Pointing superintelligence at the hardest problems in physics
2822812.7K641K607
Tags
AI agentSeed500K-1MB2BUSFunding announcementFounder-led
TLVC Rating
4.00
Hook
Editing / Creativity
Copy
Sentiment of launch
Distribution strategy
Community Rating
No ratings yet
to rate this launch.
Physical Superintelligence (PSI) is a Cambridge, Massachusetts research lab building an AI system that runs physics work at machine scale, aimed at researchers, mission planners, and industrial teams whose progress is bottlenecked by hard physical modeling. Matt Pines, Alex Klokus, and Dr. Alexander Wissner-Gross launched Physical Superintelligence (PSI) on September 1, 2026, announcing $58 million in seed funding led by Breakthrough Energy Ventures. Other backers include Dragon Global, Robot Ventures, Solari, Susa, SV Angel, Valkyrie and several individual investors. The company is structured as a public benefit corporation, and the launch matters now because PSI is stepping out of stealth with both a commercial platform and a live mission attached to it.
At the center of the lab is a platform called Emmy. Its core platform, Emmy, is named after physicist Amalie Emmy Noether and is designed as a collection of virtual physicists operating on a sovereign reasoning engine supported by a curated inventory of simulations. The idea is to compress the loop between hypothesis, simulation, and validated result so that new models and engineered systems can be produced far faster than a traditional lab could manage, with the goal of eventually surfacing new physical laws rather than only refining known ones.
The first proof point is aerospace. PSI is the founding technical partner for the Fermi Explorer Mission, which it describes as the first AI-planned interstellar mission to Alpha Centauri. PSI validated the mission's physics and identified a substantially more efficient trajectory within the mission's mass and budget constraints. From there, the team plans to productize the same platform for terrestrial and space applications, positioning PSI as an infrastructure layer for organizations that need frontier physics work done on a timeline that human teams alone cannot hit.
Comments (13)
Sign in to join the discussion.
N
Nikhil R. VoraSep 3, 2026
Naming your company Physical Superintelligence is a wild flex. If the next paper isn't titled 'we solved turbulence' I'm asking for a refund.
M
Marta KowalczykSep 3, 2026
Tech journalist here, would love 15 minutes on what 'commercializing physics breakthroughs' actually looks like in year one. Fusion? Materials? Something spicier?
D
Deandre HollisSep 3, 2026
naive question but... what counts as a 'transformative physics breakthrough'? like, room temp superconductors tier or 'we made a better solar cell' tier?
Z
Zeynep A.Sep 3, 2026
ok wait, a seed round the size of most Series Bs and the announcement is a single tweet with a link. respect the restraint.
G
grigorSep 3, 2026
so where's the demo. a landing page and a tweet does not a research lab make. show me one prediction the model got right that a grad student couldn't.
T
Tao WenliSep 3, 2026
The impressions on this are climbing fast but the reply-to-like ratio is suspiciously calm. Either everyone is intimidated by the word 'superintelligence' or the physicists are still reading.
O
Oni AdebayoSep 3, 2026
meanwhile I'm 3 weeks out from launching a todo app and these guys are pointing AGI at the Navier-Stokes equations. cool cool cool.
K
Kenji MorrowSep 3, 2026
no launch video, no teaser, just a link and a wire transfer confirmation energy. honestly iconic editing choice.
P
Priya RamanathanSep 3, 2026
curious what the smallest version of this team could look like. 3 physicists in a garage with an api key? asking for a friend who is me.
M
Mikko JärvinenSep 3, 2026
any plans to open-source the eval suite for physics reasoning? the community needs benchmarks that aren't just 'did it pass the AP exam'.
L
Lin HaoyuSep 3, 2026
building something in the same universe (ML for computational chemistry). rooting for you, also mildly terrified you'll eat my lunch by Q3.
Y
Yulia SokolovaSep 3, 2026
who's running research hiring? got a condensed matter postdoc who codes and doesn't hate industry, which is apparently a unicorn combo.
F
Femi OkonkwoSep 3, 2026
hot take: the hardest problem in physics is getting three cofounders to agree on a tagline, and they nailed it on the first try.
Work With Us
Book an intro call to explore our suite of services. We might have a waitlist.
Physical Super intelligence
Pointing superintelligence at the hardest problems in physics
Pointing superintelligence at the hardest problems in physics
Tags
TLVC Rating
Community Rating
No ratings yet
to rate this launch.
Physical Superintelligence (PSI) is a Cambridge, Massachusetts research lab building an AI system that runs physics work at machine scale, aimed at researchers, mission planners, and industrial teams whose progress is bottlenecked by hard physical modeling. Matt Pines, Alex Klokus, and Dr. Alexander Wissner-Gross launched Physical Superintelligence (PSI) on September 1, 2026, announcing $58 million in seed funding led by Breakthrough Energy Ventures. Other backers include Dragon Global, Robot Ventures, Solari, Susa, SV Angel, Valkyrie and several individual investors. The company is structured as a public benefit corporation, and the launch matters now because PSI is stepping out of stealth with both a commercial platform and a live mission attached to it. At the center of the lab is a platform called Emmy. Its core platform, Emmy, is named after physicist Amalie Emmy Noether and is designed as a collection of virtual physicists operating on a sovereign reasoning engine supported by a curated inventory of simulations. The idea is to compress the loop between hypothesis, simulation, and validated result so that new models and engineered systems can be produced far faster than a traditional lab could manage, with the goal of eventually surfacing new physical laws rather than only refining known ones. The first proof point is aerospace. PSI is the founding technical partner for the Fermi Explorer Mission, which it describes as the first AI-planned interstellar mission to Alpha Centauri. PSI validated the mission's physics and identified a substantially more efficient trajectory within the mission's mass and budget constraints. From there, the team plans to productize the same platform for terrestrial and space applications, positioning PSI as an infrastructure layer for organizations that need frontier physics work done on a timeline that human teams alone cannot hit.
Comments (13)
Naming your company Physical Superintelligence is a wild flex. If the next paper isn't titled 'we solved turbulence' I'm asking for a refund.
Tech journalist here, would love 15 minutes on what 'commercializing physics breakthroughs' actually looks like in year one. Fusion? Materials? Something spicier?
naive question but... what counts as a 'transformative physics breakthrough'? like, room temp superconductors tier or 'we made a better solar cell' tier?
ok wait, a seed round the size of most Series Bs and the announcement is a single tweet with a link. respect the restraint.
so where's the demo. a landing page and a tweet does not a research lab make. show me one prediction the model got right that a grad student couldn't.
The impressions on this are climbing fast but the reply-to-like ratio is suspiciously calm. Either everyone is intimidated by the word 'superintelligence' or the physicists are still reading.
meanwhile I'm 3 weeks out from launching a todo app and these guys are pointing AGI at the Navier-Stokes equations. cool cool cool.
no launch video, no teaser, just a link and a wire transfer confirmation energy. honestly iconic editing choice.
curious what the smallest version of this team could look like. 3 physicists in a garage with an api key? asking for a friend who is me.
any plans to open-source the eval suite for physics reasoning? the community needs benchmarks that aren't just 'did it pass the AP exam'.
building something in the same universe (ML for computational chemistry). rooting for you, also mildly terrified you'll eat my lunch by Q3.
who's running research hiring? got a condensed matter postdoc who codes and doesn't hate industry, which is apparently a unicorn combo.
hot take: the hardest problem in physics is getting three cofounders to agree on a tagline, and they nailed it on the first try.