People are rarely surprised by a screen anymore.
TVs have grown from 55 inches to 85 inches, smartphone screens are getting brighter, monitor refresh rates have surpassed 144Hz and 165Hz, and laptops are becoming thinner and lighter. These changes are happening right before our eyes, but screens have become so familiar that people gradually take them for granted. Yet, if all the screens were taken away one by one throughout the day, modern life would probably quickly lose its order.
Familiarity does not equal stagnation. According to Omdia data, global shipments of large-size display panels increased by 2.9% year-on-year in 2025, with monitor OLED and notebook OLED shipments growing by 69.3% and 26.8%, respectively. While the total volume is stabilizing, new materials and form factors are still seeking their markets.
Recently, Semiconductor Industry Insight interviewed Dr. Luo Kun, CEO of Shanghai HiSilicon Display Core. During the nearly one-hour conversation, he rarely described displays as a mere parameter competition, but rather focused on whether they could be more human-centric, more energy-efficient, and integrated into real life alongside AI. The imagination for the next generation of displays may lie not only within the screens themselves but also in the space between the screens and people.
01. Displays: Ubiquitous
From the application perspective, if we only look at TVs and smartphones, the display industry seems to have entered a mature stage: replacement cycles are lengthening, smartphone growth is slowing down, and the old path of relying on production line expansion, area increase, and pixel density is becoming increasingly narrow. From the market perspective, China has become a major hub for the global display industry. According to Omdia statistics, in 2025, BOE's share of global large-size LCD panel shipments was approximately 36.2%, and TCL CSOT's was about 16.4%, with the two Chinese companies combined accounting for over half. After establishing scale advantages, the questions change accordingly: where will the value for the next stage come from?
The answer is no longer just about producing more panels, but about adapting displays to more segmented scenarios. Following this shift, Luo Kun categorizes screens into 'large, medium, small, and micro': TVs and conference screens are large screens, computers and tablets are medium screens, smartphones are small screens, and AR glasses are micro screens. This is not just a division by size, but also corresponds to different distances and relationships between the screen and the user. Large screens face families and public spaces, medium screens support office work and creation, small screens are worn close to the body and used with high frequency, while micro screens are close to the human eye. With different demands, LCD, OLED, and Micro LED continue to evolve in their respective suitable scenarios.
Luo Kun said, 'Screens are not just ubiquitous; they are with us every moment.' Screens will continue to get larger, brighter, and faster, but the growth of the display industry is branching out from a neat parameter curve into many specific life demands.
02. The Next Screen May Not Look Like a Screen
In 2007, multi-touch entered smartphones. Swiping, tapping, pinching to zoom in and out with fingers meant no need to learn the machine's logic; for the first time, haptics entered digital interaction so naturally. Luo Kun believes that the explosion of smartphones came not only from increased computing power but also from a more human-centric interaction method. For the first time, the screen connected vision and touch so closely.
Physical AI is bringing the same question to more devices. Robot vacuums map through LiDAR, identifying pets and obstacles; NEVs use cameras, radars, and sensors to accomplish automatic parking and assisted driving. They are not just executing commands, but first perceiving the environment and then reacting. This is how physical AI begins to enter our lives.
Luo Kun pins the next step of physical AI on a very specific word: 'the five senses.' Centered around an individual's eyes, ears, mouth, nose, and body, corresponding to vision, hearing, taste, smell, and touch, this will become the direction for continuous investment in various innovations. Smart terminals will no longer just answer a question or present a picture; they need to understand what a person is looking at, hearing, where they are, and what they need next.
Glasses might be the first place where this change appears. Luo Kun describes AR in a very down-to-earth way: 'Glasses are still glasses, but when you need them, they will give you a prompt.' Navigation arrows, a line of prompt for a speech, or a bit of background information when meeting someone for the first time can all briefly appear in the line of sight, without trapping people in another screen. Counterpoint data shows that global shipments of AR smart glasses increased by 98% year-on-year in 2025; in the second half of that year, the market share of waveguide products rose from 13% in the same period last year to 38%. During the same period, global Micro LED display revenue grew by 150%, with AR glasses contributing 58%. Micro LED did not first scale on TVs or smartphones; instead, it found a commercial entry point from a micro screen close to the human eye that requires high brightness and low power consumption. Luo Kun said, 'Do not change the form of glasses; glasses are just glasses.' This sentence draws the technical boundary: the display must be bright enough, but also light enough and power-efficient enough, so that added features do not destroy the original look of the glasses. Just as the Shanghai HiSilicon AR micro-display solution also focuses on environmental adaptation: it must be clearly visible outdoors, while avoiding excessive brightness indoors; the pursuit is not simply to be brighter, but for the display chip and algorithms to adapt to the environment.
Automobiles provide another kind of change. When talking about NEVs, Luo Kun specifically mentioned smart cockpits, assisted driving, and automatic parking. What they change is not just the way of driving, but also the way information appears around people. Instrument panels, central controls, head-up displays, and rear-seat screens are turning the cockpit into an information space synchronized with humans. Omdia estimates that the global automotive display panel market revenue will grow from USD 13.6 billion in 2025 to USD 18.3 billion in 2030; behind this growth is the penetration of high value-added technologies such as OLED and LTPS.
Looking further into the future, screens may enter mirrors, desktops, and home furnishings. In the interview, Luo Kun imagined a smart makeup mirror: it knows who is standing in front of the mirror and what kind of makeup the person wants, and can call up a makeup artist or hairstylist model suitable for the individual to guide them step by step on the mirror surface. He also mentioned smart desktops, where as a child writes, the system understands their learning progress and provides timely assistance.
So is there a more direct and more comfortable display? The technical prerequisites given by Luo Kun are ultra-low power consumption, thinness and lightness, and 'not changing the form of current products, but naturally integrating into them'. A screen does not necessarily have to appear in the form of a piece of glass; it can be hidden in the line of sight of glasses, in front of a car, or on the surface of a piece of home furniture.
The next screen may not look like a screen; it is more like a perception and response system revolving around the individual.
03. Visible Screens, Invisible Technologies
Imagining new form factors is one thing; turning technology into mass-producible products is another. Luo Kun uses R and D to explain the two timelines of Shanghai HiSilicon: D stands for Development, solving today's products; R stands for Research, paving the way for the more distant future.
The DDR-free solution is a typical case of D (i.e., Development). In high-refresh-rate monitors, external DDR is used to store intermediate data generated by image processing and panel compensation. Luo Kun introduced that during the observed supply shortage phase, the price of 1GB DDR rose from one or two US dollars to twenty or thirty US dollars, while the main display chip was about five or six US dollars. Memory changed from a supporting role to a prominent variable in the overall machine cost. Removing DDR cannot just mean taking a component off the PCB. The materials and processes of the panel production line have been upgraded: conductive materials have higher purity, electron mobility has improved, wiring can be narrowed, and the aperture ratio increases accordingly. But if the chips, algorithms, and whole machines still use old designs, these advances will, as Luo Kun put it, 'remain silent in the production line.' The DTCO adopted by Shanghai HiSilicon puts the chip back into the system for redesign. The chip company, together with panel factories, monitor factories, and TV manufacturers, looks from panel materials, color filters, wiring, and aperture ratios to compensation algorithms, chip structures, and whole machine requirements. Luo Kun took the cooperative project where the screen aperture area increased by nearly 25% as an example: after the materials and structures changed, color, contrast, and compensation algorithms also had to be redone, thereby reducing the process data that needs to be written into DDR, ultimately enabling some high-end monitors to achieve DDR-free. It is understood that the two related products launched by Shanghai HiSilicon last year were originally expected to be only in the hundreds of thousands in volume, but later entered the million-level a year ahead of schedule. The rise in memory prices accelerated the volume ramp-up of the DDR-free solution, but the capability it truly demonstrates is whether chips, panels, and whole machines can change together for the same product goal.
The DDR-free solution is not an isolated case; we also see two other highly popular cases. One is that in the Zhuque RGB Mini-LED solution released by Shanghai HiSilicon, an edge-side AI engine is used for pixel-level control of brightness, color, and consistency, and technologies such as three-color light source consistency compensation and color correction are used to handle the crosstalk and color overflow of RGB backlighting. The potential of backlight technology still needs to go through chips and algorithms to be transformed into mass-producible picture quality. The other is that the tri-mode display applies system capabilities to different human tasks. Shanghai HiSilicon summarizes the three scenarios very clearly: esports focuses on refresh rate, audio-visual balances smoothness and picture quality, and professional creation relies on resolution and details. Luo Kun gave a more specific description in the interview: the FHD esports mode can achieve a refresh rate of 1000 times per second; the creation scenario pursues details of 4K and above, or even 6K to 8K level; movie watching finds a balance between 4K and about 180 to 200 frames. This solution originally came from a customer exchange. A user who does both game design and plays esports asked, since the desktop is only so big, whether one screen could cater to both creation and gaming, while also accommodating family members watching videos. The tri-mode screen does not push all parameters to the maximum at the same time, but reallocates performance among different tasks. Luo Kun's standard for evaluating products is very simple: 'Whether users like it after using it, and whether they cannot do without it.'
04. Screens, Growing Towards the Future
The more mature an industry is, the more easily innovation is understood as marginal improvements in parameters. But the changes happening in displays precisely require chips to go beyond parameters and continue to ask where the experience comes from. Luo Kun uses sports fabrics as an analogy. Today, some synthetic fiber clothes can wick sweat and breathe, not just because the materials are changed, but because after the fibers are fine and tough enough, the warp and weft and spatial structures can be redesigned. For the user, what the material is called does not matter; what matters is whether it is comfortable to wear.
Displays are the same. Eye protection cannot stop at adding a functional mode; it must continue to ask: why does vision get fatigued, how does human vision age normally, and what kind of stimulation will cause damage? Research thus enters visual principles, optics, biology, and medicine, and then in turn determines how chips and display systems should work. This is the 'root technology' Luo Kun talks about: it does not start from what functions can be added to a chip, but from what problems people truly need to solve.
Following this line of thought, Shanghai HiSilicon summarizes its future direction as '3+1': green energy saving, AI real display, healthy eye protection, and light intelligence. Green energy saving determines the battery life, thin and light form factor, and continuous working capability of edge-side devices; AI real display and healthy eye protection target the user's visual experience; open light intelligence relies on the compatible adaptation capabilities of chips and sensors to support various AI new scenarios still being explored. According to the implementation path proposed by Luo Kun, the core components will first achieve low power consumption, low cost, and small size, and after the products reach scale, they will continue to compress size and cost through higher integration.
What this technical path ultimately needs to answer is still how the industry can find new growth. After TVs and monitors enter the mature market, repeatedly lowering prices around similar products will only lock the competition into original demands. Luo Kun summarized it as: 'Anti-involution relies on innovation, innovation relies on demand migration, and relies on the migration of future demands and industrial development.'
Demand migration cannot be accomplished by a single enterprise alone; it also requires standards to turn new capabilities into industry consensus. According to Luo Kun, building on the relevant national standards in 2020, China is promoting the update of energy consumption standards for TVs and monitors according to technological evolution; Shanghai HiSilicon is also participating in the discussion of national and industry standards for low energy consumption, healthy eye protection, AI adaptation, and interference protection. Standards are not just about drawing a qualification line; they also guide the industry to invest technology and resources into next-generation products.
For China's display industry, which has already established scale advantages, this may be a more important change than refreshing another isolated parameter: competition is no longer just about who sells the same screen cheaper, but about who can discover new demands earlier and integrate materials, chips, and whole machine capabilities into the next round of growth. Facing the larger market demand for AI, Shanghai HiSilicon has also been collaborating with partners to create innovative display applications.
At the end of the interview, Luo Kun said, 'Everything we have done before is paving the way for the future and preparing for the future in advance.' The next leg of the display journey is illuminated not only by larger sizes and higher parameters but also paved by these invisible technologies. The future of displays is hard to predict, yet worth looking forward to!