My bets for the coming decade

I started angel investing in 2019 following the acquisition of my previous startup. After having met with hundreds of companies, here is what I am investing in next.

Photo by Karsten Würth on Unsplash...

A couple years ago when I started angel investing, I chose to focus on 6 industries: privacy, blockchain, quantum computing, psychedelics, microbiome and medtech. After having met a bunch of companies in the space and rethinking my thesis post-covid, here is what I am now focusing on:

Privacy

Privacy will remain one of my main topics of interest, and in particular technologies that leverage modern cryptography (homomorphic encryption, threshold signatures, post-quantum cryptography, etc..). I am quite invested via my own startup Zama, but I am looking to invest actively in other privacy-enhancing technologies.

My main bet is on a new technology called homomorphic encryption (see here for a 6 min introduction). It’s a new encryption technique that allows you to process data without ever decrypting it or seeing it. Imagine being able to do things like machine learning or database queries without knowing either what data you store, receive as input or what response you send back. In effect, it would enable end-to-end encrypted digital services, where companies and consumers can securely and privately use other companies’ services. Many applications would benefit from this, such as facial recognition, automated medical diagnosis, product recommendations, voice assistants, smart contracts, fraud detection, etc.. Basically anything that handles sensitive data.

This is the holy grail of data security, as it means that you no longer have to trust the company that you are sending your data to, as even if they got hacked or subpoenaed by a government, nobody would be able to access your data without the secret key that only you hold!

Although people have been working on homomorphic encryption for over 40 years, the first fully homomorphic scheme was only discovered in 2009. However, existing schemes have been too slow to be useful, requiring tens of minutes to do simple computations, or being limited to some basic arithmetic operations. As a result, it had remained an academic subject without industrial applications, something my new startup Zama aims to change.

I am convinced though that we have now reached a breakthrough moment, and that it is a question of 5 years at most before this is widely adopted by companies. It feels to me a lot like Deep Learning in 2010, just before it took off.

Blockchain

It’s no secret that I am an Ethereum maximalist. The reason is that contrary to Bitcoin, Ethereum enables applications to be built on top of it, creating compounding value for the protocol each time a new app becomes successful. And since these apps are built by developers, it means that the value of Ethereum is directly proportional to its developer community. As an engineer myself, I have always been convinced about the power of developer communities, and Ethereum is one of the most impressive I have ever seen.

In terms of applications, my bets are still on DeFi, NFTs and DAOs. DeFi because it makes Wall Street programmable, DAOs because they enable large scale coordination between groups of people (I wrote about it here), and NFTs because they are digital certificates for everything.

Regarding NFTs, it’s important to understand that they are simply digital certificates of ownership, just like the PDF you get when you buy something online. That an art piece is certified by an NFT does not mean NFTs are only useful for art. Sorare, Axie and other projects are not about art but still use NFTs. So when someone says “NFTs are a fad”, what they really mean is “the art piece this NFT certifies is a fad”. And while they might be right about the jpeg itself having no value, one thing is clear: NFTs are here to stay, and eventually, everything people own –from art to cars, objects to clothes– will have an NFT certificate attached to it.

High-performance computing

My initial thesis around Quantum Computing was that it could open up a lot of new doors for high-performance mathematical computing. I have now revised this thesis, and will be investing in a broad range of co-processors and high-performance computing tools.

Back when the internet was about serving content, it made sense to use commodity hardware for everything. But now that we are entering the age of data science, we need high-performance hardware that can accelerate all kinds of math algorithms, just like GPUs and TPUs have accelerated deep learning. 

One area I am particularly interested in, is tools that can accelerate FPGA development. My bet is that FPGAs will become ubiquitous in cloud platforms and that they will serve as customizable accelerators for all sorts of data workflows.

Another area I’m looking into, but still don’t have enough insights into is RISC-V. The technology looks really promising, and the open source approach resonates strongly with my thesis around developer communities eating the world. I’d love to chat with more people in that space.

Medical Hardware

One of the most striking things about Covid has been the lack of scalable medical infrastructure. Many deaths could have been avoided if we had more doctors and more machines, but governments don’t have more money to spend on expensive hardware, and training doctors takes a while.

This creates an opportunity to provide AI-powered, cheaper, smaller versions of the expensive machines used currently, such as MRIs, X-rays, vital monitors, ventilators, and so on. The benefits could be enormous:

• making machines cheaper would mean hospitals could buy more for the same price, dramatically increasing accessibility.

• making machines smaller would mean they become transportable, and hospitals could borrow machines from each other easily. This would enable a dynamically reconfigurable “just-in-time” medical infrastructure.

• providing AI tools optimized for the machines would enable doctors to save time and thus treat more patients. This would reduce pressure on the medical staff, and give us time to train new doctors.

I know investing in hardware, and in medical hardware specifically, can be scary, as it hasn’t been a startup friendly space before. But times have changed and disruption is desperately needed now.

Bioinformatics

My initial thesis was on microbiomics, or the analysis and manipulation of the microbiome. I think this thesis was too narrow, and there is a more general play that can be made here around the tools to analyze biological data. This field, called Bioinformatics, encompasses everything from analyzing DNA to analyzing the microbiome to designing new mRNA vaccines, and in the post-covid world, it will become a major part of any healthcare company. 

Fortunately, my PhD was actually in bioinformatics, and while I have been out of the field for over a decade, I believe now is the time to get back into it. Any company building data science tools and solutions for understanding and designing biological therapies is a good fit. If it’s open source, even better!

Psychedelics

Recent scientific studies have shown that of all the known active compounds, none has been as effective as psychedelics (such as Psilocybin, the active ingredient in magic mushrooms) at treating drug-resistant depression.

For having been through multiple episodes myself, I know how difficult and lonely it can be. When you consider that close to 30% of the western population takes some kind of anti-depressant drug, and that over 75% of millennials know someone who suffers from depression, it becomes obvious that we need to rethink mental healthcare. I can’t even imagine what these numbers must be after the pandemic..

Unfortunately, being illegal in most countries, psychedelics is not available as a treatment, unnecessarily leaving millions of people in pain. But this is now changing, with many companies going public in Canada and several US states legalizing its medical use. The FDA has already granted Psilocybin the status of “breakthrough therapy”, meaning it will be fast-tracked as a potential cure. Psychedelics is poised to be a much bigger market than cannabis.

For a good understanding on the topic, I recommend reading this book by best selling author Michael Pollan.

Aside from these areas, I have also been exploring foodtech, autonomous vehicles, batteries and spacetech. I still haven’t formulated a thesis though, so I will keep on learning by investing before sharing more.

Cheers,

Rand

If you liked what you read, please share this on social! My DMs are open on twitter.