Introduction

In June in Taipei, COMPUTEX 2026 prominently displayed “AI Together.” Unlike last year when discussions centered around “AI NEXT” with a focus on the next phase of technological breakthroughs, this year’s exhibition conveyed a significantly different key signal: artificial intelligence is no longer just a narrative around large models, servers, and computing chips, but is beginning to integrate into robotics, automation, edge computing, smart spaces, and real business processes.

For industry observers, this means that the center of AI discussion is shifting from “whose model is stronger” to “who can truly deploy AI into machines, systems, and organizations and achieve scaled delivery.”

On the same day, signals that echoed this trend appeared on the other end of the market. On June 5th, well-known investor Serenity publicly stated that Green Harmonics (688017) was one of the most promising Chinese listed companies in the humanoid robot sector, and based its judgments about the mass production logic of “physical AI” around core components such as harmonic reducers, rotary joint reducers, linear actuators, motors, joints, and planetary roller screws. This viewpoint quickly gained traction in the market, with the stock price of Green Harmonics experiencing significant fluctuations that day. The two threads of the exhibition scene and the capital market formed a representative intersection on the same day: as AI begins to be integrated into robots and automated systems, the focus of industry discussions has shifted from the capabilities of models themselves to the entire machine manufacturing, supply chain collaboration, and the distribution of value in key components.

Summary

• COMPUTEX 2026, under the theme of “AI Together,” further strengthened multiple directions such as AI infrastructure, edge computing, robotics, and intelligent mobility, with the exhibition featuring 1500 vendors and 6000 booths, setting a new record.

• Many exhibiting companies focused their displays not only on generative AI demonstrations but further extended into smart spaces, intelligent manufacturing, collaborative robots, smart parking, edge reasoning, edge AI devices, and enterprise scenario deployment.

• Compared to 2025, the most notable change in COMPUTEX 2026 is that the narrative of the exhibition has shifted from “AI's next step” to “how AI is collaboratively delivered and enters industrial systems.”

At the same time, market discussions surrounding Green Harmonics reflect a shift in the capital market's focus on the robotics industry chain. According to publicly circulating statements from Serenity, its core perspective does not solely rest on the sentiments of a single company, but revolves around the logic that “if humanoid robots enter mass production, which companies capable of stably entering the single machine BOM and possessing multiple core component layouts will benefit from the industry's scaling.” Whether this judgment can be fully realized remains to be seen, but it corresponds to the current most pressing questions in the robotics sector: if Physical AI transitions from proof of concept to mass production, where will the industrial value truly settle.

AI Together

From the exhibition perspective, COMPUTEX 2026 is an upgraded continuation of the Taipei International Computer Show, with the official theme set as “AI Together,” clearly focusing on AI computing and robotics, next-generation technology, and future mobility directions. Public data shows that this exhibition attracted 1500 vendors from 33 countries and regions, utilizing around 6000 booths, a scale larger than the previous edition. Besides continuing the high interest in AI computing, this year's exhibition introduced new features such as an AI robotics area and technology application and experience hall, indicating that the organizers attempt to extend AI from chip and machine-level displays into richer industrial applications and experiential scenarios.

From the content presented by companies, many exhibitors this year have downplayed the simple “showing models” or “showing parameters” approach, focusing instead on landing capability, scenario deployment, and system collaboration. For instance, MediaTek showcased its AI layout from edge to cloud in this exhibition, linking robotics, traffic, smart homes, personal devices, and data centers into the same narrative framework; InnoVEX 2026 listed “AI commercialization” as one of its key directions, mentioning manufacturing, robotics, enterprise automation, and next-generation computing scenarios multiple times in their public descriptions. In other words, the most noteworthy change in this year's COMPUTEX is not the increase in AI labels, but the clear integration of AI into deliverable, deployable, and mass-producible industrial and commercial systems.

From the market perspective, discussions surrounding Green Harmonics on June 5th also attracted attention. Many cryptocurrency and financial news platforms reposted Serenity's assessment of the company, mentioning that Green Harmonics' business covers harmonic reducers, humanoid robot rotary joint reducers, linear actuators, motors, joints, and various core components, and is entering the planetary roller screw arena. The publicly shared content also indicated that its view on Green Harmonics is based on a distinct hypothesis: once humanoid robots move towards mass production, companies that possess core component entry capabilities and occupy a certain proportion of single unit value will likely continue to benefit.

It should be noted that claims regarding “over 60% domestic market share” and “more than 1800 global clients” are currently widely circulating and primarily stem from Serenity's related articles and their secondary reposts. In publicly available research materials and industry reports at different times, various statistics exist regarding Green Harmonics' market share, client coverage, and sustainability, showing varying statistical bases and phased changes. Therefore, for formal news writing, a more prudent approach would be to clearly attribute these figures to “according to Serenity's articles” or “according to publicly circulating information” while also reminding readers that the relevant industry data still requires cross-verification with the company's regular reports, brokerage research, and subsequent order validations.

Industry Diffusion

When looking at COMPUTEX 2026 in conjunction with the current capital market's focus on the robotics chain, a clear resonance emerges: the AI industry is transitioning from an upstream competition centered around “computing” to a midstream and downstream diffusion focused on “deployment.” In the past two years, the focus of the capital market and media coverage was mainly on GPUs, HBM, high-speed interconnects, server complete machines, and large model manufacturers; however, by 2026, with edge AI, Agentic AI, robotics, and automation applications being prominently showcased at the exhibition, the market began to seriously discuss which hardware, systems, and components will truly benefit in the next stage.

This is also why the robotics sector appears particularly special in discussions this year. Unlike generative AI software, for humanoid robots to truly enter large-scale deployment, advances are needed not only in algorithms related to perception, decision-making, and control but also in the collaboration of overall machine design, actuators, reducers, screws, motors, joints, sensors, batteries, wiring harnesses, connectors, and manufacturing yields. In other words, Physical AI is not simply a software upgrade at a point; it is more like a system-wide battle that spans algorithms, engineering, manufacturing, and supply chains. This system nature resonates with the exhibition theme of COMPUTEX 2026, “AI Together.”

From this perspective, the most important industrial significance of this year's exhibition may not be to prove once again that AI is hot, but to signal the market that AI is beginning to “grow a body.” When AI enters robotics, industrial control, smart spaces, and operational systems, the methods of value creation it corresponds with will also change. Models and chips remain important, but they are no longer the only focus in the value chain; whoever can encapsulate these capabilities into real, operational machines and systems is more likely to secure a position in the industrialization process.

For investors, a direct result of these changes is that the observation framework needs to migrate in tandem. In the past, focusing on the AI industry might have involved looking more at computing power demand, chip iterations, and cloud vendor capital expenditures; however, as the narrative around robotics and automation heats up, the market will start to break down single machine BOM, component barriers, mass production feasibility, cost reduction paths, and customer certification cycles. Serenity's public statement regarding Green Harmonics quickly sparked discussions because it touched upon this new valuation issue: if future humanoid robots are not just laboratory prototypes but gradually move towards large-scale shipping, will those component companies capable of stably entering core sections hold higher weight in market re-evaluations.

However, the popularity of exhibitions and market sentiment cannot automatically translate into industry realization. Whether it is robotics complete machine manufacturers, core component suppliers, or AI companies providing algorithms and operating system capabilities, the true determinants of industry rhythm remain orders, costs, yields, stability, and customer adoption speed. Especially in the humanoid robotics field, there is often a long engineering and commercialization distance between prototype demonstrations and large-scale commercial use. COMPUTEX 2026 has revealed that this path is being accelerated, but this does not mean that all narratives will be validated in the short term.

Conclusion

Looking back at this year’s COMPUTEX, one of the most memorable changes is that the focus of AI exhibitions has shifted. It is no longer merely a stage for showcasing chip performance, model capabilities, and laptop upgrades, but increasingly resembles a window observing how future industrial systems are being restructured. From AI infrastructure, edge computing to robotics and automation systems, Taiwan's role in the global supply chain has thus become more critical.

The market discussions arising around Green Harmonics on June 5 further illustrate that as Physical AI becomes a higher-frequency industry keyword, the capital market is already attempting to find the key links that may embed into the mass production chain in advance. The exhibition provides the direction for the industry, while market fluctuations reflect capital expectations; together, they outline a proposition more specific than “generative AI continues to heat up”: AI is transitioning from the digital world to the physical world, and the real competition is extending from model capabilities to who can complete delivery, manufacturing, and scaling.