Editor-in-chief James Mawson considers the long, risky and capital-intensive processes of scaling hardware companies.
It is always worthwhile listening to really smart people – one of the perks of the job on Global University Venturing, Global Corporate Venturing and Global Impact Venturing is that basically means everyone in our network – as they engage in making the world a better place.
So, my thanks to Aaron Brandt and Nathan Pascarella at Hypertherm Ventures and its partner, Mach 49’s David Moran and Paul Holland, for the invitation to attend its webinar this week, including star speakers CEO Evan Smith, NEA managing partner Scott Sandell and Vladimir Bulović, professor of electrical engineering and computer science at Massachusetts Institute of Technology (MIT).
While not the most diverse group on certain metrics, it was fascinating how industrial group Hypertherm over the past 50 or so years has grown out of the founder’s garage into one with 1,800 people focused on people, planet and profits.
The $30m Hypertherm Ventures fund is targeting early-stage deals in advanced manufacturing, which is a difficult place to back with limited capital and an awareness of the lengthy time horizons required.
Think of the time from patent to commercialisation of say zips or velcro and it is easily beyond a decade (12 and 13 years, respectively, according to Bulović). The challenge is moving from idea or prototype to scale and cost efficiencies that can bring.
It is relatively easy for students in a laboratory to whack electrical current through a pickle and see if it glows. But far harder to turn this lightbulb moment into organic light-emitting diodes (OLED) where the voltage can change the colour.
Venture capitalists can turn to software rather than hardware for faster proof points of whether the idea can scale, which is why just 3% to 5% of VC cash goes to hardware. Corporations in physical tech, however, still need the new ideas to flow.
Bulović said the traditional 10-year innovation cycle usually looked like ideas form over a few years at university before a paper is presented at a conference where corporations can listen in for potential ideas and take them back to the corporate research and development labs for a half-dozen more years work and then production.
Long, risky, capital intensive. All of which are unappealing especially if the average tenure of a hardware company CEO is six years, Bulović added. Instead, universities are turning to startups for the middle years of development. Here, some seed or early-stage capital allied to lab space at places like Bulović’s MIT Nano facility can reduce costs and prototype ideas, such as quantum dot OLEDs or carbon nanotube chips. (In this light it is no surprise MIT committed $25m and $35m, respectively, to The Engine’s first two deeptech-focused funds. You can listen to Lesley Millar-Nicholson, director of the Technology Licensing Office and Catalysts at MIT, share more in an interview with GUV editor Thierry Heles.)
The corporation can then buy, invest in and/or license the technology and scale it through their distribution platforms, just as Samsung did with QD Vision and Kateeva for OLEDs. It is a model other universities through the TenU network are also looking to develop and GUV will be reporting on these plans in January’s update from the community.