The key to the successful business model of AI native wearable devices is whether they can continue to provide high value. There is an AI wearable device company called Whoop, which was founded in 2012 and initially focused on sleep and exercise recovery for professional athletes. At this stage, it grew into a $3.6 billion unicorn in sports wearable devices.
When generative AI technology explodes and gradually matures, it adds AI functions to its own devices and further evolves in the perception accuracy of the devices, expanding its imagination space.
Whoop completed a $200 million Series F funding round in August 2021, led by SoftBank Vision Fund Phase 2 (with a valuation of $3.6 billion after this round). Other than those involved in investment IVP、CAVU Consumer Partners、Thursday Ventures、GP Bullhound、Accomplice、NextView Ventures In addition to investment institutions such as Animal Capital, there are also football player Cristiano Ronaldo, golfer Tiger Woods, and basketball player Kevin Durant.
The Harvard team, led by professional athletes, has created AI wearable devices for sports and health
Whoop was founded in 2012 by Will Ahmed (founder and CEO), John Capodilupo (co-founder and CTO), and Aurelian Nicolae (co-founder and director of hardware engineering), who met at Harvard.
The idea for entrepreneurship came from Will Ahmed, who was the captain of Harvard’s squash team at the time. As an athlete, he felt a lack of understanding of his physical data such as training, recovery, and sleep. In this context, he came up with the idea of “creating a wearable device that can monitor recovery, sleep, and training load”.
So the initial positioning of Whoop is naturally to serve athletes and professional training groups, with a focus on monitoring recovery, sleep, and training load in these three dimensions.
Will Ahmed， Source: Whoop
As Will Ahmed himself said, “The most exciting thing about wearable technology and 24/7 monitoring is the information we haven’t discovered yet
The founding team consisted of three people, one born as an athlete and the other two knowledgeable in technology and hardware. Although they were still young at the time, they were still a good combination for founding a wearable device company.
Almost ten years after Whoop was founded, ChatGPT was released at the end of 2022, ushering in the wave of generative AI; GPT-4 will be released in March 2023; In September 2023, the Whoop 4.0 wristband was released, along with the GPT-4 based Whoop Coach. This health AI assistant can comprehensively analyze users’ personal biometric data, goals, and the latest sports science knowledge, providing highly personalized and conversational health and fitness guidance for users.
Afterwards, Whoop officially embarked on the development path from wearable devices to AI wearable devices. It launched Whoop 5.0 and Whoop MG in 2025.
What can AI bring to healthy wearable devices?
The hardware advancement of Whoop 5.0 compared to Whoop 4.0 is primarily due to sensor upgrades. Its tactile engine, accelerometer, and gyroscope have all been upgraded, making it more sensitive in perception. Secondly, its processor has been redesigned to increase energy efficiency by 10 times and extend battery life to 14 days. If paired with a wireless power bank, it can also provide an additional 14 days of battery life. After all, always on is an important advantage and feature of wearable devices, and battery life is crucial.
Whoop MG is a newly launched hardware series, although it can also count steps and record basic data, it is not an ordinary fitness tracker, but more inclined towards medical assistive devices. Compared to Whoop 5.0, Whoop MG includes a medical grade electrocardiogram (ECG) sensor.
Source: Whoop
Whoop’s health functions focus on three core pillars: sleep, recovery, and consumption. These input data collectively generate daily recovery scores, helping users understand their physical and mental readiness for the day.
The uniqueness of Whoop lies in how it combines sensor data with users’ self-reported habits. Through a feature called ‘Insight’, the platform extracts information from a customizable daily log that offers over 160 tracking options – covering everything from supplements, medication to mental health.
For women, Whoop 5.0 also features a physiological cycle tracking function that provides insights into hormonal changes and how they affect sleep, stress, and physical recovery.
Sleep quality and physical exertion are key driving factors for recovery scores, mainly calculated through heart rate variability (HRV), which is the change in time interval between heartbeats. Simply put, the higher the HRV compared to the user’s baseline, the better the user’s physical recovery is likely to be.
Of course, consumption has a significant impact on this value. In the context of Whoop, ‘consumption’ refers to the combination of activity and physical exertion. The device will automatically detect most exercises, but users can also manually start training within the app – or make up for it afterwards.
Taking strength training as an example: Whoop allows users to build their training by inputting weight, number of times, and number of groups. Using heart rate data during training, Whoop estimates the effort put in and provides a daily consumption score. The higher the heart rate during the activity, the higher the calculated expenditure.
In addition, the recovery score also takes into account resting heart rate, respiratory rate, and sleep performance.
On top of these basic features, Whoop 5.0 also adds a new dataset called “Healthspan”. It focuses more on long-term health, calculating users’ “Whoop Age” and “aging rate” by analyzing nine different indicators.
Compared to regular products, Whoop MG introduces on-demand electrocardiogram readings, estimated blood pressure tracking, and biological age estimation.
Whoop MG can perform heart screening, and its electrocardiogram function has been approved by the FDA. Users can touch the watch button electrode, wait for 30 seconds, and the application will give them a medical grade reading. It can also passively monitor arrhythmia and issue alerts, but these two functions are more of a safety net than a clinical solution.
For blood pressure, this is still a testing phase function, which is estimated based on algorithms after one-time calibration using a standard cuff blood pressure monitor. This feature is not entirely accurate and cannot replace true medical grade monitoring, but it still has a preventive effect.
Whoop now offers the ‘Whoop Age’ feature. This feature uses nine key indicators and takes six months to estimate your Whoop age. It is not intended to provide medical advice, but rather an intelligent, data-driven snapshot of the overall health status of users.
It can be seen that Whoop is expanding its territory from the field of sports and health to the semi clinical field.
In terms of AI functionality, Whoop has a “Your Daily Outlook” feature based on OpenAI’s advanced model support. Every morning, users will receive a personalized summary, including weather, key points from the previous day, and factors that may affect their recovery. Users will also receive activity suggestions and can directly raise follow-up questions in the application.
It is very helpful in discovering trends. For example, some users noticed that their breathing rate gradually increased based on this feature and made corresponding adjustments according to suggestions.
Through the Whoop Coach feature, users can delve deeper into data, understand their physical condition and possible reasons behind it, and further obtain feasible suggestions, such as breathing exercises, hydration reminders, or light activities like yoga and walking.
Based on long-term memory, it can also manage users’ unhealthy habits, such as when was the last time they drank alcohol? How many times did you drink it last month? Whoop Coach can tell you.
In terms of business model, Whoop is a membership based service. Users do not need to pay hardware fees in advance – the cost is already included in the subscription fee, and each membership level unlocks different features. Whoop MG belongs to the highest level, with an annual fee of $359. For those who particularly wish to monitor cardiovascular health, it is a useful assistive device – especially when used in conjunction with clinically validated tools that users already have.
The key to the successful business model of AI native wearable devices lies in whether they can continue to provide high value
Smart devices related to health and fitness have been facing bottlenecks and limitations in the past if they adopt a hardware cost+subscription business model.
For example, in the field of fitness mirrors, there were previously products such as Tonal and Mirror, with Tonal’s highest valuation reaching $1.6 billion, while Mirror was acquired by Lululemon for $500 million.
Their main function is fitness assistance, which can be considered as a “fitness private tutorial” at home that allows users to see their fitness movements and correct them at any time. However, from the current situation, both Tonal and Mirror have experienced significant valuation declines and asset impairments.
Why haven’t their business models been successful? Because there is a question that is difficult to answer: does the user feel that long-term subscriptions are worth the sustained value you bring to them?
This question also needs to be answered by intelligent wearable devices in the AI era. Because most hardware devices were previously paid for at once, users would not continue to spend money on them after purchasing the device. Spending money on hardware while continuously paying subscription fees is currently a challenge to user perception.
However, until the end-to-end AI models and computing power are powerful enough, the AI capabilities of AI wearable devices still need to be provided in the cloud. Without subscription fees, manufacturers cannot afford the long-term model inference costs, and it is even more difficult to iterate on the experience of intelligence.
So, how could this paradox be resolved? It still needs to be implemented in terms of experience, truly making users recognize the value of long-term subscriptions, which may come from the collection and processing of data.
Taking health wearable devices represented by Whoop as an example, their sensors must become increasingly accurate, the data collected must be more precise, and the insights and suggestions obtained through AI processing will be more valuable. In addition to precision, it is long-term. If the device can collect users’ health data accurately for a long time and enable AI to insight into users’ trends in exercise or health, it will be very helpful for users in chronic disease management and exercise rehabilitation. Finally, there is personalization. Everyone’s health condition is different. If AI can provide truly useful exercise or rehabilitation suggestions based on each person’s personality data, it can also provide users with a sense of value.
For AI native wearable devices, they have not yet converged to a specific form or exploded in a specialized field, but what is certain is that they will become a portable entry point in the future. When they collect enough richness and personalization, they can find their own focus areas beyond generalization.
Alpha Corporation is very optimistic about the opportunities in the AI native hardware field and has completed early investments in nearly ten AI native hardware startups, including Guangfan Technology, Looki, X-Origin AI, Pixboom, and Noyiteng Robotics. The vast majority of them are first round investors in the projects, and several of them have already completed new rounds of financing. We look forward to communicating and collaborating with more start-up teams on this track.
暂无评论内容