Soma Capital
Sleep & Why BCIs are Cooler Than You Think
Investment Thesis
Ismail Mardin
Ismail Mardin
CS+BioE @ Stanford
Published on 1/25/2025
image

Sleep - a wake up call

We spend roughly a third of our lives sleeping - around 26 years. Sleep is pivotal to our quality of life, not only on a day-to-day basis (where the quality of the previous night’s sleep can determine an entire day’s mood) but also on a longer term, being one of the leading determinants of longevity. It is at least as important as exercise and nutrition, and so, imagining the utility of a hypothetical sleep company compared with other health companies, it is not hard to imagine valuations similar to that of FitBit, Garmin, or Peloton. And yet, this hasn’t really happened (an argument can be made for WHOOP, Oura etc. but these are more so overall diagnostic tools, without actively improving one’s sleep); there are no well-established companies that - as a primary function - actively tackle the sleep problem. Rather, the current landscape is of small-mid sized companies, often still in initial growth stages, and so there is significant opportunity in the marked difference between the importance of the product offered (better sleep) vs the quality of what companies are actually offering currently.

Before going any further, it’s important to consider why now is the right time to pursue this. If, in the last 10-20 years, there has been no company that has managed to fully capture this opportunity, what is currently shifting to now facilitate this? I’d put forward a few reasons:

  1. The COVID-19 pandemic. It disrupted normal routines, triggering widespread sleep issues and creating a more fertile ground for sleep solutions. With work-from-home policies, there was no longer a clear separation between work and rest. Screen times increased dramatically. Workloads increased too, and all these factors made healthy sleeping much, much harder. There was a surge in demand for sleep aids, with one in five Americans now relying on over-the-counter sleep products - clearly, the need for healthy sleep, and a willingness to pay for products that improve it, has increased.
  2. Simultaneously, there has been growing recognition of the link between sleep and mental health. Popular mental wellness platforms like Headspace and Calm have expanded into this intersection, again creating a precedent for paying a company to improve one’s sleep.
  3. The financial benefits of improving sleep quality are also becoming increasingly clear. For employers, they lead to more productive workers. In particular, in industries where mistakes are disastrous and/or fatal - doctors, workers dealing with dangerous machinery, bus/truck drivers - it is pivotal for employees to be performing well. For insurers too, they lead to lower overall health costs.
  4. Lastly, and perhaps most importantly, there is an overall trend of becoming more long-sighted with our health. We only need to open Instagram or TikTok to see influencers with millions of subscribers promoting health, longevity etc. - from college students to Bryan Johnson. Importance on our healthspan is quickly becoming level with our longevity, and new companies are formed simply on the differentiation of being more health/sustainability focused. Sleep naturally fits in with this trend.

So - now is the time to build/bet on companies in the sleep space. What features would lead to success?

1. Precision: Sleep is an incredibly personal aspect of people’s lives: sleep schedules, whether you sleep alone or with others, the mattress you use, the lighting you’re used to, and so on. Unlocking the full potential of sleep solutions, therefore, would lie in the ability to provide a deeply personalized approach.

Of course, it would be near impossible for a company (especially a startup with limited resources) to cater to every single need, and so a work-around would be to initially target people with similar sleep and lifestyle preferences. Examples might include doctors (their preferences may include short bouts of sleep around shifts), athletes (who might need to time sleep around flights, jetlag etc.), soldiers (who need the best quality sleep possible in the shortest amount of time in combat situations), and so on.

By focusing on specific market segments, startups can gain relevant insights into unique user segments. As a result, their sleep interventions would be far more effective than one-size-fits-all alternatives; if a startup can identify a large enough customer base with similar specific contexts, they can more effectively specialize their product.

(Crucially, this is what a lot of big players nowadays have not considered - Oura, WHOOP, 8sleep etc. all cater to the general consumer. This may be ok if no one else is specializing, but the quality of the product is severely limited. Once a company really masters a niche - for example, being the #1 tool for deployed soldiers to get a good night’s rest while fighting in unfamiliar, foreign terrain - the difference in quality will be striking. And for something as important as sleep, having the best available matters.)

Once this niche has been mastered, built out, and a reputation is established, a startup can expand into other sectors of sleep.

2. User Experience: As with most products, the user experience is everything - particularly so for something as personal as one’s bedroom/sleep conditions. A dedicated part of the startup’s workforce (doesn’t have to be too extensive, but needs to be responsive) would ideally provide a kind of white-glove service: on the software side: personalized schedules and sleep tracking/plans, on the hardware side: personalized headsets, EMGs, mattresses etc.

Such attention to individual user needs fosters loyalty, builds community, and encourages user advocacy - all crucial elements for long-term success. This is particularly important for a health-focused company - more users leads to more data and insights, which leads to a better honed product, which leads to more users and so on.

Existing companies within sleep:

Actively improving sleep/sleep habits:

Passively improving (with sleep scores, diagnostic tools etc.)

---------

Why BCIs are even cooler than you thought

In recent years, we’ve witnessed a pretty significant shift in how we see healthcare - from focusing solely on physical well-being to recognizing the importance of mental and cognitive health. There is not only greater attention on issues like dementia, Alzheimer’s disease, depression etc., but also on our ability to focus and be productive, targeting issues like brain fog, better sleep (see above!) and attention spans.

With this transition comes a new source of health metrics - the brain. BCIs, either through direct brain sensing (EEGs, invasive electrodes) or indirect (eye-tracking, EMGs, MMGs) are quickly becoming more and more widespread. From this data, we can (rudimentarily) track correlations between brain activity in certain areas with their corresponding effects - which signals relate to speech, movement, and at a more abstract level, emotion, stress, etc.

What’s most interesting to me, however, is how far we can take this. If this field can reach the same rates of acceleration as in other tech+adjacent fields, it’s perhaps not too ambitious to imagine, within our lifetimes, bio-sensing capabilities to a single-neuron resolution and beyond, with all the implications that follow. Imagine being able to measure which specific neurons activate, as well as their corresponding membrane potentials, when you move a particular muscle? When you’re looking at a particular image? Or even when you’re thinking about a particular memory?

Understanding the brain - and how it processes information at a very mechanistic level - is not just about improving health and productivity. Understanding and adapting its algorithms - how it processes data so efficiently - can revolutionize computing. The implications of this knowledge will also stretch into evolution, psychology and philosophy - into free will, and the development of consciousness. Just think about it: we could potentially unravel the very fabric of our thoughts, memories, and decision-making processes. We might finally understand why we are the way we are, how we became conscious beings, and what truly drives our actions - all things which, in my humble opinion, are firmly within the reach of human innovation.

So - that’s the vision. What’s stopping us from getting there?

If we look at the wearable health sector, we have many accurate and versatile products which produce great data: the Apple Watch, Fitbit, Oura, Whoop and so on. These allow researchers - both from within and outside these companies - to find insights and produce reliable algorithms to optimize health, and maximize a user’s understanding of how daily activities (such as working out or sleep) affect the body. There is, however, no comparable standard piece of hardware that provides high-fidelity, useful and actionable data for neurological metrics, even despite a growing trend towards concern of mental deterioration, cognitive impairment, stress and burnout etc.

Consequently, any company building within the brain biosensing space must build their own hardware setup. This is slow, and EEG and EMG data is significantly noisy and not simple to interpret. Companies in the space also take a very siloed approach - where each one safeguards their progress in hardware as their defensive moat - and this stifles the overall rate of innovation.

Not even considering the further implications mentioned in the first section, and looking at this from a purely health-related perspective: the brain is our most important asset - it makes us who we are, and controls our very perception of the world around us. We should have the tools to understand, diagnose, and ultimately, control it.

I’ve had the great fortune to meet a few companies working on this frontier. Sonera Magnetics is one - they’ve got an interesting approach to biosensing using magnetomyography (MMGs) instead of electromyography (EMGs). I’d highly recommend checking them out! Others which I’ve found fascinating include NextSense, NextStem, and Kernel - each pushing the boundaries in their own unique ways.

AIConsumer