takeaways from bob mumgaard, ceo commonwealth fusion

keynote talk at mit tough tech summit 2023

michael raspuzzi
7 min readJan 4, 2024

commonwealth fusion (cfs) is one of the few generational startups that could change how the world works in the next hundred years by what they accomplish in the next decade.

as a spinout of mit and funded by the engine, they’ve developed a system to explore how to create fusion — what powers the sun — in a box. moving beyond a science r+d project, they’ve hit major milestones while scaling to 500+ employees in the last 5 years with $2B capital raised:

from cfs.energy

at mit’s tough tech conference, bob’s keynote talk was the highlight of the whole summit. he shared his candid journey from phd student to ceo with bits of wisdom from building cfs as well as perosnal character development.

diagram of the arc with hts magnets from cfs.energy

here are my top takeaways from his talk

1. build an ecosystem from day 1

the team knew from day one that it would take more than a private company’s effort to reach commercial success in fusion: science r+d, talent, support from regulators, and more.

they took an ecosystem approach by looking beyond immediate commercial partnerships, to build relationships to connect people who wanted (and needed) to support their mission.

for example, they are still closely connected to mit’s plasma science and fusion center, where the initial core idea and technology spun out, for continued scientific r+d.

mit’s the engine fund has helped spin up ecosystem support as all of their portfolio companies are tackling hard problems with different community programs.

2. clear vision = motivation

vision is a unity of purpose. he shared that the team has been continuously motivated by the long term vision of carbon free, abundant energy, and fueled by the purpose to tackle all of the shorter term challenges to make that vision a reality.

he shared that when they started team building and raising capital that the story that told itself as ‘getting fusion to work on earth is one of the greatest stories ever told that respects the laws of physics. if it’s possible with the sun, then it’s inevitable we’ll figure it out.’

pair the bold vision of the technology with one of the best market opportunities (energy is 13% of all GDP), and it’s a compelling story for investors to support too.

side note, it makes me think of yuval’s book sapiens where he shares that entire epochs are defined by collective stories like this one.

3. 10 year master plan

this one surprised me. it also was a point that came up more than once for the more mature companies from the engine’s portfolio.

from napkin sketch over dinner for the initial idea to setting up the core plan for all the milestones needed to make cfs possible, it took about 2 years to map it all out.

6/10 years into it, he said they are still on track.

let that sink in. they are on track with what they mapped out pre-day 1.

most people when they think of startups they think of the chaos of uncertainty where different market forces or constraints play a role in figuring out product roadmap or organizational direction.

for companies with more ambitious scientific and engineering roadmaps, it seems important to forego the ‘lean startup’ playbook and focus more on master plan like thoughtfulness, which makes me think of tesla and this piece from tim urban on what makes the chef v. line cook: being able to reason up from first principles.

4. start with the end in mind and work backwards

this is especially important for tough tech as bob defines tough tech as building “$B things,” from fusion generators to large scale manufacturing for green steel or clean cement, these are capital intensive builds.

as they require technical de-risking, and sometimes market de-risking, the end goal needs to be clear in order to understand the milestones to get there as well as the resources that need to come together.

5. work in milestones, concurrently

proving hts magnets could be commercially produced was a milestone that helped them raise more funding. reaching one helps push the other, and each can be developed alongide each other in parallel v. in series.

6. break out of simple and build what it needs to be to work

every step requires a series of problems to be solved.

he mentioned they can’t move at the speed of a marketing budget, like many software companies, they move at the speed of a physical technology company.

as ceo, he’s had to ensure that things were not over simplified and turned in to binary thinking. he had to ensure early team members and investors were onboard with the journey.

along the way, he’s mentioned different mindsets that he’s found particularly helpful:

a. figure it out

b. find the truth, and do what makes sense

c. think clearly about next steps

d. seek perspective from thought partners to challenge assumptions

e. constantly seek self critique

f. be resilient

g. act courageously

7. understand phase of development to anticipate what kind of problems

for example, as they build sparc now they have ‘jack hammer’ problems as he called them ie. construction challenges. these are different than novel engineering breakthroughs to make the initial magnets work in 2021.

8. be so bold it makes (some) people uncomfortable

they knew how much capital they needed to raise in the first few years of development, so they initially explored if they could raise it all up front.

one way for early stage startups to raise money is called a s.a.f.e. note. this was pioneered by y combinator, an accelerator currently offering $500k on their safe.

the cfs team asked themselves if they could raise $2B on a safe note initially… 4,000 times the best offer in the current market.

wild questions and ideas extend beyond pushing the limits of science + engineering as they explore building every part of their business.

9. stay one step ahead

he mentioned one of his top goals as ceo is to always be one (sometimes two) steps ahead.

on the organization side, that includes partnerships, funding, and hiring. on the technology side, that connect back to #3 as well as anticipating challenges in #7.

he’s always thinking what and who needs to be involved to increase the probability of success for keeping up momentum to the next step.

10. think interesting thoughts

he said he is always feeding his curiosity to steer his thoughts in interesting directions. the latest being the history of how the physical infrastructure of the post office was designed and developed.

11. when doing something new, there is no pattern recognition

while he could draw inspiration from aerospace or architecture for the construction of facility, there is no direct blueprint to copy for a first of a kind fusion plant.

the team at commonwealth fusion had to be sure to bring together the right people with different backgrounds to figure it out.

12. good characteristics of a tough tech ceo

in a candid story, he shares that initially as they raised more capital, they were unsure who would be ceo.

while they had brought in different candidates for the week to work directly with the team to see how it went, they ended up deciding that bob was the best candidate.

these criteria, which he mentioned for making the decision, can be generalized for other technical ceos:

  1. understands the fundamental science
  2. know where the technology and team is headed to help lead through uncertainty
  3. committed to being there a long time
  4. knows the technology and the market

13. different people at different times will need new roles.

it’s part of human nature to evolve with a new system and adapt. when that doesn’t happen, it’s important to ask ‘what needs to be true now for the company to succeed?’

14. check in what’s important daily, make it a ritual

ensuring the days don’t get mixed up with the decades, he says he checks in with himself to see if his values are put into action. it’s a practice to be honest with himself to increase self awareness on what’s working and what’s not working.

this is similar to marshall’s goldsmith (ceo coach) daily question exercise.

one of bob’s core values is self critique, so much so it even made it to one of the core company culture values.

15. explore meta-processing as processes get designed

for new processes:

1. make it work -> proof of concept

2. replicate -> prototype (ex 2)

3. design systems to replicate -> scale (ex 3)

when building a tough tech company, it also builds you

it’s interesting to see how the journey from phd student to ceo shaped bob’s mental models, values, and what he thought about. in less than an hour, he packed in more stories and more principles for innovating than entire podcasts and research papers focused on the topics.

bonus: blog on unconferences if you want to run your own summit

if you’re curious how to run a 600+ summit that feels as intimate as a great dinner party, here’s my breakdown of the unconference style event covering behind the scenes for mit’s tough tech summit.

hey, i’m michael, and i’m building new ways to help the next generation of innovators. connect with me on linkedin to see what i’m cooking up next.



michael raspuzzi

building something new. previously @tks @harvard @culinaryai