Lenny's Podcast

Why we’re at the beginning of the AI hardware boom | Caitlin Kalinowski (ex–OpenAI, Meta, Apple)

with Caitlin Kalinowski
17 May 2026 18 min read 1h 47m
TL;DR

We're entering the AI hardware boom because digital AI is approaching saturation—the next frontier is physical systems like robots, drones, and manufacturing. The biggest challenge isn't the AI; it's rebuilding U.S. manufacturing capacity and supply chains for critical components like actuators and magnets, which have been outsourced to Asia for 25 years.

Key Moments
Caitlin Kalinowski
“There's a dawning realization, especially in the labs, the acceleration is going so vertical that what you can do behind a keyboard with AI is going to saturate. When that happens, the next frontier is the physical world. Robotics, manufacturing, industrialization.”
Explaining why hardware and robotics are suddenly the focus of every AI lab
Caitlin Kalinowski
“In hardware, we only get to compile our code quote unquote like four or five times. And... total, ever.”
Contrasting hardware development with software's iterative approach
▶ 10:20
Caitlin Kalinowski
“I do feel that we need to re-industrialize the country significantly in order to be safe in a military sense. You really never know what's going to happen in the future. And people that are your allies now may not be in the future.”
On U.S. supply chain vulnerability and geopolitical risks in robotics
▶ 21:46
Lenny
“Just imagine 100,000 drones coming out of China just at us. I do feel that we need to re-industrialize the country significantly to be safe in a military sense.”
Discussing the existential threat of advanced drone warfare
▶ 0:25
Caitlin Kalinowski
“The humanoid robots are still prototypes. Um, and they're advanced prototypes. What we need to do is show that this works at all, which is kind of where we're at right now.”
On the current state of humanoid robots and safety concerns
▶ 15:11
About Caitlin Kalinowski
›
Caitlin is a legendary hardware leader who was part of the original unibody MacBook Pro team and technical lead on MacBook Air and Mac Pro at Apple. She led Meta's AR glasses hardware team (including Orion) and their VR hardware division, and most recently built OpenAI's robotics and hardware division from scratch. She's spent decades at the intersection of AI, consumer hardware, and physical systems.
Takeaways
1
Hardware has 4-5 compile cycles maximum Unlike software engineers who can deploy daily, hardware teams redesign in CAD and physically manufacture only 4-5 times before final mass production. This forces extreme conservatism—every reliability check, tolerance stacking, and variance must be solved upfront because you cannot ship updates to millions of devices already in the field. This is why hardware development is fundamentally slower but also why it creates defensible moats.
2
Actuators and magnets are the critical bottleneck The supply chain for physical AI depends on foundational components like motors (actuators) and rare earth magnets that have been almost entirely outsourced to China and Asia over 25 years. Without domestic actuator manufacturing and magnet processing, the U.S. cannot scale robotics, drones, or manufacturing at speed—making this a national security issue, not just a business problem.
3
Soft robotics is a safety requirement, not a feature Humanoid robots operating near people must prioritize compliance and mass reduction to lower impact energy if they malfunction or are compromised. Lighter, softer arms dramatically reduce injury risk compared to rigid industrial designs. This design constraint forces roboticists to rethink actuator types and materials—another reason supply chain independence matters for safe deployment.