Investing in the Brain: a Generalist Investor Guide to Spotting Winning Neurotech Startups
Neurotech stands at the precipice of growth, a field burgeoning yet constrained by the scarcity of startups attracting significant venture capital, going public, or being acquired for substantial amounts. In my observation, only a dozen neurotech companies have crossed the $100M mark in total funding over their entire lifetimes across multiple investment rounds — a clear indicator of investor interest. This pales in comparison to the broader U.S. tech scene. In 2021 alone, 42 tech companies each garnered over $100M in Series A rounds, and are at the beginning of their journeys to raise significantly more (see here for an explanation of the different stages of startup funding).
Evaluation frameworks
The blueprint for a stellar neurotech company remains a puzzle. Investors have not yet established quantifiable methods of evaluating neurotech companies comparable to those applied in other industries. For example, companies operating in the software as a service (SaaS) market can rely on what is known as ‘SaaS napkins’, a collection of metrics that a company should hit to be considered by venture capitalists (VCs). Building companies and raising capital are particularly challenging in the absence of a consensus on criteria and metrics for what constitutes success — or at least the promise of it.
Founders and investors need a reliable framework for evaluating neurotech startups. Without established metrics to lean on, the focus should shift to investing in startups uniquely poised to solve industry-specific challenges. These challenges vary in their intensity and relevance when compared to other sectors, such as consumer internet or B2B SaaS.
“Striking the right balance between friction and value is a defining factor of the neurotech startups that I consider seriously for investment.”
Friction vs value
A notable hurdle in neurotech is the invasiveness of its most advanced hardware. For example, at present the most advanced neurotech applications for speech reconstruction are invasive. End users’ reluctance to embrace invasive technologies, even when medically advised, underscores a vital concern — risk and friction may outweigh an offering’s value proposition. For instance, the value of deep brain stimulation (DBS) for Parkinson’s disease was questioned by 45% of studied individuals who subsequently opted for the procedure.
With this concern in mind, neurotech startups have begun to employ dual strategies. Some are innovating to reduce invasiveness and friction through non-invasive modalities like EEG and fNIRS and targeting specific use cases of relatively limited value, such as EEG for neonatal seizure monitoring or fNIRS for tracking motor recovery in stroke patients. Others are betting on delivering substantial value from invasive technologies, akin to Neuralink’s ambitious brain chip app store concept. In my own investment practices, striking the right balance between friction and value is a defining factor of the neurotech startups that I consider seriously for investment — both inside and outside the healthcare space.
Scalability
However, minimising initial user friction alone is not sufficient for venture capital viability. Scalability, too, is paramount — the ability to expand from a few patients or users to potentially millions in the healthcare domain (in the UK alone, millions of people suffer from mood, sleep, and other brain disorders, costing healthcare systems hundreds of billions), and to potentially hundreds of millions in the consumer space. For example, a startup enhancing the attractiveness of DBS needs to address the shortage of qualified neurosurgeons required to perform the surgery, considering that there are only about 3,500 neurosurgeons presently in the US.
The challenges of scaling non-invasive technologies differ from the challenges associated with scaling invasive ones, primarily due to the fact that non-invasive technologies do not involve surgery. For example, assuming that scalp EEG can achieve speech reconstruction quality comparable to that of invasive modalities in the near-term, that technology will face significantly greater demand from a much wider audience. Its scaling challenges will thus revolve around ensuring reliable EEG functionality in a greater diversity of environments and use cases, the mass production of electrodes, and minimisation of the need to replace them.
Accurately predicting the trajectory of scaling in neurotech is quite challenging. It could evolve through automated surgeries, simplified manual procedures using stents, or a shift towards completely non-invasive methods. Regardless, a key consideration for a neurotech startup is its strategy for serving a large user base effectively.
Scaling up also introduces further engineering challenges to be solved, such as the requirement for frequent upgrades and updates. As the number of devices in use by patients or consumers grows, it becomes logistically more challenging to ensure that device hardware and software are updated in a timely and smooth fashion. In healthcare use cases, the rate of upgrades and updates is primarily driven by the patients’ needs — fewer or indeed no updates may be required if the existing technology functions effectively and meets patients’ needs.
In contrast, it’s harder to put upgrades and updates on pause in the consumer markets, where startups and corporations need to frequently innovate with new features in order to retain users and monetise their products.
When it comes to managing the product life-cycle, despite the neurotech industry being in its infancy, founders must go beyond exclusively focusing on fundamental needs, such as implant extraction and energy supply. They must additionally place early emphasis on designing robust mechanisms for device hardware and software updates in order to periodically deploy new releases with improved performance or expanded feature sets. For example, a new generation of Neuropixels probes was released in 2021, four years after the original hardware’s release. As generations of these and other systems evolve more quickly, the need to replace already implanted devices may occur.
“A key consideration for a neurotech startup is its strategy for serving a large user base effectively.”
Venture capital investors place a strong emphasis on scale, recognising that conquering large markets is key to achieving the substantial investment returns they seek. “If a seed fund is $100M, an investor generally wants to see a path for any portfolio company to become a $1B enterprise value company” in order to return the fund for an investor. That, in return, requires the market to be worth at least a few, if not tens, of billions. In my own practices, I prioritise startups that target existing multi-billion dollar markets, or bet on a vision of creating new markets.
Platformisation
Platformisation has been a successful scaling strategy for many tech companies, enabling them to expand across various sectors to gain market share, as well as to build barriers or moats that protect them from competitors. In fact, according to the Financial Times, “the top 43 publicly-listed platform companies had nearly twice the operating profits, growth rates and market capitalisations of the 100 largest firms in the same businesses over a 20-year period — with half the workers.”
At a high level, platformisation creates value through connecting with each additional element produced by either the platform owner themselves or third-party firms. These elements may be products, components, software applications — anything that creates value for a customer. For example, the value of MS Windows goes up with a new Word, Excel, or Skype offering. Alternatively, value can be created through connecting two or more market sides for the purpose of linking buyers and sellers or exchanging information. Think of Airbnb or Uber, or Google in its early stages of matching websites with visitors.
“Value can be created through connecting two or more market sides.”
The application of platformisation to the neurotech industry, however, is complicated. For tech companies of the internet boom, code tweaking allowed them to expand to various verticals and industries, such as from developing customer support software to software for human resource management. The ability to reduce transaction costs through software allowed tech platforms to conquer markets where these costs were otherwise historically high: eBay in commerce, Airbnb in hospitality, Uber in transportation, and so on. It is, as of yet, unclear whether similar strategies will work for neurotech. It might be difficult, for example, to expand between different medical indications, or to repurpose electrodes initially positioned in a given brain location for one intended use, such as controlling a mouse cursor, for use in a different application, such as speech reconstruction. Moreover, no direct applicability of neurotech to lowering transaction costs has so far been reported.
Some neurotech startups are exploring various platform models — either through vertical integration, by targeting multiple indications, or incorporating marketplace elements in their offerings. For instance, Neuralink’s product includes a vertically integrated platform with value generated through several core components, including the implanted brain-computer interface (BCI) itself and a surgical robot used to perform the BCI implantation procedure. Similarly, in 2023, computational neurodiagnostics and electroencephalogram (EEG) analytics company Beacon Biosignals acquired Dreem, a developer of an EEG headset, allowing them to integrate various software and hardware components from each product suites to create a more end-to-end offering.
Platformisation doesn’t only apply to the vertical direction; horizontal expansion — to new indications or capabilities — is also possible. For instance, Bionaut Labs develops microscale robots to tackle a range of indications, including brainstem glioma and neurodegeneration, using the same core platform. Meanwhile, MintNeuro is developing modular circuits with for a range of brain sensing and stimulation applications.
Using a marketplace approach typically associated with e-commerce companies can also be a successful strategy for neurotech ventures. NeuroStar, developer of a transcranial magnetic stimulation (TMS) system, allows patients to find and book appointments with clinicians experienced in the technology directly on their website. Muse/Interaxon, a developer of an EEG headset for meditation, created a subscription offering through which customers can access meditation courses provided by field experts.
It is up to company founders to choose a suitable approach to platformisation. The art of selecting and executing a platformisation strategy is something I closely observe and assess in founders I consider for investment.
Technology
Regardless of their platform strategy, neurotech startups must integrate with and rely on existing technology stacks. Reliance on cloud-centric architecture — typically a staple for tech giants — may not be as effective for BCIs, where continuous and reliable operation is paramount. Parallel challenges have been evidenced in the autonomous vehicles space: if a self-driving car goes offline, a human driver takes over or the car might come to a halt. Similarly, if a speech decoding BCI goes offline, a locked-in user should still be able to communicate.
“Recognising technological dependencies, navigating around them, or addressing them in-house is a key trait I look for in startup founders.”
Understanding the limitations of current technology is only one aspect. Another important one is the identification of critical enabling technologies and strategies. For example, in order for brain-imaging technology magnetoencephalography (MEG) to become portable, innovations in so-called optically pumped magnetometers were required. In order to implant sophisticated BCIs at scale, robotics innovation will likely be required. Recognising technological dependencies, navigating around them, or addressing them in-house — like Neuralink with their surgical robot — is a key trait I look for in startup founders.
Equally important is understanding the broader product landscape for effective market positioning. For example, a few neurotech startups focusing on EEG overlook the highly competitive environment and existing solutions. Knowing how a neurotechnology performs compared to established, sometimes ‘lower-tech’ competition is very important — as is being able to envision what a trajectory to 10x improvement over competitors involves. For example, if a user can get insights about their breathing and meditation technique from a smartwatch, what will compel them to switch to your headset?
The ability to develop and market breakthrough technologies and to identify and then expand beyond initial markets into dominant industry platforms requires diverse expertise. Depending on the startup’s starting point, skills in materials science, machine learning, neuroscience, clinical affairs, data science, electrical engineering, and product management may be highly valuable. For example, one startup may emphasise materials science to develop on novel electrodes; another may double down on machine learning to devise cutting-edge stimulation protocols. A well-balanced founding team with a mix of various backgrounds is something I value highly as an investor.
While expertise is one important angle for evaluating a founding team, another is founder-market fit. Founder-market fit, in addition to knowledge and experience, emphasises “obsessive commitment to the goal of your startup as the problem [that] resonates so deeply.” However, as raising VC money frequently involves hype, sometimes staying true to your company’s ‘north star’ can be difficult. My advice here is to stay authentic. Like many investors, I am guilty of projecting my generalist biases on neurotech founders and nudging them towards consumer-focused business models. The ability to listen to investors yet push back when what you hear contradicts your obsessions and goals is essential.
Conclusion
In summary, I seek authentic teams with complementary skills who can develop products with a clear trajectory for significantly outperforming existing solutions, can scale to tens of millions of users, and which have the potential to become the platforms of the future — much like Apple’s iPhone, Microsoft’s Windows, or Ford’s conveyor did in their respective eras.
Written by Peter Zhegin, edited by Lewis Bates, Dilara Parry and Lars Olsen, with AI-generated artwork prompt-engineered (apologies for the spelling…) by Sophie Valentine.
Peter Zhegin is an angel investor at approx.vc, an angel collective, where he focuses on investing in the data stack and frontier technologies, including neurotech. Formerly, Peter worked at 7percent Ventures, an early-stage trans-Atlantic frontier tech venture capital fund. You can follow Peter on X (@peterzhegin).
Lewis Bates is a software quality engineer in the medical imaging field, currently studying neurotechnology through Queen’s University. Also a certified mindfulness and emotional intelligence trainer, he is interested in mindfulness and mental health-related neurotechnology and neuroscience research.
Dilara Parry is a data scientist at a global consultancy. She has a neuroscience (MSc) academic background with BCI research experience and the curiosity to explore the overlap of her two disciplines.
Lars Olsen is a regulatory medical writer. He works in the pharmaceutical industry writing submission documents, and has additional experience with medical devices. He has a biology background and is interested in AI, AGI/ASI, and BCI/HCI.
Sophie Valentine has a background in experimental psychology and cognitive neuropsychology research, with degrees from Bristol University. Her work is focussed at the intersection of tech-for-good, product, digital health, and neurotechnology.
