Sony Hands Image Sensor Manufacturing to TSMC in Landmark Joint Venture, Marking the End of Its Fully In-House Era

Industry News

by Nino LeitnerMay 11^th, 2026

Subscribe to News - Industry

Sony Semiconductor Solutions and TSMC have signed a non-binding memorandum of understanding to form a joint venture for the development and manufacturing of next-generation image sensors, with Sony as majority and controlling shareholder. CEO Hiroki Totoki framed the deal as Sony’s first move toward a “fab-light” model, a structural shift that has direct implications for the cameras filmmakers will be buying for years to come.

The agreement covers development and production lines inside Sony’s newly built fab in Koshi City, Kumamoto Prefecture, on the Japanese island of Kyushu, the same region where TSMC already operates its own Japanese chip plant via its JASM subsidiary. Sony is also a minority shareholder in that TSMC site, which gives both partners a pre-existing footprint in the area. The two companies say the joint venture is structured so that Sony’s sensor design expertise will be paired with TSMC’s process technology and manufacturing capacity, with the joint venture expected to scale in phases tied to market demand and Japanese government support. A definitive, legally binding agreement still has to be signed before the structure is final.

Why Sony is willing to share the crown jewels

For decades, Sony has been the dominant force in CMOS image sensor design and manufacturing, supplying its sensors to most of the smartphone industry (including Apple, Samsung, and Huawei flagship devices), to most of its competitors in the camera industry, and of course to its own Alpha and Cinema Line cameras. The Imaging & Sensing Solutions division generated roughly $11.8 billion in sales in fiscal 2024 and remains the largest single recipient of Sony’s R&D spending. The image sensor business is, in other words, not a sideline; it is one of the most strategically important units in the entire Sony Group.

That makes the framing CEO Hiroki Totoki used on the company’s post-earnings call notable. According to a Bloomberg report, Totoki told analysts that “the joint venture with TSMC will be our first step to becoming fab-light,” adding, “until now, we have handled everything in-house, from R&D to manufacturing, but going forward, we hope to advance manufacturing not only on our own but also by bringing in partners.” That is a meaningful change in posture from a company that has historically guarded its sensor manufacturing tightly.

The strategic logic, read between the lines, is straightforward. Modern stacked CMOS sensors (the kind that power cameras like the Sony BURANO, the a7 V, and the recently announced LYTIA 901 smartphone sensor) are increasingly difficult and expensive to fabricate. They require precise wafer bonding, on-chip logic and AI processing circuits, multiple stacked layers, and yield rates that fall as complexity grows. TSMC has spent a generation building exactly the advanced process technology and capacity that next-generation image sensors require. Bringing TSMC inside the tent gives Sony access to that capability without having to build (and finance) every new fab from scratch.

What “fab-light” actually means for Sony’s chip business

The TSMC tie-up is the most concrete step yet in a wider transition Totoki has signaled since taking over as CEO. Sony has been steering the group toward an asset-light, IP-driven structure, leaning more heavily on music catalogs, film and television franchises, and gaming, while reducing the capital intensity of its hardware businesses. The Bravia television line was previously folded into a manufacturing partnership with TCL, a similar gesture toward outsourcing production while keeping brand and design control in-house.

The image sensor joint venture applies the same logic to silicon. Sony retains the design IP, the relationships with camera and smartphone customers, and (crucially) majority control of the joint venture. TSMC contributes process expertise, capacity, and a manufacturing model the entire fabless semiconductor industry has been built on. Sony’s chip arm reported a 33 percent increase in quarterly profit in its latest results, but the unit has also been cutting costs to offset weaker demand from smartphone makers. A capital partner that can absorb part of the next investment cycle is, from that vantage point, a useful thing to have.

The geographic concentration matters too. Kumamoto sits in southern Japan, an area that has lived through the disruptions caused by the 2016 earthquake that temporarily shut down Sony’s sensor production and rippled across the entire camera supply chain. By co-locating the new joint venture with TSMC’s existing Kumamoto fab, both partners benefit from a deepening supplier ecosystem, but they also concentrate risk in a single region that is no stranger to seismic events.

AI, robotics, and where the next sensors are really pointed

The press release is unusually direct about where this partnership is aimed, and it is not primarily at filmmakers. The two companies state that the joint venture will “explore and address emerging opportunities in physical AI applications, such as automotive and robotics.” TSMC’s Senior Vice President and Deputy Co-COO Kevin Zhang made the point even more clearly in his prepared statement, calling the partnership “a key step forward in driving future sensing technology in the AI era.”

Read alongside Sony’s recent LYTIA 901 announcement, which embedded an AI remosaicing engine directly into the sensor, the direction of travel is consistent. Sony has been pushing on-chip intelligence since the IMX500 Intelligent Vision sensor was unveiled in 2020, and the Kumamoto joint venture looks like the manufacturing platform for that next generation: stacked sensors with logic, memory, and AI inference layered onto a single device, scaled across automotive, robotics, smartphones, and (eventually) cameras.

For working filmmakers, the takeaway is mixed. On one hand, more advanced fabrication should, over time, mean better sensors in mirrorless and cinema cameras, with improvements in dynamic range, readout speed, on-chip noise reduction, and global shutter implementations of the kind we’ve previously covered from Sony. On the other hand, the strategic priority is clearly being set by automotive and AI customers with much larger volume requirements than the entire global camera industry. That has historically been true of smartphone sensors as well, and it is one of the reasons our cameras have benefitted from sensor improvements paid for by phones for over a decade.

A plan still subject to change

It is worth keeping in mind exactly what was announced. This is a non-binding memorandum of understanding, not a finalized joint venture. The investment numbers are not disclosed, the phasing is contingent on market demand, and Japanese government support is being assumed rather than confirmed. Definitive, legally binding agreements still need to be signed, and the closing conditions have not been spelled out. In other words, the architecture is set, but the financials are not.

For an industry already navigating the global memory chip shortage and a wave of consolidation across manufacturing and post-production, a Sony-TSMC partnership of this scale is a clear signal that the economics of advanced semiconductor manufacturing have shifted permanently. Even Sony, the world’s largest image sensor maker, has decided it cannot, or should not, do everything alone.

Does Sony’s move to a “fab-light” model excite you, or does sharing sensor manufacturing with TSMC raise concerns? Don’t hesitate to let us know in the comments below!

Featured image illustration by CineD.