Zhu Mingming on Smart Glasses and Long-Term Choices in Hardware

Zhang Xiaojun interviews Rokid founder Zhu Mingming about his two startup ventures, smart-glasses products, hardware competition and organizational choices.

Zhu recalls that his first company was later acquired by Alibaba. He sees the opportunities and help he received as important elements of luck in his startup journey.

He describes meeting Jack Ma, Joe Tsai and Wang Jian and seeking investment when the company was short of cash. These details are his personal recollections.

He recalls that at the most difficult point payday was two weeks away and funds were limited. Subsequent investment gave the team more time to continue operating.

He says his first company continued after receiving support and later joined Alibaba, giving him an opportunity to understand platforms and organizational operations.

During his time at Alibaba, he recalls that Wu Yongming proposed exploring AI; M Lab was then formed to focus on inputs such as cameras and microphones and related capabilities.

He explains leaving a large company to start again: the direction he wanted to pursue was highly uncertain, required talent and patience, and might not fit an established company’s pace.

Discussing Rokid’s early fundraising and advice he received, he says capital should serve a long-term product direction and founders have responsibilities to investors and the team.

He recalls a major decision around 2019 to move from the existing AI hardware direction toward AR, requiring the team to reassess markets, technical maturity and resource allocation quickly.

He had followed virtual and augmented reality but preferred to pursue AR and made a judgment at the time about when the technology might mature.

He says the team weighed continued spending against potential value and chose to direct limited funds toward what he considered a stronger long-term opportunity.

He distinguishes two kinds of glasses products: one emphasizes rich content and interaction; the other must first work as comfortable glasses while offering audio, camera and assistant functions.

He expects the paths might converge as content-focused devices become lighter and everyday glasses gain capabilities. This is a product forecast, not an accomplished outcome.

On phones and glasses, he discusses different wearable-device use cases. Sustained adoption of glasses also depends on weight, appearance, functions, privacy and user experience.

He imagines voice and AI helping users complete some tasks more directly, reducing navigation through graphical menus; actual usefulness still depends on recognition and execution reliability.

He is optimistic that people who already wear glasses will accept AI features and that this may change phone habits. His timetable is a 2025 prediction, not a guarantee.

The transition involved layoffs. He recalls wanting to compensate departing employees while preserving enough cash to test the new direction.

He discusses searching for product-market fit after the pivot: a team must not only build a device but observe whether people will buy and keep using it.

He compares smart glasses to the period between BlackBerry and the first iPhone, arguing that the category is still seeking a more mature form. This is his industry judgment.

He divides AI hardware applications into embodied systems interacting with the environment and devices carried by users; both must bring model capabilities into practical settings.

On competing with giants, he says startups should look for opportunities large firms may not want, cannot quickly pursue or struggle to accommodate organizationally, rather than imitate their scale.

He discusses how smart-glasses definitions differ across markets, emphasizing the influence of ecosystems, use cases and supply chains.

He argues a developer ecosystem requires joint investment in technology, product experience and partnerships, and that a small company must consider how to attract outside participants.

He sees hardware competition as different from software: as supply chains mature, later entrants can use existing capabilities, so early movers must keep improving.

Discussing culture, he values curiosity and playfulness and says he will correct decisions based on product experience or data, while someone must bear the consequences.

He wants a product that does more than showcase technology: users should understand and use it without an engineer beside them. That requires further work on hardware and interaction.

On fundraising, he says capital can support capacity and research but also brings responsibility; the appropriate amount depends on the company’s plans and needs.