A modern military EO/IR system is a remarkably dense piece of engineering:

Optics → detector → electronics → processing → stabilization → thermal management → mechanics

Every one of these blocks consumes size, weight and power. And the challenge becomes particularly severe when the system must also include: visible imaging, thermal imaging , laser range finding ,laser designation ,stabilization ,GPS/INS ,image processing ,automatic tracking. The market is therefore moving toward a very different question: How much capability can we put into every gram of payload?

That question is becoming particularly important for small UAVs, launched effects, robotics, soldier-worn systems and increasingly small satellites.

This is Part 2 of our analysis of EO/IR sensor miniaturization. Part 1 covered why sensors are shrinking and the engineering limits in optics, detectors, stabilization and cooling. Here we look at the companies leading the race, the three miniaturization frontiers, and where the technology goes over the next decade.

9. Who Is Technically Doing Something Different? The competitive landscape is surprisingly fragmented because there is no single route to miniaturization.

Some companies are attacking the detector and optical core. Others are specializing in micro-gimbals. Others are integrating increasingly complex combinations of EO + IR + LRF + designator + INS + processing into one stabilized system.

The most interesting companies are therefore not necessarily those with the smallest published product. They are the ones that have found a different way to preserve useful capability while reducing size, weight and power (SWaP).

9.1 LYNRED — miniaturizing the thermal core itself: LYNRED is probably one of the most important European companies to watch at the detector/core level. Its strategy is different from NextVision or Trillium. LYNRED is attacking the problem before the gimbal manufacturer even starts. The company produces: Cooled MWIR detectors , uncooled LWIR detectors ,SWIR detectors ,detector modules ,integrated thermal imaging modules ,compact PlugUp modules ,wafer-level/package technologies.

The recent ARGO MW SL is particularly relevant. It is a 640×512 MWIR module designed specifically for integration into small UAV gimbals. It operates at up to 130 K, has a stated NETD below 33 mK, and requires 4 W regulated input power during operation, with 12 W cooldown input.  Then in 2026 LYNRED introduced ARGO SL, explicitly targeting SWaP-constrained tactical drones and soldier-borne systems. It uses a T2SL MWIR architecture and is being positioned as a turnkey module rather than simply a bare detector.

Even more interestingly, LYNRED introduced YOCTO, an 8 µm-pixel uncooled microbolometer, and announced the SPIRIT European project to develop the next generation of sovereign European IR detector technology.

What is LYNRED's miniaturization philosophy? Effectively: Move more of the detector, correction, packaging and imaging functionality into a compact standardized module. This reduces the amount of engineering that a gimbal manufacturer has to build around the detector.

That is important because

detector → electronics → packaging → calibration → optics

can otherwise become a large portion of the payload.

9.2            Teledyne FLIR — make thermal imaging an OEM building block: Teledyne FLIR OEM is taking a somewhat different approach. The Hadron 640 is an excellent example. It combines: 640×512 thermal ,visible camera ,60 Hz operation ,IMU ,onboard electronics , standardized interfaces into a module weighing just 56 g. Its steady-state power is 1.8 W, with a published size of approximately 35 × 49 × 45 mm.  The interesting part is not simply the 56 g. FLIR provides software drivers and integration support for platforms including: NVIDIA Jetson ,Qualcomm RB5 ,other embedded compute architectures.

So FLIR's strategy is essentially: Make the sensor a standardized OEM building block that can be dropped into drones, robots and autonomous systems. This is exactly what you need if the market moves from hundreds of sophisticated military systems to thousands or millions of smaller autonomous platforms.

9.3 NextVision — perhaps the clearest micro-gimbal specialist: NextVision Stabilized Systems is one of the most interesting companies for this article because its product portfolio almost looks like a miniaturization experiment. Its current portfolio runs from:

125 g → 145 g → 180 g → 250–350 g → 640 g → 1.5 kg → 2.15 kg

with products such as:

The DragonEye2 is particularly impressive at the lower end: it provides EO/IR imaging with up to ×40 visible zoom and ×4 thermal zoom in a 125 g stabilized package.  But NextVision's strategy doesn't stop at optics. Its TRIP mission computer provides: EO/IR tracking ,object detection ,image enhancement ,geolocation ,video compression ,recording ,object/scene tracking ,dual VIS/IR modes.

This makes NextVision particularly interesting. It is moving toward: sensor + stabilization + navigation + processing as one compact product.

9.4 Trillium Engineering — stabilization through electronics: Trillium Engineering takes another route. Its HD25-LV weighs only 350 g and measures 71 × 109 mm.

It combines:1280×720 EO ,640×512 LWIR ,12× visible zoom ,2× thermal zoom ,electronic ,stabilization ,onboard scene tracking ,GPS/INS ,H.264 compression while consuming 10 W average / 75 W peak.  The IMU is co-located with the cameras and stabilization is performed electronically. So Trillium is attacking the problem we discussed earlier: Replace some mechanical precision with sensing and computation. This allows the mechanical gimbal to become smaller without completely giving up stabilized imagery.

9.5 Safran — integration rather than extreme miniaturization: Safran Electronics & Defense represents another strategy. The JIM Compact is not a micro-gimbal. It is a handheld multifunctional system weighing less than 2 kg including battery, with: Cooled 640×480 MWIR ,15 MP visible camera ,low-light channel ,12 km eye-safe LRF, GPS , digital magnetic compass ,image stabilization, image fusion. The interesting thing is how much functionality Safran puts into that mass. Its philosophy is essentially:

Don't just shrink the camera; combine several capabilities into one device.

That is especially important for soldier systems, where the user cannot carry separate: thermal imager ,day camera ,rangefinder ,GPS ,compass.  Safran is therefore optimizing capability density, rather than chasing the absolute minimum mass.

9.6 Thales — modular multifunction systems Thales is pursuing a similar strategy with the SOPHIE family. The Sophie Ultima combines: thermal imaging ,visible imaging ,infrared ,target location ,laser rangefinding ,teleconverter ,geolocation functions in a package weighing approximately 2.5 kg. Thales explicitly described the design as a modular concept intended to accommodate future technologies and capability upgrades.

This is important because modularity becomes increasingly valuable when sensors are being deployed across multiple platforms. Instead of redesigning the entire system every time a detector or processor improves: replace the module while keeping the overall architecture.

Thales has also pushed this philosophy into small-UAS systems. Its Spy'Ranger architecture combines lightweight UAVs with mature optronics and allows the sensor-derived target coordinates to feed directly into other systems.

9.7            CONTROP — integrating the entire targeting chain: CONTROP is particularly interesting because it illustrates the transition we discussed earlier: EO/IR → EO/IR + LRF → EO/IR + LRF + designator. Its T-STAMP family includes increasingly complex combinations of: FHD EO ,cooled MWIR ,stabilization ,LRF ,laser designator ,laser pointer ,laser illuminator.

The T-STAMP-XD, for example, integrates a 640×512 MWIR, FHD EO, 1064 nm / 50 mJ laser designator, 1550 nm LRF and optional pointer/illuminator in a 5.9 kg, 220 mm turret. This is not extreme miniaturization. It represents something arguably more difficult:

How much targeting capability can be integrated into a single stabilized optical architecture?

9.8 HENSOLDT — high-end sensor fusion rather than micro-gimbals: HENSOLDT is at another point on the spectrum. Its ARGOS-II family is a large airborne multi-sensor system integrating: HD thermal imaging ,HD multispectral TV ,SWIR options ,IMU ,GPS ,image processing ,autotracking ,laser rangefinder ,optional laser pointer/illuminator.

HENSOLDT says more than 800 ARGOS-II systems have been in operation globally, making it a mature fielded architecture.  The standard ARGOS-II HD is around 50 kg, so it should obviously not be compared directly with a 125 g NextVision camera. But that contrast is useful. HENSOLDT demonstrates the upper end of the capability-density problem: multiple sensors + long-range optics + laser + stabilization + navigation + processing. The question for the industry is how much of this capability can migrate downward into smaller platforms.

9.9 Leonardo — system-level integration: Leonardo is less of a pure micro-gimbal specialist and more of a system integrator. Its strength is combining sensors into complete ISR architectures. Leonardo's unmanned systems, for example, integrate EO/IR alongside radar, AIS, communications and mission systems. Its AWHERO maritime ISR configuration uses an 8-inch EO/IR together with radar and AIS.  Leonardo's current portfolio explicitly emphasizes open architectures and integration of electro-optic subsystems into broader ISR/ISTAR systems. So its miniaturization philosophy is less: "make the smallest camera" and more: "make the sensor interoperable with the entire mission architecture." That becomes increasingly important as distributed sensing develops.

9.10 ASELSAN — common aperture and single-LRU integration: ASELSAN is worth including because it has taken a very explicit approach to optical integration.  Its CATS UAV targeting system combines: 640×512 MWIR ,1920×1080 HD day camera ,low-light NIR ,LRF ,laser designator ,laser pointer ,laser illuminator ,IMU ,image processing ,automatic tracking ,geolocation. inside a single LRU/common-aperture architecture. The published system uses a 220 mm common aperture, weighs approximately 55 kg, and has an LRF/designator range capability up to 35 km according to the manufacturer specification.  The important innovation is not that CATS is small—it isn't. It is that ASELSAN is trying to put multiple optical functions behind a common aperture rather than simply adding separate optical assemblies. That's the same architectural principle that will eventually matter at much smaller scales.

ASELSAN's 2024 annual report also says CATS is actively fielded and had completed detection, tracking and designation tasks for more than 7,500 ammunition firings.

9.11 Tonbo Imaging — interesting Indian small-UAV strategy: Tonbo Imaging is one of the more interesting South Asian companies for this story. Its Hummingbird family is specifically aimed at small UAVs. The company says the HB-3500 integrates: cooled MWIR ,day imager ,LRF in a package under 4 kg.

This is a very interesting middle ground: 125 g micro-gimbal->350 g->640 g->1.5 kg-><4 kg EO/IR + cooled MWIR + LRF. Tonbo is effectively trying to preserve high-end thermal + ranging capability while bringing it into a much smaller UAV class than traditional airborne turrets.

9.12 Bharat Electronics — India is building the complete stack. Bharat Electronics (BEL) is different again. BEL has substantial experience integrating EO/IR, laser and navigation functions. Its CoMPASS Rev III program, produced through technology collaboration with Elbit Systems' ELOP, integrates: Colour TV ,third-generation MWIR ,SWIR ,laser target ,designator ,laser rangefinder ,GPS ,automatic tracking ,image processing.  BEL explicitly describes the Rev III as having reduced weight, modularity and space-saving packaging.

BEL is also pursuing smaller systems. Its product portfolio includes eye-safe LRFs, lightweight laser target designators, uncooled thermal systems and EO/IR payloads. More recently, BEL has also partnered with RRP Group on semiconductors, electro-optics and unmanned systems, showing that India is increasingly trying to move from system integration toward domestic component capability.

9.13 Hanwha Systems — Korea: small EO/IR plus autonomy: Hanwha Systems is an important Korean player. Its division-level UAV EOTS is explicitly described as a small-sized/lightweight electro-optical tracking system providing: EO/IR imaging ,automatic tracking ,rangefinding.

Hanwha is also developing EO/IR systems for Korea's medium-altitude UAV and small attack-drone programs. Its small attack-drone program specifically emphasizes:

high-performance, ultra-small and lightweight EO/IR

alongside small datalinks and autonomous/clustered operations. This makes Korea particularly interesting because the miniaturization effort is happening alongside autonomy and network integration, not purely optical miniaturization.

9.14 Japan — Tamron and the optical-module approach: Japan has a somewhat different strength. Tamron Business Products has developed compact camera modules using its optical and stabilization expertise. Tamron specifically describes its small lightweight zoom modules as using optical image stabilization and being suitable for applications including drones.

Its LW10F90-U thermal module is a good example of compact thermal integration: uncooled microbolometer ,80×80 ,8–12 µm ,1.0 W maximum ,46 × 46 × 37.4 mm ,~132 g.  This is not a military targeting gimbal, but it demonstrates a different Japanese strength:

compact optics + compact thermal imaging + optical stabilization.

Japan's NEC has also demonstrated the same philosophy historically. Its small-UAV sensor-gimbal work reached approximately 600 g, combining visible and uncooled IR sensing with a lightweight motor and compact structure.

9.15 Taiwan — NCSIST is already integrating the targeting stack: NCSIST provides an interesting Taiwanese example. Its EO payload combines, third-generation thermal imaging, HD TV ,eye-safe LRF ,stabilized platform ,target tracking  with published 1920×1080 EO + 640×512 thermal and at least 8 km ranging. It is designed for UAVs, helicopters and vehicles. So Taiwan is not only developing individual detectors; it is integrating the complete EO/IR + ranging + stabilization architecture.

9.16 Singapore — ST Engineering is approaching the problem from the platform side: ST Engineering is less directly comparable with NextVision or LYNRED. Its strength is system architecture.

ST Engineering's DroNet UAV platform uses a configurable payload bay supporting a basic 2-in-1 EO/IR camera, while its architecture is designed around autonomous operation and different sensor configurations. Its drone-agnostic operating platform also integrates autonomous/multifunction UAS networks and AI. So Singapore's approach is interesting because it moves the optimization one level upward:

Don't just miniaturize the sensor; make the UAV architecture capable of accepting different small sensors and processing their information autonomously.

9.17 China — potentially the most important volume ecosystem: China deserves a separate treatment because the market is much harder to characterize using public military specifications. There are major state-owned defense-electronics organizations such as CETC, AVIC and NORINCO, while a large private/commercial ecosystem supplies thermal cores, camera modules and small gimbals.

CETC describes UAV electronic equipment as one of its major defense-electronics product areas.  NORINCO's Sky Saker FX400A, for example, has an EO/IR turret with published 6–10 km detection/recognition, an 8 km laser designator and 15 km ranging capability.

Raytron: Raytron Microelectronics announced in August 2026 a <3.5 g 640×512 uncooled LWIR camera core with 8 µm pixels and a 13 × 13 mm footprint. The company explicitly targets commercial drones and micro-gimbals.  

Dali Technology: Dali Technology is another important Chinese thermal-imaging company.

Its SF740 module is: 640×512 ,12 µm ,120 Hz ,30 g excluding lens ,32 × 32 mm ,1.4 W.

Dali also says it has industrialized its own uncooled detector manufacturing, including 640×512 12 µm and 17 µm VOx arrays. This is exactly the kind of vertical integration that could become important if the market moves toward millions of inexpensive sensors.There are also Chinese micro-gimbal specialists such as Wuhan Kimbal and Wuhan JOHO, with products ranging from micro three-axis gimbals to larger multisensor systems.

The caveat: public information about military qualification, fielding and actual production volumes for many Chinese EO/IR suppliers is much less transparent than for European/US suppliers. It is therefore important to distinguish commercially demonstrated capability from military maturity rather than assuming the two are equivalent.

What the competitive landscape really looks like: This is how I would organize the market conceptually:

Layer 1 — Detector / sensor core

·        LYNRED

·        Teledyne FLIR

·        Raytron

·        Dali

Their battle is: How small can the sensing element become without destroying SNR, sensitivity and useful optics?

Layer 2 :Micro-gimbal

·        NextVision

·        Trillium

·        Tamron

·        Chinese micro-gimbal suppliers

Their battle is: How much stabilized imaging can we put into 100–500 g?

Layer 3 — Compact multisensor payload

·        Tonbo

·        NCSIST

·        Hanwha

·        Safran

·        Thales

Their battle is: How much EO + IR + navigation + ranging can we fit into 1–4 kg?

Layer 4 — Targeting system

·        CONTROP

·        ASELSAN

·        BEL

·        HENSOLDT

·        Leonardo

Their battle is: How much sensing + ranging + designation + stabilization + processing can we integrate into one precision optical architecture?

Market landscape table

The most important observation from this landscape: There are three distinct miniaturization frontiers developing simultaneously:

Frontier 1 — 1–100 g; Detector/core miniaturization: Examples: Raytron <3.5 g ,Teledyne FLIR 56 g ,Dali 30 g ,LYNRED compact detector modules . The objective is: Make sensing almost disappear into the platform.

Frontier 2 — 100 g–2 kg; Micro-gimbal miniaturization: Examples: NextVision 125 g → 2.15 kg ,Trillium 350 g ,compact Tamron modules ,small UAV EO/IR payloads ,The objective is: Add stabilization, optics and processing without losing the SWaP advantage.

Frontier 3 — 2–6 kg+; Capability integration: Examples: Safran JIM Compact  , Tonbo HB-3500 ,CONTROP T-STAMP,NCSIST EO payload , BEL/CoMPASS.

The objective becomes:Add LRF, navigation and eventually designation while remaining small enough for tactical platforms.

And above that sits the traditional high-end world: HENSOLDT / Leonardo / ASELSAN / large WESCAM-class systems

10. The Most Important Market Trend

I don't think the future belongs simply to the company that produces the smallest camera. The more important metric will be: Capability per kilogram.

Consider these three levels:

125 g:

·        EO + LWIR

·        Stabilization

·        compact UAV

350–650 g:

·        higher-resolution EO/IR

·        better stabilization

·        larger optics

1.5–4 kg:

·        multispectral

·         cooled IR

·        LRF

·        Illumination

·        advanced stabilization

↓

5–20+ kg:

·        large aperture

·        high-performance cooled sensors

·        long-range laser systems

·        sophisticated stabilization

The market is not replacing large EO/IR systems with small ones. It is adding an entirely new volume layer underneath them. The available industry data suggest that stabilized gimbal cameras are already a six-figure-unit global market when commercial and defense applications are combined, with approximately 162,000 units produced in 2025.

The military/tactical subset is much smaller and cannot be measured precisely from public data, but manufacturer production figures indicate that tens of thousands of lightweight stabilized systems can now be produced annually. NextVision alone reported production capacity of approximately 2,000 cameras per month at the end of 2025, spanning roughly 100 g to 2 kg products.

At the other end of the market, high-end EO/IR targeting turrets remain relatively low-volume systems, but their unit value and individual capability are much higher.

The result is a pyramid of capability:

Thousands to tens of thousands of lightweight sensors ->  Thousands of multisensor tactical systems -> Hundreds to low thousands of high-end targeting/ISR systems

The market is therefore not eliminating the 5–20 kg sensor. It is adding an enormous number of lower-cost performance tiers underneath it.

Unit-volume estimates are Photonics Intel estimates based on publicly available industry production data and market studies; no authoritative global database reports military EO/IR shipments by payload weight. The broader gyro-stabilized gimbal-camera market was estimated at ~162,000 units in 2025, including commercial applications.

11. What Happens Over the Next 5–10 Years?

My expectation is that the biggest change will occur at the bottom and middle of the market, rather than at the extreme high-performance end.

2026–2028: The 100–500 g era: Dual EO/IR stabilized systems become increasingly common on small UAVs, with roughly 100–500 g becoming a practical sweet spot for combining visible imaging, LWIR, stabilization and onboard processing.

2028–2032: The 50–250 g era: More capable stabilization, higher-resolution thermal imaging, onboard tracking and AI processing migrate into systems previously requiring substantially larger payloads.

2030–2035: The 5–50 g module era: The major change moves below the level of the complete gimbal. Thermal and EO sensor cores become small enough to be embedded directly into micro-gimbals, autonomous vehicles and distributed sensing nodes.

2035–2040: The sensor-network era: The question increasingly becomes not “How capable is this camera?” but: “How much useful information can we generate per kilogram, per watt and per dollar across the entire sensing network?”

The important caveat is that 5–10 g does not mean 5–10 g complete stabilized military turret. It is much more likely to mean: detector + optics + electronics with stabilization and processing handled elsewhere.

12. The Real Strategic Shift: The most important conclusion is therefore not:

"EO/IR sensors are becoming smaller." That is already obvious.

The more important shift is: EO/IR sensing is becoming a distributed semiconductor-scale capability.

The historical architecture was: Large platform → large stabilized turret → powerful sensor

The emerging architecture is increasingly: Small platform → micro-gimbal → compact EO/IR core → onboard AI

And eventually: Tiny platform → integrated multispectral sensor → onboard perception

This changes what military systems can do. A sensor that weighs 5 kg restricts you to a relatively small number of expensive platforms. A sensor that weighs 500 g can go on thousands of UAVs. A sensor that weighs 50 g can potentially go on tens of thousands. And a sensor approaching 5–10 g begins to make something fundamentally different possible: sensing becomes cheap enough, light enough and small enough to be distributed almost everywhere.

That is why the future will probably contain both extremes: tiny 5–50-g sensing modules for short-range/distributed systems, and kilogram-class EO/IR payloads for missions where aperture, cooling, laser energy and long-range identification still matter.

The real competitive race is therefore not simply "Who makes the smallest EO/IR system?"

It is: "Who can preserve the most operational capability while removing the most mass, power, volume and cost?"

And that is a much more interesting market to watch.