Retrofitting Old Factories for AI Won't Work | Adarsh Kulkarni
MTS
November 1, 2026
Foundry Robotics founder and CEO Adarsh Kulkarni discusses raising a $100M Series A to build software-defined, AI-orchestrated factories from scratch.
Adarsh Kulkarni
Something that's 90% reliable fails one out of 10 times.
Something that is 99% reliable fails one out of 100 times. And so something that is 99.99% reliable fails one out of every 10,000 times. So the difference between 90% and 99.99% is a thousand times fewer errors. That's the way you need to think about it. That's why it's so hard.
Sophia
Everyone, and welcome back to MTS. We are back and live with Adarsh Kulkarni, who is the founder and CEO of Foundry Robotics, that just announced their $100 million Series A to build autonomous factories for America's critical industries. Adarsh, welcome to MTS.
Adarsh Kulkarni
Thank you so much, Sophia. I appreciate it. We're excited to build more factories.
SPEAKER_3
Excited to build more factories.
Sophia
I know you're calling in today from Emeryville. What's going on at Emeryville right now?
Adarsh Kulkarni
Everything. What's not going on? We've got parts coming in from customers. We have robots running. We have technicians building. We have rollouts running. We have roboticists collecting data and deploying models. I mean, this is how we build manufacturing for the 21st century full of robots, people and AI and everything in between.
Sophia
This is how we build it all. I want to talk a little bit because in your launch video, you were sharing about this larger problem that's happening right now. You said we've offshored two thirds of all our manufacturing overseas. You say America is not ready and prepared for this new robotics and AI revolution. What's going on in the wider space here?
Adarsh Kulkarni
Yeah. I'll talk about it from the problem space, and I'll talk about it from the solution space. Problem space, you can't just retrofit old factories. One of the things that we think about at Foundry is, you have to build net new factories. We need to build more capacity. We need to build more factories than exist today.
If you're going to build factories from scratch, you might as well figure out what's the most efficient way to do it that includes robotics, AI, and everything in between. So we have robotics foundation models that we deploy. We also have an AI native orchestration layer that does a ton of agentic supply chain and management of the facility itself.
We think about, okay, let's zoom out. The factory is the product. What does that look like? And so there's a lot of tailwinds here. We have a lot of customers in defense, energy, telecom, infrastructure, industrials. People need to build more things, and we need capacity that doesn't exist. On the other side, robots are starting to show up, right? The foundation model labs are pushing the frontier of what's possible every single day. Our job is to take the base models and deploy them in really, really, really important settings like manufacturing. So laundry folding is great. I want one, so does everybody else. But we have to do the work to go from a laundry folding model to a satellite building model and deploy it at scale.
Sophia
Now, you bring up an important point because oftentimes we're seeing these really cool demos of robots doing great things. But then what it actually takes to get these demos into production is a whole other can of worms. So tell me a little bit about what's the hardest part here.
Adarsh Kulkarni
So I think it's worth calling out for robots to really, really make a dent in the economy. We need to talk about very, very real scale and the scale comes in four different types. One is uptime, right? We want to be running 24 seconds, 24-7. The second is in speed. How can we be superhuman in our speed? The third is how can we have four nines of reliability? The fourth is how can we do that with a million robots? All four of those things are complicated. There are four metrics that we hold ourselves to at Foundry that are different from frontier robotics research.
What is challenging is taking what is the most cutting edge research in the world, hardening it to that extent across those metrics and then scaling it up. I think today, the first metric we're tackling is really reliability. How do we get to one nine and then two nines and then three nines and then four nines? A lot of how we think about our harnessing of the models, a lot of how we think about the safety critical guardrails, a lot of how we think about like, all right, cool, we can't have a robot mess up and what happens when it does.
That is a lot of the engineering prowess at Foundry today.
Read the other 11 minutes of this episode here.
$15 gets you this episode and 99 more. Credits never expire. After checkout, one link brings you back here to read it.
Not what you expected? Full refund within 14 days.
Thousands of transcripts fetched by people building searchable podcast archives
Fetch transcripts with code
The same credits work from a script or an AI agent. The demo key returns a sample episode in full, no card needed:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000799226892