You will not find a shortage of industry reports claiming that Time-of-Flight sensing is about to conquer the smartphone. Most of them are written by component manufacturers who have a catalogue of ToF modules to sell. The real picture, looking across flagships shipped between 2020 and 2026, is more specific. ToF has a handful of genuine sweet spots, a handful of categories where it lost decisively to alternative approaches, and a quietly expanding long tail in industrial and robotic applications that is where the volume growth is actually coming from.

The basic physics is simple. A ToF sensor emits a pulse or a modulated wave of light — typically near-infrared at 850 nm or 940 nm — and measures the time or phase shift of the returned signal. From that, it calculates distance. Build an array of those pixels and you get a depth map of the scene. Compared to structured-light systems (which project a known pattern and infer depth from its deformation) and stereo-vision systems (which triangulate from two cameras), ToF promises simpler optics, faster frame rates, and better performance in low light. Those advantages are real. They are also not, in every application, sufficient.

Anyone evaluating 3D sensing for a consumer product is really choosing between three families of technology. Each has a distinct profile of strengths and costs, and the choice is rarely as clean as a single spec-sheet comparison suggests.

This is the core of why the smartphone ToF story unfolded the way it did. For the specific job of face authentication at arm’s length, structured light is simply more accurate — and Apple’s Face ID system has used a structured-light dot projector since iPhone X, not a ToF sensor. For photographic bokeh, computational approaches using dual cameras and neural networks have closed most of the quality gap that ToF was supposed to solve. For outdoor AR, stereo vision with neural depth refinement has proven more robust than ToF under direct sunlight, where ambient infrared swamps the sensor.

Apple introduced a rear-facing LiDAR scanner on the iPad Pro in early 2020 and brought it to the iPhone 12 Pro line later that year. The marketing framed it as a foundation for augmented reality — faster autofocus in low light, better portrait photography, and spatial mapping for AR apps. For several product cycles, the LiDAR module was a standard feature of the Pro-tier iPhone. Developers received a new set of APIs for room-scale scanning. Third-party apps appeared for home measurement, object capture, and accessibility.

The adoption curve for consumer-facing use cases, however, was flatter than Apple had hoped. AR measurement apps turned out to be a one-time novelty for most users. Object-scanning workflows remained the domain of professionals in specialised fields — estate agents, industrial inspection, accessibility research — rather than mass-market features. Vision Pro, Apple’s spatial computing headset, ultimately relied on a different sensing stack rather than porting the iPhone’s LiDAR architecture wholesale. By the iPhone 15 Pro launch cycle, Apple had begun a quiet walk-back, and rumours circulating through the supply chain suggested the module’s future on the iPhone was not secure.

The Android flagship response to Face ID and Apple’s LiDAR initiative was a scramble. Samsung, Huawei, LG, Honor, Sony, and several Chinese OEMs shipped ToF sensors in their top-tier 2019 and 2020 devices. The Samsung Galaxy S10 5G, Galaxy S20+, and Note 10+ all carried dedicated rear ToF modules. Huawei’s P30 Pro and P40 Pro included ToF. LG’s G8 ThinQ attempted front-facing ToF for hand-wave gestures. For a brief period, it looked as though ToF was on its way to becoming a standard flagship spec alongside optical image stabilisation and telephoto lenses.

Then the sensors started disappearing. The Samsung Galaxy S21 line removed the ToF module. So did the Note 20 Ultra’s successors. LG exited the smartphone business entirely. Huawei’s trajectory was disrupted by US export controls that were only incidentally related to optical sensing. By 2023, ToF had become a feature that appeared selectively on specific models for specific reasons, not a default flagship expectation. The reasons were unglamorous and largely economic. ToF modules added bill-of-materials cost — the VCSEL emitter, the specialised CMOS receiver, the dedicated illumination optics, and the processing overhead all sat above the rest of the camera stack. The feature differentiation they delivered was small enough that consumers did not reliably notice its absence.

Google’s position is worth flagging. The Pixel line has never shipped ToF, and has consistently produced industry-leading computational photography — including excellent portrait-mode bokeh — using a combination of dual-pixel autofocus data, neural depth estimation, and careful image processing. That is the real competitive threat to consumer-grade ToF in smartphones. If a pure software approach can produce 90% of the visible quality at zero additional bill-of-materials cost, the ToF module struggles to justify its inclusion.

The smartphone story is the most visible part of the ToF market but, in shipment-volume terms, no longer the largest. Three adjacent categories have quietly become the real drivers of ToF demand.

The unglamorous truth is that robotic vacuums are now one of the largest consumer-facing ToF markets in unit terms. The category has moved up-market rapidly — high-end models from Roborock, Dreame, Ecovacs, and others now routinely include ToF-based obstacle avoidance and sometimes dedicated LiDAR turrets for mapping. The sensing requirements of a floor robot are almost perfectly matched to what ToF does well: medium-range depth mapping, indoor lighting conditions, continuous operation, and sub-centimetre accuracy in the environment the robot actually has to navigate. Pet-detection models added in 2023–2024 further validated the sensor load.

Meta’s Quest line, Apple Vision Pro, Pico, and the various Chinese entrants all use depth sensing — often a combination of ToF and stereo-vision — to handle hand tracking, room mapping, and guardian-boundary setup. The sensors are less visible than on a phone because they sit inside the headset rather than being visible through an aperture, but the aggregate shipment volume has grown steadily. Meta sold tens of millions of Quest units across its product lines. Each one contains multiple depth-sensing cameras. This is a quiet but meaningful pull on the ToF and IR imaging sensor supply chain.

Driver-monitoring systems, occupancy detection, and in-cabin gesture control have become standard features on mid-to-high-end vehicles, driven partly by regulatory mandates in Europe requiring driver-attention monitoring in new cars. ToF is a strong fit for this application — it works in darkness, it is robust to changing ambient lighting, and it can operate at the frame rates needed for continuous monitoring. Several Tier-1 automotive suppliers have built in-cabin camera systems around ToF or hybrid ToF + IR architectures. The unit volumes here are smaller than smartphones but the design-in cycles are longer and the margins considerably better.

A modern ToF module is a stack of specialised photonic and semiconductor components that have each gone through their own compression curve over the past five years. Understanding what lives inside the module helps clarify why cost trajectories and form-factor improvements look the way they do.

The single most important hardware trend underlying the current market is the maturation of VCSEL arrays at 940 nm. The older 850 nm wavelength sat closer to the peak of ambient solar infrared, which made outdoor performance a persistent problem for early mobile ToF. The shift to 940 nm — where atmospheric water absorption reduces ambient IR — combined with tighter receive-side filtering has materially improved outdoor performance. It has not eliminated the problem, but it has raised the ceiling of usable conditions.

The second important trend is the move toward indirect Time-of-Flight (iToF) architectures in consumer modules, with direct Time-of-Flight (dToF) reserved for higher-end applications. iToF measures the phase shift of a continuous-wave modulated signal, which simplifies the receiver electronics at the cost of a fixed unambiguous range. dToF measures individual photon arrival times using single-photon avalanche diode (SPAD) arrays, producing longer-range and higher-accuracy data at substantially higher component cost. Apple’s iPhone LiDAR is a dToF system. Most Android ToF modules have been iToF. The split reflects a genuine architectural trade-off, not a hierarchical “better or worse” ranking.

The most important competitive force acting on consumer ToF is not another depth sensor — it is neural depth estimation from ordinary images. Monocular depth networks now produce startlingly good dense depth maps from a single RGB frame. Multi-frame approaches, dual-pixel parallax, and stereo-from-motion pipelines close the gap further. For the core consumer uses of ToF in a smartphone — bokeh, segmentation, measurement — the purely computational path is now competitive on quality and vastly cheaper in hardware.

This does not make ToF obsolete. Neural depth networks are excellent at producing plausible depth but poor at producing verifiably accurate depth. For any application that needs ground-truth distance — AR placement, volume estimation, accessibility features, robotics, driver monitoring — a physical ToF measurement retains its edge. What has shifted is the set of applications that actually require that ground truth. Most consumer photography applications do not. Most industrial and robotic applications very much do.

The ToF sensor supply chain is concentrated. Sony dominates high-performance mobile ToF sensor shipments through its CMOS imaging fab capacity. STMicroelectronics has become a leader in dToF modules for consumer applications, including the sensor inside the iPhone LiDAR module. Infineon, pmd, Melexis, ams OSRAM, and Analog Devices hold important positions across automotive and industrial segments. VCSEL production for ToF sits with Lumentum, II-VI (now Coherent), and a handful of Chinese suppliers including Vertilite and Everbright Photonics.

This concentration has two practical consequences. First, the supply chain is genuinely strategic — ToF modules are among the photonic components now subject to scrutiny under Western export-control regimes. Second, price trajectories over the next several years will depend significantly on whether Chinese sensor and VCSEL manufacturers continue their trajectory of closing the quality gap with incumbent suppliers. If they do, and political conditions permit cross-border trade, consumer module prices compress further. If they do not, or if trade fragments, prices stabilise at current levels with regional supply bifurcation.

Smart glasses are the category most likely to become the next genuine volume driver for consumer ToF — and also the category where the technology faces its hardest design challenges. The Meta Ray-Ban product line demonstrated that lightweight, socially acceptable smart glasses can reach meaningful consumer adoption when they deliver a narrow, well-chosen set of features. The next generation of products — from Meta, Samsung, Apple, and a growing Chinese field — is expected to add display and spatial sensing capabilities that will almost certainly require some form of depth input.

The design envelope for glasses is brutal. Every gram matters. Every milliwatt of battery matters more. Optical apertures are tiny, housing volumes are vanishingly small, and industrial design considerations often overrule what engineers would prefer. This pushes hard against conventional ToF module form factors. Expect a wave of increasingly miniaturised, often hybrid, depth-sensing modules — combinations of dual cameras, sparse ToF dot illumination, and neural fusion — rather than the large rear-mounted modules that appeared on 2020-era smartphones. The photonics engineering problem is genuinely interesting and not yet solved.

The neat narrative — “ToF is going to be in every consumer device by 2025” — never made physical or economic sense, and did not come true. The actual trajectory is more interesting. ToF has become a niche-but-growing component in smartphones, a near-ubiquitous component in higher-end robotic vacuums, a standard element in VR and AR headsets, a rapidly expanding part of automotive in-cabin systems, and an open question in smart glasses. The aggregate picture is healthy growth without the consumer-facing saturation the 2020 forecasts predicted.

The technology itself will keep getting better. Pixel counts will rise. Module heights will fall. Power draw will drop. SPAD-based dToF will continue its migration down-market from iPhone-tier products toward mid-range devices. What will not change is the underlying truth that a sensor needs an application to justify it. Hardware that answers a question nobody is asking gets designed out of the bill of materials, regardless of how elegant it is. The ToF industry has learned that lesson the hard way and is now — quite sensibly — chasing applications where depth sensing is a load-bearing feature rather than a marketing asterisk.