Mobileye Global Inc. (MBLY)

Mobileye Global develops computer-vision systems that help cars sense the world and drive autonomously. Its software runs on purpose-built chips that interpret camera feeds to detect pedestrians, lane markings, obstacles, and hazards — everything needed for a vehicle to navigate without human intervention. The company was acquired by Intel in 2017 for over three billion dollars, but remains partly independent as a division within Intel and partly (as of 2023) a public entity again, with a complex corporate structure that reflects the difficulty of scaling autonomous-driving technology while holding multiple stakeholders’ interests.

Field notes: the camera-first bet

Mobileye’s founding insight was that cameras, not lidar (the expensive rotating laser scanners used by some rivals), could become the primary sensor for autonomous driving. A camera is cheap, proven, and already standard on every vehicle; it sees in visible light the way humans do; and with the right algorithms, it can extract depth and position without specialized hardware. This bet was radical enough in 2005, when the company was founded, that few believed it would work. The company spent years proving that pure vision could match or exceed the reliability of costlier sensor suites, and that economic and logistical edge became its moat.

By the time Intel bought Mobileye, the company had already proved the concept at scale. Its EyeQ computer-vision chip was in hundreds of thousands of vehicles worldwide, running in passenger cars, trucks, and buses. The chip does real-time object detection and lane tracking, feeds warnings to the driver, and logs vehicle and environment data. Mobileye’s software stacks multiple models — trained on enormous datasets of road footage — to recognize pedestrians at night, interpret hand signals, predict the behavior of nearby cars, and navigate complex urban scenes. None of this is trivial.

The autonomous bet and the money problem

The money Mobileye makes today comes almost entirely from selling safety-driver-assistance software. A major automaker integrates the EyeQ chip and the accompanying software into a vehicle model, and Mobileye earns a per-unit royalty, sometimes supplemented by upfront licensing or custom-development fees. This is recurring revenue: every generation of a popular model includes the system, creating a predictable revenue stream that ties to automotive production volumes.

The larger, more speculative bet is fully autonomous driving — vehicles that need no human at the wheel. Here the timing and economics have proven treacherous. The company has been developing its autonomous taxi platform, called Chauffeur, for more than a decade. It has conducted thousands of hours of testing on roads in Israel, the United States, and elsewhere. But the journey from lab to commercial deployment has been repeatedly delayed, hamstrung by technical challenges, regulatory uncertainty, and the enormous capital required to build a whole transportation business on top of the software.

That capital problem matters. Mobileye spends heavily on research and development — engineering teams across multiple countries, hardware design, simulation infrastructure, and the sheer human effort of teaching machines to drive safely. The company has shifted between operating at a loss (especially in the years after Intel’s acquisition) and modest profitability, depending on how much it is spending to advance the autonomous platform. Intel, as the parent, has subsidized those losses while betting that the autonomous-driving market would eventually materialize and justify the investment.

Dependent on automotive cycles and chip supply

Mobileye’s present business — the driver-assistance and safety systems that pay the bills — is tied entirely to the automotive industry. When car makers reduce production because of a recession, semiconductor shortages, or shifting demand, Mobileye’s revenue declines directly. The company also depends on the health of Intel’s chip manufacturing and supply chains, which created a secondary risk during the semiconductor shortage of 2021–2023. Though Mobileye is Intel’s subsidiary, it buys some of its own components from external suppliers and competes for capacity.

The other structural challenge is that driver-assistance systems are becoming commodified. Competitors like Nvidia (with its Drive platform) are offering their own end-to-end vision stacks, and traditional automotive suppliers like Bosch and Valeo have advanced their own systems. As the technology matures, automakers have more options, which puts downward pressure on unit economics and forces Mobileye to invest in next-generation features to stay ahead.

How a reader would track it

Start with the company’s quarterly earnings calls and investor updates, where management discloses safety-system revenue, unit volumes shipped, and progress on the Chauffeur autonomous taxi program. Watch for uptake announcements — newly announced automaker partnerships or contract wins signal whether the installed base is expanding. Track Intel’s total revenue and operating results, since Mobileye’s share is reported as a segment within Intel’s financials, though Mobileye now files its own annual reports and can be studied independently.

The 10-K filings (SEC CIK 0001910139) contain the most detail on revenue by customer, geography, and product line, and lay out the risks the company considers existential: regulatory changes, competition, the timeline and feasibility of autonomous driving. The key technical announcement to watch is progress on on-device AI — using the same chips for safety features and autonomous functions, rather than requiring separate systems.

For investors, the existential question is whether Mobileye’s vision-first approach will prove sufficient for full autonomy, and whether the company can commercialize Chauffeur and earn enough from it to justify the years of R&D spending and the capital requirements of a ride-hailing operation. The safety-assistance business is profitable but not transformative. Autonomy, if it works, could be worth ten times as much. Until that bet either succeeds or fails, Mobileye will be valued as a highly speculative play on an uncertain future, alongside a stable but modest cash-generating present.