Two years ago, the case for a software-defined camera with an open operating system seemed clean. Decouple hardware from software, expose the OS to third-party developers, and Canon, Nikon, and Sony would face the same fate that hit BlackBerry and Nokia: a closed ecosystem hollowed out by a marketplace of apps and a community of tinkerers. The thesis was directionally right about one thing — software would dominate the next decade of camera differentiation — and wrong about everything else.
The disruption came from inside the wall. Nikon spent $85 million acquiring RED Digital Cinema in early 2024, and by October 2025 it shipped the Nikon ZR: a 6K cinema body with internal R3D recording at $2,200, undercutting the Sony FX3 by nearly half. The R3D codec — RED’s defining proprietary asset, the thing that made the brand a legend among cinematographers — now ships as a feature inside a Z-mount mirrorless body designed and priced for working creators rather than rental houses. There is no open marketplace, no sideloaded plugin, no community fork. There is a closed pipeline that Nikon owns end-to-end and prices like a consumer product. That is what disruption in this industry actually looks like.
Sigma’s BF, released in 2025, made the second case against the open-OS thesis. It is the most radical rethink of the camera interface in a generation: ten top-level parameters, ten menu items, three buttons and a dial, the mode dial deleted entirely as a holdover from the film era. It runs on closed firmware. The interface revolution that the open-OS argument predicted would come from third-party developers came instead from a single Japanese lens manufacturer carving aluminum blocks for seven hours apiece. The lesson is not that simplicity beats openness. The lesson is that interface rethinks are bottlenecked on taste and conviction, not on developer access.
The actual open-OS project, Apertus AXIOM, has been in development since the late 2000s. Its status page still reads “In Development.” A few hundred Beta Developer Kits shipped to backers without enclosures or a finished operating system, and the AXIOM Gamma — promised for 2016 — never arrived. Magic Lantern, the Canon-DSLR firmware hack that was supposed to be the wedge into the mirrorless era, never made the jump. The community-driven flywheel that the thesis depended on never spun up, and the institutional appetite for an open camera platform — among manufacturers, among working photographers, among capital — turned out to be near zero.
What did spin up was firmware-as-product inside the closed systems. The Nikon Z9 has received feature updates over its lifetime that would have been new SKUs a generation ago: bird detection, pixel-shift high-resolution capture, expanded video codecs. Sony’s AI processing unit handles subject recognition that no third-party app could match because the model is welded to the silicon. Fujifilm’s film simulations have become a marketing centerpiece, sold as software identity rather than hardware capability. Each of these is exactly the kind of evolution the open-OS thesis predicted — except delivered by the incumbents, on their terms, inside their walls, with the upgrade cadence and roadmap entirely under corporate control.
The smartphone is the place where the software-defined camera actually shipped. Computational photography — multi-frame fusion, neural denoising, generative fill, scene-aware HDR — all of it runs on iPhones and Pixels and the Chinese flagships, on operating systems that are closed in every meaningful sense but open enough to host third-party imaging apps. That is the partial vindication of the original thesis: the SDC won, but it won as a phone, not as a camera body, and it won inside two operating systems controlled by the largest companies on earth. The flexibility came; the openness did not.
Lens mount licensing has loosened — Sigma and Tamron now ship autofocus lenses for Sony E and Nikon Z under contractual terms — but this is industrial cooperation, not platform openness. No camera body in 2026 lets the user install a third-party autofocus algorithm or a custom raw pipeline. None is on the roadmap.
The thesis was that software would eat the camera and that the eating would be done by an open ecosystem. Half of that turned out to be true. Software did eat the camera. The eating was done by the same three companies that dominated in 2024, plus one cinema acquisition, plus the entire smartphone industry. The open ecosystem never showed up, and at this point there is no structural reason to expect it to.
