On June 28, 2026, an event titled 'AI Smart Hardware's New Opportunities, The Battle for Interaction Entry Points Begins,' hosted by ITjuzi and co-organized by Yihui Capital and Zhongguancun Science City Company, was held in Beijing. The event gathered executives from institutions such as ITjuzi, Yihui Capital, SoundAI Technology, Ling Universe, and Zhongbo Jili.
I. Report Release:
Insight Report on Newly-Emerged Companies in the AI Hardware/Smart Hardware Track
The event kicked off with a presentation by Ms. Wu Meimei, Senior Analyst at ITjuzi, who delivered ITjuzi's industry analysis report on smart hardware. The report indicated that in recent years, China's AI hardware and smart hardware sectors have experienced a wave of concentrated entrepreneurial activity.
By examining newly-established companies founded after 2023, the report summarized four major trends in this track:
1. Track heat exceeds expectations. Among 431 startup companies, 327 have already secured funding, with a financing rate as high as 75.9%. In the first half of 2026 alone, 179 companies received investment, an extremely rare consensus among capital.
2. Financing ecosystem dominated by intelligent robots, with diverse developments. Embodied intelligent robots stand out for attracting the most capital. Wearable tracks like smart rings, AI glasses, and sports/health wearables, though smaller in scale, show significant growth and considerable potential. In terms of financing structure, angel rounds account for more than half, indicating the sector is still in its early stages, the landscape is far from settled, and the window of opportunity remains open.
3. Geographical distribution: Shenzhen leads the pack, with Beijing-Shenzhen as the dual core. Shenzhen ranks first with 95 companies (22%), followed by Beijing with 44 and Shanghai with 50. China's Shenzhen possesses world-leading and unique advantages in its global-class hardware supply chain.
4. Typical cases: Micro-innovations across six sub-sectors. Smart rings are evolving from health monitoring to tactile AI interaction; AI glasses have multiple vertical scenarios like cultural/heritage guiding and outdoor imaging; AI toys incorporate emotional engines and virtual personality systems, emphasizing companionship attributes; these companies essentially apply new AI technologies to existing categories. Within humanoid robots, diverse approaches exist (interaction-focused, embodiment-focused, brain-focused), with companies like Zhiyuan taking the lead in mass production and commercialization, indicating the competition in this track has entered a fierce stage.
In summary: "The entire track is vast. As technology matures and becomes more accessible, entrepreneurs don't need to 'reinvent the wheel.' They can focus on micro-innovation, establishing barriers in specific scenarios and user groups."
II. SoundAI Technology:
Advancing Perception and Interaction through Acoustics
Ms. Huang Yunhe, Chief Product Officer of SoundAI Technology, in her keynote sharing session on 'AI Technology Penetration into Terminals Drives Perception and Interaction Intelligence Upgrade,' pointed out that AI hardware startups commonly face the dilemma of 'a hammer looking for a nail'—having the technology but struggling to find a mass-producible product carrier. She emphasized that product innovation must balance technical capabilities with market acceptance. Stage-by-stage success should be based on balancing current technological thresholds, user needs, cost control, and form-factor micro-innovation, rather than blindly pursuing disruptive innovation.
Current voice interaction still relies on wake-up commands, and true multi-turn natural interaction has not yet been fully realized. Ms. Huang Yunhe stated that future AI hardware needs to rely on sensor upgrades and the development of physical AI to achieve wake-up-free, proactive perception-based interaction. Furthermore, innovation should be rooted in local demands to create China's next-generation AI terminal products.
III. Roundtable Dialogue:
Observations and Insights from Smart Hardware Entrepreneurs and Investors
Ms. Huang Yunhe, CPO of SoundAI Technology; Ms. Liu Cuixue, Senior Product Manager at Ling Universe; and Ms. Liu Hongyan, Brand Strategy Head at Zhongbo Jili, participated in a roundtable discussion themed 'AI Smart Hardware's New Opportunities, The Battle for Interaction Entry Points Begins.' The discussion was moderated by Mr. Wang Simeng, Director at Yihui Capital.
The following is the content of the roundtable dialogue:
Wang Simeng, Director at Yihui Capital: In the past, when talking about interaction entry points, everyone would first think of smartphones, think of Apple, Samsung, HMOV (Huawei, Xiaomi, OPPO, vivo), and other major players. In the AI era, I'm curious to know everyone's thoughts: will future interaction entry points still converge towards a unified super-hardware, or will they fragment into more specialized, scenario-specific hardware? Against this new backdrop, what will be the competitive relationship between major players and startups?
Huang Yunhe, CPO of SoundAI Technology: Unified terminals will still occupy an important position, but vertical function carriers for specific scenarios will also continue to emerge. Currently, there are two development paths for AI terminals—matching hardware carriers for AI, or empowering scenario-specific hardware with AI capabilities. Startups can define their product form based on multiple factors such as target audience, technical capabilities, and market environment. When market demand is clear and user willingness to pay exists, there's no need to wait for technology to be perfect; they can enter the market first for validation.
Liu Cuixue, Senior Product Manager at Ling Universe: Startup teams can, in the short term, choose more niche scenarios and more vertical user groups to accumulate sufficient data for product iteration. Regarding hardware selection, they can examine whether a particular product category's user mindshare is already occupied by leading brands. If choosing forms like watches or rings where strong players already exist, users have high expectations for basic experience. It's necessary to first excel in basic functions and then layer AI capabilities as differentiation, rather than competing on intelligence from the outset.
Liu Hongyan, Brand Strategy Head at Zhongbo Jili: The future AI hardware field might form a multi-layered ecosystem. While major players have advantages in ecosystem, supply chain, and resource aggregation, their reach may not cover all vertical areas. Entrepreneurial companies can, based on their own resource endowments and core strengths, formulate medium- to long-term strategic plans. After selecting the right track, they can persistently deepen their efforts in hardware-software synergy and product commercialization. It is entirely possible to stand out through a single point of breakthrough.
Wang Simeng, Director at Yihui Capital: Currently, the competition for entry points appears on the surface to be a contest between hardware products, but at a deeper level, it's a competition of operating systems, interaction algorithms, and user data. Regarding the barriers for AI hardware, are they in software or hardware? From a commercialization perspective, is it better to first deploy hardware to capture user mindshare, or to refine interaction technology before mass-producing products?
Liu Cuixue, Senior Product Manager at Ling Universe: The main force in current AI hardware entrepreneurship comes from individuals with backgrounds in large central state-owned enterprises and ByteDance. Their financing logics differ, but they face common challenges in product implementation: one is a deep understanding of AI models, and the other is practical hardware experience—how to navigate non-transparent information like hardware selection and cost differences. Regarding business models, AI hardware can explore models where hardware isn't profitable but relies on subscription revenue. However, the prerequisite is simultaneously controlling hardware BOM (Bill of Materials) costs and model inference costs; otherwise, losing money on both ends is unsustainable.
Liu Hongyan, Brand Strategy Head at Zhongbo Jili: The core of investment lies in verifying the landing carrier and application scenarios for AI technology. While product perfection is important, whether the market will accept it, whether sales can take off, and whether cash flow can be sustained are fundamental. Currently, AI companies generally burn cash. The key is to find the right entry scenario, adapt to the market through edge-side computing power and hardware-software synergy, ultimately achieve sales growth and sustainable development, and complete the commercialization loop from technology to scenario—this is what truly matters.
Wang Simeng, Moderator from Yihui Capital: Ms. Huang mentioned return rates earlier. I'd like to ask Ms. Huang and Ms. Liu about the average return rate in our industry. What is the return rate range for a relatively successful hardware product?
Huang Yunhe, CPO of SoundAI Technology: Regarding the current return rate situation in the earphone industry, it's reported that the average return rate for AI earphones reaches 30% to 50%. Overseas, it might be lower due to concerns about return costs.
Wang Simeng, Director at Yihui Capital: As large language models are now moving to the edge side, many local hardware devices are upgrading from passive response to actively perceiving user needs. So, I'd like to ask everyone to envision: what new hardware categories will proactive human-computer interaction give rise to? In the next 2-3 years, where will the biggest industrial dividends for domestic AI smart hardware be concentrated?
Huang Yunhe, CPO of SoundAI Technology: Due to large model hallucinations, lack of physical constraints, and insufficient environmental perception, proactive interaction technology is not yet mature in actual user experience. Only health wearable devices have achieved limited proactive reminders. The industry should learn from the early internet experience: find specific user groups and amplify strengths rather than dwell on universally existing technological limitations. Chinese entrepreneurs possess unique advantages in marketing, product definition, etc. The core is to find target users, establish a coherent user experience, first achieve market traction and iterate gradually, patiently waiting for technological maturity and inflection points.
Liu Cuixue, Senior Product Manager at Ling Universe: Proactive interaction is an important evolution direction for AI hardware. Devices judge opportune moments through multimodal perception (visual, auditory, physiological signals, etc.) to initiate meaningful interactions and build user understanding and emotional connection. Achieving this capability requires solving two core problems. First is timing judgment: to "speak when it's time to speak," avoiding disturbance. Second is long-term memory: accumulating user preferences and interaction history to make interactions increasingly natural. Engineering implementation is a current challenge, especially in balancing edge-side inference capabilities and privacy protection. Hardware forms need to vary by user group; for example, women value appearance, while children need lightweight, easy-to-operate devices.
Wang Simeng, Director at Yihui Capital: I'd like to ask Ms. Liu from Zhongbo Jili: currently, the financing heat for AI hardware remains high, and trendy projects emerge one after another, but bubbles are also appearing simultaneously. From an investor's perspective, what core criteria would you focus on to select quality projects and avoid homogeneous bubbles?
Wang Simeng, Director at Yihui Capital: Please, three guests, summarize in one sentence: in this battle for interaction entry points, what is the most critical key to victory for a company to ultimately succeed?
Liu Cuixue, Senior Product Manager at Ling Universe: For hardware startups, team layout should be tailored to location. Brand, marketing, and financing teams are suitable for Beijing, Shanghai, and Guangzhou, but the supply chain team is recommended to be based in Shenzhen as a priority. From hardware selection, mold development to module procurement, Shenzhen's complete and mature industry chain offers outstanding advantages, significantly reducing trial-and-error costs during hardware implementation.
Liu Hongyan, Brand Strategy Head at Zhongbo Jili: Future AI companies that stand out must first select promising tracks and be able to execute a perfect closed loop from technology, product to commercial implementation. Although currently in a cash-burning phase, the key is whether they can burn their way to a sustainable and profitable model. Companies need their products to gain market recognition, forming a growth flywheel of revenue and profit, achieving long-term value enhancement. This is not only a key focus for investors but also fundamental to whether a company can develop steadily and sustainably.
This article is from the WeChat public account: ITjuzi , Author: ITjuzi
