Many of you know that one of my passions is helping technology cross the so-called “valley of death”. I love science and the potential impact it can have on people and the planet. But there are so many barriers in the way before this impact can happen.
And so, I’ve been thinking deeply about when a technology idea should move from academia to industry. There are different incentives and goals in academia than industry. Academia values new knowledge creation and wow factor in the form of an academic publication. Grants are given to PI’s based on their publication record and citations. Graduate students need to ultimate graduate and demonstrate their own publication record for future positions. Note that there isn’t a true incentive for whether the science touches the end user. Companies, on the other hand, only make money if there is a paying customer at the end of the day that benefits from the technology. And so there is a strong incentive to get the proof of concept to a useful product stage.
I’ve been lucky to speak with employees/founders from several deep-tech startups over the years including Astranis, Heirloom, insitro, Ayarlabs, Perceptra, Glass Imaging, etc. Some of these are spinouts of academic research (Ayarlabs, Perceptra, Heirloom), while others are founded by seasoned industry professionals. After speaking with people on both sides of academia and industry, here are some questions I think every scientist should answer before they take the leap out of academia:
“Gates” to the valley
Does the technology need commercial scale effort to grow? Has it outgrown academia? Commercial scale effort means focused resources, large amounts of funding, domain expertise, industry partners.
Is there a strong proof of concept? Is the concept scalable?
Has there been an actual market identified? Some of this work happens through the startup formation process and programs like iCorps. But ultimately if you cannot find a market, there’s no viable commercial path for the technology.
What is limiting the technology? Is there just “Engineering” left? i.e. the technology will make it to the customer with appropriate development requiring time, money, people, and resources.
One technology successfully crossed these gates from my previous undergrad & masters lab at MIT and became Ayarlabs. The POEM (Photonically Optimized Embedded Microprocessor) project was a DARPA funded decade long research project. The goal was to demonstrate a superfast photonic link between processor and memory that could beat copper. You can read about the success story here about how the founding team found the their value proposition (low energy) and beachhead market (data centers). (Disclaimer - I had NO involvement and claim no credit. These are entirely my own naive opinions.).
But I want to focus on the above questions that may have qualified the technology to enter the MIT Energy Ventures class in the first place. After 10 years of development across multiple labs, the technology had been sufficiently demonstrated. It now needed a focused team and funding to take it the next stage. The proof of concept had been met. The solution was scalable since it can be mass produced via semiconductor manufacturing. Some initial applications (UAVs, satellites) had been identified (after all, it was a military funded project). The next need was developing the technology further for specific applications and building a team to do so. The fundamental science limitations had been overcome (I think).
Some times it may actually be better for technology to stay in academia. Perhaps the market is not primed or ready. Perhaps there are better non-dilutive funding sources available through grants. Perhaps the technology needs to be further de-risked. For example, one of my research projects has a provisional patent filed and could be potentially useful for commercial applications, but my team is still struggling with accurate reconstructions of our biological samples. The technology risk hasn't beens sufficiently mitigated, so this is still a research problem. We need to demonstrate a stronger proof of concept before jumping into the next stage.
Death in the valley
Even after passing through the gate with flying colors, many startups fail through the “Valley of Death”. During the Valley of Death, startups undergo initial customer acquisition, partnership management, fundraising, team growth, prototype development, etc. Now I want to point out that the valley exists partially for a good reason.
Some “correct” reasons for technology to fail during this period are:
Solution isn’t scalable
Underlying science knowledge/assumptions were flawed
The market doesn’t have a need for the technology. alternative markets also don’t.
The solution is not sufficiently safe or policy compliant
However, some “unfortunate” reasons that deep-tech startups fail:
Founders are inexperienced/not knowledgeable about how to cross.
Investors aren’t interested in deep-tech that require large capital and long turnaround (longer than fund cycle)
Difficulty hiring/maintaining a good team and creating a strong positive culture
Policy/regulatory barriers are too high for a startup to overcome
Deep-tech pilot and product development require infrastructure and resources that aren’t accessible. (ex: lab space, specialized equipment)
Last, but not least, running out of $$$$.
Hope for the future?
There are some resources available that assist with these pitfalls. Such as the Activate Fellowship to train scientist entrepreneurs and specific venture funds for deep-tech like Breakthrough. However, part of me becomes a little bit more jaded every time I see a good idea and good team fall through the cracks because they lacked connections, resources, and funding.
How can we do better so that science breakthroughs ultimately help people?
More reading: