TCL CSOT's Zhao Jun: Printed OLED Yield Rises to About 70% in Half a Year; in the AI Era, We Don't Want to Make Just a Screen
TCL CSOT CEO Zhao Jun revealed that about half a year into mass production of printed OLED products, yield has risen from single digits in the initial production phase to about 70%, and capacity at the Wuhan t12 line has increased from 3K large substrates to 9K large substrates. The t8 8.6-generation line had production equipment moved in on October 31 and is planned for mass production in the fourth quarter of 2027. Zhao Jun said that in the AI era, display companies need to understand end devices earlier, participate in product definition, and expand into AI infrastructure such as glass-based packaging.
Body · AI translation
TCL CSOT's Zhao Jun: Printed OLED yield rises to about 70% in half a year; in the AI era, we don't want to make just a screen. Ifeng Technology Download the app for exclusive early access. November 14, 2026, 17:27:56 From Beijing. Updated November 14, 2026, 17:24:27
Author / Liu Yukun
Editor / Yang Ruiqi
Ifeng Technology News, November 14 — About half a year after a printed OLED product entered mass production, its yield has risen from single digits at the start of production to about 70%; capacity at the Wuhan t12 line has increased from 3K large substrates to 9K large substrates; and only after 13 years of technology R&D did TCL CSOT decide to build an 8.6-generation line. Behind these three numbers is this panel maker's weighing of technology routes, mass-production feasibility and investment pace.
On November 12, on the eve of the opening of the 2026 TCL CSOT Global Display Ecosystem Conference (DTC2026), Zhao Jun, senior vice president of TCL Technology and CEO of TCL CSOT, gave an interview to media in Wuhan, responding to questions on printed OLED mass production, OLED supply and demand, AI terminals and panel price cycles.
After leading in scale, where are the shortcomings?
China's display industry has achieved leadership in scale, but scale does not equal full say. Zhao Jun summarized the support for say and product definition rights as capacity, market share, profitability and technological leadership. He said mainland China's LCD capacity already accounts for more than 70% of the global total, and "we have firmly grasped the dominant position in the LCD industry."
OLED presents a mixed picture. In the small-size FMM OLED (evaporation OLED) field represented by mobile phones, Chinese companies' capacity and shipments are on par with South Korean companies, but there is still a gap in technology, product competitiveness and recognition by top global customers. Printed OLED focuses more on medium sizes; in new display fields such as Mini LED direct display, Micro LED and other MLED technologies and printed OLED, Zhao Jun judges that Chinese and foreign companies are basically at the same starting line, and Chinese companies' future growth opportunities and competitiveness are more cause for optimism.
The real areas where catching up is needed remain upstream in the industry and in the ecosystem. "Overseas companies, especially South Korean ones, have built capabilities in intellectual property, key materials and key equipment in the OLED industry that are worth our learning and reference," Zhao Jun said. He also stated frankly that Chinese companies remain relatively weak in industry standards.
Printed OLED: answer the mass-production question first
TCL CSOT is investing in LCD, FMM OLED, printed OLED and MLED at the same time. Zhao Jun's resource allocation logic is to let mature businesses such as LCD continue to contribute profit and cash flow and form the operating foundation, then provide investment resources for growth businesses. Over the past two years, this approach has produced results. The role of a multi-business portfolio is to reduce the impact of fluctuations in a single market on operating stability.
Technology choices face three thresholds: whether there is a real demand scenario, whether demand has enough room to grow, and whether it can be commercialized. "If this technology is purely for showing off, purely for self-indulgence in specs or parameters, then it is better not to do it," Zhao Jun said. He disclosed that in recent years the company has cut many projects internally, an important reason being the lack of application scenarios.
Printed OLED's path from R&D to the t8 decision followed exactly this validation path. Zhao Jun said TCL CSOT has been working on printed OLED for 13 years and only in the 12th year did it decide to invest in an 8.6-generation line. The key was not only that the technology was becoming mature, but also mass-producibility, manufacturability and large-scale mass-production capability. Only after t12 achieved stable mass production and yield reached a certain level did the company preliminarily judge t8 feasible.
At present, t12's first product entered mass production in April, and yield has risen from single digits to about 70%; multiple products are in mass production at the same time, two of which achieved mass-production shipments two to three months ago. Several top customers have released positive signals and specific product requirements. For the 27-inch 500Hz printed OLED gaming display, Zhao Jun said it is still being developed and refined, and the launch time will need to be arranged according to brand customers' product plans.
t8 construction is also advancing. On October 31, the line officially moved in production equipment, shifting from factory construction to equipment move-in and mass-production preparation, with mass production planned for the fourth quarter of 2027. Zhao Jun said the launch of t8 is a turning point for ecosystem cooperation, with upstream equipment and material partners and downstream customers shifting from a wait-and-see attitude to greater confidence, but specific names were not disclosed due to commercial agreements.
OLED competition is "brutal"; where will growth come from?
The OLED market currently faces pressure from both supply and demand. On one hand, soaring memory prices are suppressing end demand; on the other, new capacity is being released gradually. Zhao Jun described the current competition as "in a relatively brutal state." But he also listed possible sources of demand recovery: memory prices returning to rationality, OLED penetrating entry-level products, and the expansion of medium-size applications such as automotive and tablets.
This is also the backdrop for TCL CSOT still launching t8. Zhao Jun believes that, besides picture quality, printed OLED has significant advantages in material cost, investment cost and operating cost, and also features eye protection and environmental friendliness. "As long as we bring out the competitiveness of printed OLED and take it to the extreme, no matter what changes occur in the market, we will have the confidence to face the challenges head-on."
He did not shy away from the gap between the advantages of the principle and those of the product: "To translate its advantages in principle into technological advantages and product advantages, a great deal of work is still needed." For medium-sized OLED, he listed three routes—evaporation OLED, printed OLED and photolithographic OLED—and regarded competition among different routes as benign competition, to distinguish it from the homogenized imbalance caused by a rush of capacity expansion in the same technology and similar products.
Foldable screens are another incremental observation point. Zhao Jun said Apple's entry is an "accelerator" for foldable phones. Market information indicates that global demand for foldable phones in 2026 will be about 28 million units, up about 36% year on year; it is forecast to reach 46 million units by 2031. Beyond large foldables that fold left-right and small foldables that fold up-down, domestic brands have previously launched wide-fold products. Market forecasts say the share of wide folds among foldable phones will rise from the current 50% to 70% by 2031.
When foldable phones reach more than 40 million and approach 50 million units, Zhao Jun judges that they will change from a novelty product into a mainstream model. Zhao Jun believes Apple's crease handling is relatively good; in response to this pain point, TCL CSOT already has technology to reduce creases and will work with customers to drive mass production.
Zhao Jun divides AI terminals into three categories: traditional terminals adding AI capabilities, such as AI PCs, which place higher demands on interaction and low power consumption; native AI terminals whose form is unchanged but whose substance has changed—he cited Nvidia's PCs based on the Spark platform, which can run large models locally; although powerful with high computing specifications, they are also very expensive, and the first batch of products all use OLED screens; and AI-native terminals with fully updated form and substance, represented by AI glasses with screens, commonly known as AR glasses, which may require entirely new technologies such as silicon-based Micro LED.
Corresponding to products, ultra-wide screens are form innovations using relatively mature technologies such as LCD or OLED. Beyond mainstream 16:9 screens, users need extra display area to converse with agents. Zhao Jun judges that first-generation native AI PCs will basically use OLED screens, the second generation may see LCD join in, and the display-quality requirements of high-end native AI PCs provide an opportunity for printed OLED.
Glass-based packaging, meanwhile, jumps out of displays and enters AI infrastructure. Its foundation is not building from scratch, but the spillover of TCL CSOT's capabilities in large-size glass-based micron-level processing, glass wiring and flatness control, while also requiring new capabilities such as glass drilling. Zhao Jun summed it up as: "From an interface for intelligent terminal displays to a link in the infrastructure."
Conversely, AI is also transforming display companies. Zhao Jun introduced that TCL CSOT released the Star Intelligence vertical-domain large model three years ago. He said Star Intelligence's performance in authoritative tests exceeds similar vertical-domain models and the top general large models. The company's agents can already identify complete or vague customer needs and translate them into internal standardized design guidelines, with accuracy exceeding 90%, reaching the level of a relatively mature engineer. The next step is to support automated design with AI, agents and dozens of software tools; the hardest part is connecting the design solution with the manufacturing system.
At DTC2026 held on November 13, TCL CSOT showcased "silicon-based employees" across the entire display R&D and design process, integrating large models, multi-agent systems and professional engineering tools, covering solution planning, design, simulation, verification and quality management; achievements including a 27-inch 500Hz printed OLED esports display and a 14-inch printed OLED ultra-thin, lightweight, low-reflection notebook display were also unveiled. The latter is less than 0.66 mm thick, weighs less than 66 grams, and has a reflectivity below 1%.
Structural opportunities in 2027
Regarding this round of panel price increases, Zhao Jun did not define it as a new demand-driven cycle, but as cost-driven price repair. Upstream glass substrate manufacturers have proposed price increases of no less than 15% in the fourth quarter; some capacity and resources for PCB, IC and passive components have been squeezed by AI demand; in the first three quarters, TCL CSOT's material costs rose significantly, and combined with exchange rate factors, the company needs to release some cost pressure.
Zhao Jun expects that in 2027, total demand for large-size panels represented by TVs will decline slightly, but average size growth will drive a slight increase in area, with global display panel shipment area possibly growing 1%. Structurally, shipments of high-end TVs priced above $1,000 in North America may grow about 9%; some Taiwanese manufacturers selling factories and transferring capacity may also reduce supply. Small-size FMM OLED is in a stage of bottoming out and consolidating; foldable phone growth and OLED penetration into entry-level products are worth watching, and the biggest variable remains memory prices.
"On the one hand, we are not blindly optimistic about the future, but there is also no reason to be blindly pessimistic." Zhao Jun summarized 2027 as a relatively weak balance in total volume with opportunities in some segments.
Cumulative investment in Wuhan exceeds 80 billion yuan
Industrial opportunities also need manufacturing and ecosystem support. TCL CSOT has laid out two LTPS production lines, t3 and t5, the t4 evaporation OLED production line and the t12 printed OLED 5.5-generation line in Wuhan, with cumulative investment exceeding 80 billion yuan, making Wuhan its second-largest investment cluster after its Shenzhen headquarters. What Zhao Jun values is the industrial foundation formed by the "Optics Valley" around "light, chips, screens, devices and networks," full upstream and downstream chain support, and the concentration of university talent.
The Wuhan New Display Industry Innovation Joint Laboratory, approved for construction in August 2024, is led by Wuhan China Star Optoelectronics and jointly with Jingli Electronics, Jingwei Technology, Huarui Optoelectronics, Huazhong University of Science and Technology and others, carrying out chain-based research on display devices, new materials, testing equipment and automotive displays. After three months of trial operation, the world's first mass-produced 21.6-inch 4K printed OLED professional display was delivered in Wuhan.
Choosing technology from real scenarios, supporting long-term investment with mature businesses, and then turning principle advantages into product competitiveness through mass production and the ecosystem—this is the path Zhao Jun laid out. The changes brought by AI are not just continuously rising screen specifications, but that display companies must understand terminals earlier, participate more deeply in definition, and extend toward R&D, manufacturing and even AI infrastructure. After leading in scale, the answer to the next round of competition lies not only in capacity tables, but also in real scenarios, mass production curves and the industrial ecosystem.