May 17, 2026 · 1h 39m
Hardware veteran warns of imminent memory price shock and robotics bottlenecks
Why we’re at the beginning of the AI hardware boom | Caitlin Kalinowski (ex–OpenAI, Meta, Apple)
As artificial intelligence transitions from digital models to physical robotics, critical hardware bottlenecks and supply chain strains threaten to stall the next wave of tech innovation.
- 1A looming memory price shock threatens to disrupt the supply chain and inflate costs for next-generation AI hardware.
Don't miss
Caitlin Kalinowski details the looming memory price shock and the technical friction of building humanoid robots.
The brief
Hardware pioneer Caitlin Kalinowski, who shaped iconic devices at Apple, Meta, and OpenAI, warns that the robotics revolution faces immediate, severe roadblocks that the tech sector is largely unprepared to handle.
The most pressing threat is a looming memory price shock. As AI hardware demands skyrocket, supply chain bottlenecks and surging component costs threaten to stall the development of next-generation physical computing.
Beyond raw components, building humanoid robots introduces unprecedented engineering friction. True mobility and utility require solving massive power, weight, and mechanical trade-offs that software-centric firms often underestimate.
Navigating these complex physical constraints requires a distinct style of leadership. Drawing on her time with Steve Jobs and Mark Zuckerberg, Kalinowski emphasizes that hardware success demands relentless, uncompromising product design.
What was said on this episode
43 statements · 22 positive · 14 negative · 2 mixed · 5 neutral
AI-enabled digital work will eventually saturate, shifting innovation toward the physical world.
“what you can do behind a keyboard with AI is going to saturate”
Listen at 0:05
War technology will change more than consumer electronics during the next two years.
“There's probably more change in war than there is in consumer electronics. In the next two years.”
Listen at 0:16
The military should invest substantially more in drones than aircraft carriers.
“We need to invest a lot more in drones than in aircraft carriers.”
Listen at 0:21
The United States should significantly reindustrialize to improve military safety.
“we need to re industrialize the country significantly, significantly to be safe in a military sense”
Listen at 0:30
Pixar and Disney are probably the world’s best at designing approachable characters and devices.
“Pixar, Disney are probably the world's best at doing this type of design work.”
Listen at 1:12
VR gaming is an interesting but relatively niche market.
“VR gaming's a really interesting, it is kind of a niche”
Listen at 3:55
VR-derived spatial technologies are being used in robotics for navigation and perception.
“in robotics, all of these technologies are being used because you need to understand how the robot is moving through space”
Listen at 4:01
AR glasses will become part of the future of personal computing.
“I believe in AR glasses as part of the future”
Listen at 5:14
Meta’s Orion AR glasses are not yet ready for mass production because waveguides and microLEDs have insufficient yields.
“Orion, the AR glasses we worked on, I worked on most recently are a bit ahead of their time because they're using waveguides and microleds that are not quite ready for mass production”
Listen at 5:32
Hardware, robotics, manufacturing, and industrialization will follow saturation of digital AI capabilities.
“when that happens, the next frontier is hardware. The next frontier is robotics, manufacturing, industrialization”
Listen at 13:18
Softer robots are safer for operation near people.
“Softer robots is safer.”
Listen at 14:23
Soft, compressible robot limbs reduce impact impulse during collisions.
“if it's soft and compressible, then the impulse is lower”
Listen at 15:01
Humanoid robots remain advanced prototypes rather than production-ready systems.
“in my worldview, the humanoid robots are still prototypes and they're advanced prototypes”
Listen at 15:09
Scaling humanoid robots to hundreds of thousands or millions is a major challenge.
“At Scale is a huge challenge now for me. My background at scale means millions usually, but let's even say hundreds of thousands.”
Listen at 16:25
The United States currently lacks strong actuator companies for robotics.
“we don't have great actuator companies here yet, for example”
Listen at 17:27
Robotics supply-chain layers have largely shifted to China, Japan, and Korea over 25 years.
“each layer of this chain has essentially been outsourced over the last 25 years to countries like China, like Japan, like Korea”
Listen at 18:33
Unavailable actuators can prevent companies from manufacturing robots.
“the actuators themselves are important because if for some reason we can't buy them, then we don't get to make robots”
Listen at 19:42
Military technology should rely on an independent domestic supply chain wherever possible.
“we need to, at least on the military side, have an independent supply chain as much as possible”
Listen at 21:34
Future warfare will increasingly involve rapidly updated drones and 3D printing.
“the future of where war is headed, unfortunately”
Listen at 23:29
Robotics and drone systems must control adversarial threats to their hardware layer.
“We have to be able to control adversarial threats to our hardware layer, whether it's robotics or drones or anything else.”
Listen at 25:09
Quest 2’s cost-focused redesign led to the highest-selling VR headset, in Kalinowski’s assessment.
“what it required is a redesign of the entire product essentially for cost, which then I think led to the highest selling VR headset of all time”
Listen at 32:26
Hardware development is less adaptable to frequent changes than digital product development.
“Hardware is not as adaptable to lots of changes throughout its development as anything digital.”
Listen at 33:29
Customers cannot reliably specify novel products they have never encountered.
“if you want to build something new, customers don't know what they want because they haven't seen it”
Listen at 42:16
Hardware companies should prebuy memory to withstand supply-driven price spikes.
“I have been advising startups and companies to pre buy memory and to have enough memory in stock if they can afford it, to ride out price spikes”
Listen at 45:45
Memory prices will probably double.
“I think prices are going to double probably.”
Listen at 47:17
A single missing component can halt hardware production.
“you can't build anything if you have one component missing”
Listen at 49:28
Verticalized manufacturing helps companies adapt to supply-chain shocks.
“if you have verticalized and you have a lot of the components in house, or you're building a lot of things in house, you can actually adapt to supply chain shocks better”
Listen at 52:27
Hardware teams should use off-the-shelf components whenever possible during prototyping.
“I use off the shelf whenever I can, especially in the prototyping phases.”
Listen at 53:13
AI-assisted CAD is currently at an early stage.
“We're right at the very, very beginning of AI being able to do cad.”
Listen at 55:43
Current language and video models lack physical understanding needed for engineering.
“These LLMs and even video models, they don't know how to do that. They don't have the ability to understand friction or weight or contact, pressure, friction, surface texture.”
Listen at 59:06
Humanoid robots are unsuitable for specialized repetitive manufacturing tasks.
“you wouldn't want a humanoid to do that”
Listen at 1:01:49
Advanced manufacturing needs more dedicated robots rather than humanoid replacements for workers.
“we don't actually need to replace humans with humanoids. We just need more of these dedicated robots.”
Listen at 1:02:37
Large incumbent manufacturers will adopt AI-driven CAD more slowly because of tools and IP concerns.
“the big incumbents are going to be slower because they have dedicated tools and a lot of IP privacy”
Listen at 1:05:11
Robots that signal intent before moving are less alarming to people.
“if a robot looks before it turns and then goes, it's much less alarming”
Listen at 1:08:12
Waymo autonomous vehicles will cause a fraction of the deaths caused by human-driven cars.
“you're going to have a fraction of the deaths using a Waymo”
Listen at 1:11:04
The number of robots operating in public spaces will continue increasing.
“you're going to see more and more robots”
Listen at 1:14:11
There will not be 20 million robots within five years.
“I'm not somebody who thinks that in five years you're going to have 20 million robots.”
Listen at 1:14:14
OpenAI’s Department of War deal announcement lacked adequate governance guardrails.
“the lack of defined guardrails around the announcement of the Department of War deal is not how I thought it should have been done”
Listen at 1:16:23
Effective AI hardware teams require both senior and junior employees.
“when you actually build these teams is you have to have both. You must have both.”
Listen at 1:22:12
Mission alignment helps unify AI research and hardware teams.
“mission alignment, which actually unifies a team”
Listen at 1:22:47
Meta maintained speed by delegating technical decisions to the lowest possible organizational level.
“the decisions were made at the lowest level possible in the company to maintain speed”
Listen at 1:26:01
Daily AI-tool use will help individuals remain at the forefront of future work.
“if you embrace it and are daily using these AI tools right now and daily applying them to what you're doing, you'll be at the forefront of whatever comes next”
Listen at 1:32:06
AI helps with basic learning but inadequately explains cultural context and significance.
“reading and communicating with AI is very helpful on the basics, but then understanding what was happening culturally and what the significance of the work isn't adequate”
Listen at 1:37:10
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.