ugly and needs context: Mobileye took a roughly $3.8 billion goodwill impairment in the quarter, producing a GAAP net loss of about $3.8 billion. That impairment is a non-cash writedown of acquisition goodwill (largely the legacy Intel/Mobileye carrying value) — it does not touch the operating business or the cash. On an operating basis the quarter was strong: adjusted EPS rose roughly 51% year over year, and management raised the full-year 2026 revenue outlook to about $1.975 billion. The cash position is the differentiator in this batch — Mobileye is profitable on an adjusted basis, generates operating cash, and used roughly $591 million of cash to close the Mentee acquisition while still authorizing a $250 million share-repurchase program. Forward consensus puts the stock around a high-20s forward P/E (the ~29.2x in our data), which is a real earnings multiple — unlike every other name in this batch, MBLY is valued on profits, not on revenue hope. The next binary is Q2 2026 earnings, expected on or around July 23, 2026.

Customer mix today

Mobileye's customer base is its structural strength: roughly three dozen OEM customers across the EyeQ ADAS franchise, with no single OEM dominant — there is no customer-concentration cliff of the kind that sinks single-program suppliers. The mix shift that matters in 2026 is geographic and product-tiered rather than customer-named. Geographically, Chinese-OEM export demand was explicitly called out as the swing factor in the Q1 2026 revenue beat — Chinese automakers exporting vehicles fitted with EyeQ, plus a normalization of Western dealer inventory after a soft 2024–25, drove the 27% growth. By product tier, legacy EyeQ ADAS remains the overwhelming majority of revenue and is the cash engine; SuperVision and Surround ADAS are the growth layer, still small in absolute revenue but with the Porsche, Audi and now Mahindra wins building a 2027-onward ramp; Drive (robotaxi) is pre-meaningful-revenue, monetizing through the VW/MOIA program toward 2027 series production; and Mentee/humanoid robotics is effectively $0 revenue today, a 2028-commercialization option. The honest framing: roughly 90%-plus of today's revenue is the mature ADAS franchise, and the entire "robotics theme" case is the 2027–2028 layer on top.

What's actually happening at the robotaxi and humanoid end-markets

The robotaxi mechanism is the VW/MOIA program. Mobileye supplies the Drive autonomous stack — perception, maps, driving policy — for Volkswagen's MOIA robotaxi service, which is advancing toward series production, and at CES 2026 VW Autonomous Mobility committed to a roadmap of commercial robotaxi service in six cities by end-2027. The significance is that this is a named OEM partner with a production timeline and a fleet operator behind it, not a science project — if it ships on schedule it converts Drive from a cost center into a per-vehicle revenue stream. Separately, SuperVision is the nearer-term proof point: management noted at Q1 2026 that SuperVision is demonstrating positive out-of-the-box US performance on unplanned routes and in challenging weather, which is the kind of generalization result that matters for the Chauffeur eyes-off launch with Audi in 2027.

On humanoid robotics, the Mentee mechanism is earlier and more speculative. Mentee builds an AI-first humanoid robot; Mobileye's pitch is that its vehicle-grade perception, world-modeling and sensor-fusion stack ports onto that platform, compressing Mentee's development timeline. The disclosed milestones: first on-site customer proof-of-concept deployments during 2026, running autonomously without teleoperation, with series production targeted for 2028. The skeptical read is that Mobileye paid $900 million for a pre-revenue startup founded by its own CEO — a related-party dynamic the 2026 proxy statement flagged — and the synergy claim, while plausible, is unproven. The bull read is that owning a humanoid platform gives Mobileye a captive design partner to prove the perception-stack-transfers thesis, and that 2028 commercialization is option value the market is barely paying for inside a 29x forward multiple.

The competitive threat / NVIDIA, Qualcomm and the in-sourcing OEMs

Mobileye's competitive set has three layers. The first is merchant silicon-and-stack rivals: NVIDIA's DRIVE platform is the most credible, pushing a high-compute, software-defined autonomy stack with strong developer mindshare, and Qualcomm's Snapdragon Ride is taking ADAS-and-beyond sockets with aggressive pricing. Both are larger, better-capitalized companies for whom automotive is one vertical among many. The second is OEM in-sourcing — the structural threat Mobileye has lived with for years. Tesla's vision-only FSD stack is the proof of concept that an OEM can build perception in-house and skip the merchant supplier entirely; several large OEMs have stood up internal autonomy teams. The third is the robotaxi-operator stack — Waymo most prominently — which competes for the eventual robotaxi market Mobileye's Drive program targets, though Waymo is an operator rather than a stack-supplier and the overlap is partial. The competitive rebuttal that supports the bull case: Mobileye's EyeQ volume, its REM crowd-sourced maps, and two decades of validated road data are genuinely hard to replicate, and most OEMs do not have Tesla's appetite or scale to in-source — which is why Mobileye keeps winning sockets (Mahindra in Q1 2026) even as the in-sourcing narrative runs.

The terminal risk

The terminal risk for Mobileye is that the EyeQ ADAS franchise — the cash engine that funds the robotaxi ramp and the Mentee bet — erodes faster than the growth layers scale to replace it. The erosion vectors are real: OEM in-sourcing (the Tesla model), merchant competition from NVIDIA and Qualcomm taking next-generation sockets, and the long-run risk that "vision-only with enough compute" commoditizes the perception layer Mobileye sells. If EyeQ revenue and margin compress through the late 2020s while SuperVision/Chauffeur are still ramping and robotaxi/Mentee are still pre-scale, Mobileye becomes a melting ice cube funding two long-dated options. The hedge is that this is precisely why Mobileye is the safest name in this batch despite the risk: it is profitable and self-funding today, it has a $250 million buyback and a fortress balance sheet, and it has roughly three to four years of visible EyeQ cash flow to land the transition. The multiple you can pay is constrained by the in-sourcing risk; the survival of the company is not in question the way it is for the micro-caps in this batch.

Bull / Gap / Optionality (Photoncap framing)

Bull

1. It is the only self-funding name in the batch. Mobileye is profitable on an adjusted basis, generated the cash to pay roughly $591 million for Mentee, and still authorized a $250 million buyback — all in a quarter when it also took a $3.8 billion non-cash goodwill writedown. Every other name in this perception batch is burning cash on a thin balance sheet; MBLY funds its robotics optionality out of operating cash flow.

2. The core business beat and the outlook was raised. Q1 2026 revenue of $558 million was up 27% year over year, adjusted EPS up roughly 51%, and management lifted full-year 2026 revenue guidance to about $1.975 billion. The ADAS franchise the bears keep writing off is currently accelerating, driven by Chinese-OEM export demand and Western inventory normalization.

3. The 2027 product ramp is contracted, not hoped-for. SuperVision integrates with Porsche and Audi by late 2026; Chauffeur launches with Audi in 2027; the Mahindra win adds SuperVision and Surround ADAS across six-plus models from 2027. These are named OEMs with priced products ($1,300 SuperVision, $2,500–$3,000 Chauffeur) and dated timelines — a visible step-up in revenue mix toward higher-ASP autonomy.

4. Robotaxi has a real production partner. The VW/MOIA Drive program is advancing toward series production, with VW Autonomous Mobility targeting commercial robotaxi service in six cities by end-2027. This is a funded OEM partner with a fleet operator and a timeline — the most credible robotaxi-supplier position outside of Waymo's in-house stack.

5. The Mentee humanoid bet is cheap option value inside the multiple. At ~29x forward earnings the market is paying for the ADAS-and-SuperVision business; the 2028-commercialization humanoid platform is barely priced. If the perception-stack-transfers thesis works even partially, Mentee is a free call option on physical AI bought with cash MBLY could spare.

Gap

1. ~90%-plus of revenue is still the mature ADAS franchise the bears want to short. The robotics theme case is the 2027–2028 layer; today MBLY is overwhelmingly a legacy EyeQ company. If in-sourcing and NVIDIA/Qualcomm competition bite before SuperVision and robotaxi scale, the cash engine compresses on schedule with the bear thesis.

2. The $3.8 billion goodwill impairment is a bad look even if non-cash. A writedown of that magnitude is the accountants confirming that the Intel-era carrying value was too high — it does not touch cash, but it is an admission that prior-regime expectations were not met, and it will be used against the stock.

3. The Mentee deal has a related-party stench. Mobileye paid roughly $900 million for a pre-revenue humanoid startup founded by its own CEO, Amnon Shashua — a conflict the 2026 proxy statement had to address. The synergy claim is plausible but unproven, and the price was set inside an obvious related-party dynamic.

4. RSI 77 and +30% above the 50-day MA — this is a momentum entry, not a value one. MBLY has run hard into the robotics narrative. At a high-20s forward multiple with the stock technically extended, a Q2 print that merely meets expectations could trigger a sharp mean-reversion; the entry has to respect that.

Optionality

The trade

Mobileye is the Bucket-B anchor of the sensing/perception batch — the one name you can own without lying awake about the balance sheet, because it funds its robotics optionality out of operating cash flow. Initiate at $9.98–$11.04 (current price ± ~5%, the correct construction for a name running at RSI 77 well above its 50-day MA), size at roughly 1.5% of risk capital — a notch above the speculative micro-caps in this batch because MBLY is profitable and self-funding, but still measured given the in-sourcing overhang — and set the stop at $8.80, below the 50-day EMA cloud and the pre-breakout consolidation shelf. The named catalyst is Q2 2026 earnings around July 23, 2026, the binary on whether the Chinese-export-driven ADAS acceleration holds and the FY26 raise is conservative. The cleaner way to think about MBLY: it is not really a robotics pure-play yet — the Mentee humanoid layer is a 2028 option — so own it as the profitable AV-perception compounder with a free physical-AI call option attached, and accept that the robotics theme membership is mostly forward-looking. If you want unhedged robotics-sensing torque and can stomach the burn, OUST and AEVA give more beta to the theme; MBLY is the quality leg. Conviction: 6 / 10.

Sources referenced inline throughout. Reference v1 of this template format: _Watchlist/hanmi-photoncap-style.md.

\newpage

OUST — Ouster, Inc. · WATCH (Tier-2) · Conv 6/10 · Bucket B

ticker: OUST name: Ouster, Inc. theme: Robotics bucket: B conviction: 6 entryzonelo: 31.60 entryzonehi: 34.90 currentprice: 33.26 pricedate: 2026-05-14 positionsizepct: 1.0 stoploss: 26.50 thesisoneline: Highest-volume digital lidar pure-play pivoting hard into physical AI and robotics; revenue inflecting but still loss-making on a thin balance sheet. catalystnext: Q2 26 earnings catalystdate: 2026-08-05 rsi: 67.2 vs50ma: 41.7 forwardpe: 0.0 themecycleposition: early customermixsummary: No single customer >10%; smart infrastructure + industrial automation now largest verticals, robotics/edge-AI fastest-growing, automotive a long-tail. terminalriskoneline: Chinese lidar (Hesai, RoboSense) cost curve collapses ASPs faster than Ouster can scale volume, or solid-state/camera-only perception displaces spinning/digital lidar in robotics. bulldriverscount: 5 gapriskscount: 4 optionalitycount: 5 lastearningsdate: 2026-05-05 nextearnings_date: 2026-08-05