Insights from Leqi GlassesAI + Hardware First Bucket of Gold
Many opportunities were decided five years ago. By 2025, the AI trend has already lifted several pigs into the sky! The choices we make today directly determine the market opportunities of the next five years and the business direction of the next ten years! Let’s first analyze the miracle of Leqi AI glasses selling over 100 million in just five months: how close are we to the first bucket of gold in AI + hardware?
In fact, Leqi’s success is not accidental; it hides three major golden tracks for making money with AI hardware: empowering traditional hardware with technology, scenario-based solutions, and ecosystem service sharing.
Many people feel that Leqi has a strong background as an Alibaba-invested project. Ordinary people like us do not have huge funds, top-notch technology, or access to Alipay’s 1 billion customer resources, making it difficult to replicate Leqi’s miracle.
First, let’s see what Leqi did right.
First, clear positioning.
They chose the price range of 3000 yuan. Anyone who understands finance knows that if you choose a product priced at 100 yuan, you need to find 100,000 customers to achieve 10 million in sales. However, at a price of 1000 yuan, you only need 10,000 customers. Leqi’s price range is over 3000 yuan, which is considered a middle-class harvesting machine in the market. With 30,000 customers, they have achieved a small goal. In reality, selling houses, 1,000 customers can earn a net profit of 100 million.
Secondly, focusing on essential needs.
Leqi selected only high-frequency scenarios among the complex functions of AI glasses, such as asking for directions, ordering food, and communication, focusing on scenarios used more than five times a day. They eliminated the flashy but useless applications.
Payment: Integrating with Alipay, “tap to pay” allows glasses to pay by tapping the phone (solving the problem of scanning codes abroad).
Navigation: AR real-time arrows + bone conduction audio (no need to look at the phone while riding/driving).
Translation: Double-tap the temple to translate automatically, supporting dialects (Sichuan dialect → English with an accuracy rate of 92%).
There are no universal products, only precise pain points — this is the core of Leqi’s 35% repurchase rate!
Third, dimensionality reduction in technology: using a “beggar version” solution to create a flagship experience.
Leqi did not use Qualcomm’s flagship chip but chose Rockchip RK3566 (costing only $8), paired with the open-source AI framework Tengine, compressing local voice recognition latency to 280ms (faster than AirPods). The most impressive part is the translation module: instead of developing their own algorithm, they integrated Youdao’s offline translation SDK (annual fee of 200,000 yuan, costing only 2 yuan per unit for 100,000 units), achieving real-time translation in 37 languages — achieving 80% experience at 20% cost, which is the essence of grassroots entrepreneurship!
Finally, leveraging the minimum order quantity in the supply chain, they used pre-sale funds to pay factory deposits.
In fact, Leqi is quite frugal with costs: they spent 8000 yuan on a website to buy open-source circuit diagrams, only modifying the interface to fit their lenses. Production: They found Huawei’s supplier, Wingtech Technology, for OEM, with a minimum order quantity of 500 units, reducing the cost per unit to 320 yuan.
Then they used JD crowdfunding with the offer “1 yuan off 100 yuan,” locking in 20,000 orders in three days, using user funds to pay factory deposits. They basically achieved zero inventory, maximizing the survival strategy of a small team!
To summarize, the core logic behind Leqi’s 100 million sales in five months is something we can directly replicate: first is dimensionality reduction in technology, using a 30 yuan AI translation module to turn ordinary headphones into “talking” smart devices, increasing the price by 3-5 times; then focusing on essential scenarios, zeroing in on payment, navigation, and translation, which can be used more than five times a day, avoiding those “jack-of-all-trades, master of none” pseudo-needs; finally, adopting a light asset model, outsourcing hardware production, starting with pre-sales, and stabilizing cash flow through crowdfunding!
Next, here are three opportunity tracks that ordinary people can directly enter, which are easy to understand:
The first track: traditional hardware AI transformation, can be mass-produced in three months, and grassroots can do it!
First, choose the right category, focusing on those priced around 500 yuan, above 300 yuan, because the 300 yuan price range is highly competitive.
Next, select hardware with a large user base, such as Bluetooth speakers, desk lamps, and dash cams. Then integrate AI functions; for voice interaction, use iFLYTEK’s offline voice module, costing only 25 yuan, which can also customize wake words; for visual recognition, directly connect to Baidu’s AI open platform, which offers 200 free calls per day, easily achieving object recognition. A merchant in Shenzhen’s Huaqiangbei transformed a regular rearview mirror into an AI driving assistant, adding front car warnings and voice control, with only an additional cost of 80 yuan, but the selling price increased by 300 yuan, which is a great profit! The supply chain is also simple; they spent 5000 yuan on a website to find engineers for circuit diagrams, and Alibaba for hardware solution providers, with a minimum order of 500 units, controlling the cost of smart speaker modifications to under 120 yuan per unit. Sales rely on Pinduoduo for 9.9 yuan pre-sale market testing, and Douyin for AI function comparison videos, investing 1 yuan to earn 3 yuan, ensuring profit!
The second track: vertical scenario solutions, light asset service, earning 500,000 a year is not a dream!
Let me give you two existing examples: in the elderly care scenario, if you have resources in this area, you can create a smart bracelet that integrates Huawei Cloud’s fall detection algorithm, costing only 0.01 yuan per call, with hardware sales earning a 40% gross profit, plus a monthly data service fee of 19.9 yuan per person, cooperating with community elderly care stations, installing one can earn a 200 yuan commission;
If you lack resources and are purely market-oriented, consider the education scenario, using Raspberry Pi and microphone arrays to create an AI speaking practice machine, integrating Youdao translation API, offering “free hardware + course subscription,” charging 99 yuan per month, with a 55% revenue share from school channels, making money effortlessly! The key to this is selecting the right algorithm, prioritizing zero-code platforms like Baidu EasyDL and Tencent TI-ONE, which allow you to train models without programming knowledge, and ensuring data compliance by signing data usage agreements with users, avoiding the collection of sensitive information.
The third track: AI hardware ecosystem profit sharing, zero-risk “lying down to earn” model!
First, become an agent, becoming a regional agent for brands like Shumai Technology and Q&A, earning a 15-30% markup; you can also provide AI voice ordering + cash register transformation for restaurant merchants, charging 5000 yuan per store, and later taking a 3% share of the revenue.
Second, develop paid skills for Tmall Genie and Xiao Ai, creating children’s stories and corporate training courses, earning 0.1 yuan for each call. A developer named “Xiao Ming AI” created a workplace English conversation skill, with 100,000 monthly active users, earning 80,000 yuan in revenue share each month!
Third, do data labeling and testing, taking tasks on JD Crowd Wisdom and Baidu Crowdsourcing, such as labeling gesture data for smart glasses, earning 30-50 yuan per hour; testing night mode for AI camera manufacturers, earning 200 yuan per unit, simple and stable!
Finally, let me highlight some key points to avoid pitfalls and provide some free resources:
Cost control is crucial; hardware BOM costs should not exceed 40% of the selling price. For example, for a product sold at 500 yuan, the hardware cost must be controlled within 200 yuan; for software development, prioritize using cross-platform tools like APICloud and uni-app to save costs. Compliance with policies is also essential; smart hardware must pass the CCC certification, which costs about 5000 yuan through a third-party agency; data cross-border must comply with the Data Security Law, and servers should be selected from domestic nodes.
I’ve also organized some free resources for you: open-source hardware like Raspberry Pi (300 yuan) and ESP32 development boards (50 yuan); AI models like TensorFlow Lite Micro and Baidu Paddle Lite; for the supply chain, visit Shenzhen Huaqiangbei SEG Electronics Market or Lichuang Mall to buy components, directly connecting with solution providers.
Finally, here’s a startup suggestion from 0 to 1, using AI translation headphones as an example:
In the first month, use ESP32 and iFLYTEK modules to create a prototype, costing only 150 yuan, and shoot a comparison test video;
In the second month, crowdfund on Taobao with a 99 yuan early bird price, targeting 500 units, preparing product renderings and function demonstrations;
In the third month, based on order volume, find a Dongguan factory for mass production of the first batch of 5000 units, collaborating with Douyin influencers for sales, offering a 30% commission; by the sixth month, iterating to add offline translation features, entering JD’s AI smart pavilion, and if monthly sales stabilize at 2000 units, monthly profits could reach 150,000 yuan!
The golden period for AI hardware is within the next 2-3 years. Our advantage as ordinary people is the ability to quickly test and focus on scenarios. Don’t get bogged down in technical details; use open-source solutions to create the minimum viable product and validate the market with pre-sale data. This is the starting point for Leqi’s transformation from a startup to a unicorn! If you want to delve deeper into designing product solutions for specific scenarios like education and elderly care, we can discuss further, breaking down requirements and cost structures to seize this wave of AI hardware profit opportunities!
#LeqiGlasses #AIOpportunities #FirstBucketOfGold#