Brad Porter

Excited to introduce Proxie, our first cobot, to the world today!

@tonyzzhao 1 million robots at Amazon says this can be solved. But very few outside Amazon know how to do it. I’m hosting the people who know how to do it on my podcast, Deployed. podcasts.apple.com/us/podcast/dep…

Deployment is what matters. Deployment is not solved.

@samanfarid @tonyzzhao I’ve deployed a few hundred thousand robots myself.

@tonyzzhao Love what you guys are doing… but speak for yourself sir. We’ve deployed autonomous robots in 100+ factories running 24/7

"Does OpenAI go to zero the way Netscape did?" @fubini asked me this. We both interviewed at Netscape the same day in '95. His answer: "The “homepage” for AI hasn't shown up yet." Deployed EP4 is live. Link in comments.

@hamptonism 😂 in 1996 I spent the summer with 4 dudes in a 2br apartment because I wanted to help bring the interment to everyone at Netscape. I don’t regret it. x.com/bradporter_/st…

this could have been you europe-maxxing this summer but instead you chose to share a 1bdr with four other dudes in sf to walk around the city with your laptop slightly open so your agents won’t stop running.

Two nerdy kids interviewed for Netscape internships the same day in 1995. One built Amazon's robotics org. The other wrote the first check into Anduril. Tomorrow we talk about what 30 years taught us about spotting opportunities before they're obvious. Deployed EP4 with @fubini drops tomorrow.

@bilalfarooqui Khosla

Lots of DMs asking “who is a tier 1 VC” so in no particular order: - a16z - thrive - general catalyst - accel - benchmark - sequoia - founders fund - kleiner - first round - lightspeed - bessemer Who did I miss?

@samanfarid This alone could drive a 15% improvement in national productivity!

How is google slides dropping the ball sooooooo incredibly horribly.

@bznotes sees this pattern constantly: Founders who solve the science and assume the business follows. It doesn't. Product → customer → revenue → margin. Most deep tech pitches break somewhere in that chain. The founders who make it are the ones obsessed with closing the gap between what the technology can do and what someone will actually pay for. That's the conversation in Episode 3 of Deployed. Link in comments.

@lukas_m_ziegler @adcock_brett So much speculation off a chart with no Y axis and uneven variable height bars.

Forbes assumed so based on Omdia report; quote: „However, given that an Omdia research report recently estimated Figure’s 2025 shipments at 150 and that the chart Adcock shared includes data all the way back to 2023…” the rest in the linked Forbes article. Curious indeed too whether it’s really shipment data.

BREAKING: Figure is doubling humanoid robot shipments EVERY SINGLE MONTH. @adcock_brett shared the data himself. The company shipped around 150 units in all of 2025. Based on that baseline, the last few months likely look something like this: → Feb 2026: ~60 units → Mar 2026: ~120 units → Apr 2026: ~240 units And their BotQ facility is rated for up to 12,000 units per year. The hardware is serious too. Figure 03 has cameras in each hand, fingertip tactile sensors detecting forces as light as 3 grams, and achieved a 90% component cost reduction vs Figure 02. That last number is the one that matters. That's the cost curve that makes consumer-scale deployment possible. For context, the ENTIRE global humanoid robot market shipped under 14,000 units in 2025. It's forecast to hit 2.6 million annual units by 2035. Exponential curves are strange things. If Figure keeps doubling for even two or three more months, the gap between it and the leading Chinese competitors starts to look a lot less permanent. THE RACE IS HEATING UP. The Forbes article with the numbers: forbes.com/sites/johnkoet…

@samanfarid @JonMSchwartz So many rumors, so few deployments.

@JonMSchwartz lol the number of companies that got multi-hundred $M LOIs from UPS then promptly delivered ~nothing but raised a few hundred $M in the process....

“Wants to buy” is far from “bought” or “is currently using”… Who wouldn’t want to buy tens of thousands of useful and appropriately priced robots?

There's a type of founder most VCs don't know how to evaluate. They carry technical risk and commercial risk simultaneously for years, Even before they know if the market is real. @bznotes spent $700M on those founders. His best seed bet is now worth $15B. Deployed EP3 is live. Link in comments.

@aphysicist 💯 the constraint in robotics is the number of people with actual field deployment experience is shockingly low. This was the inspiration for my podcast, Deployed. co.bot/news/Introduci…

time for a gemba walk boys. we need to tour these guys around some factories so they can see what roles robots have in production. tbh we should do this with most sf ai ppl who want to do robotics, their heads would explode. should add this to the summit

My investor @bznotes told me something that still sits with me: "I'm not judging your smartness. You passed that bar long ago. What I'm looking at is whether you're prepared for what's coming at you." Then: "I don't know what's coming. And you don't know either." That's the real filter in deep tech. Not brilliance. Preparedness for the unknown. Full conversation on Deployed Ep 3 Live Today.

The venture industry spent a decade perfecting SaaS. Meanwhile, the physical world: automotive, manufacturing, construction, logistics, just sat there. Trillions of dollars. Barely touched by software, just waiting. @Bznotes didn't wait. In Episode 3 of Deployed, he breaks down why the founders entering deep tech right now are fundamentally different, and why the returns will be too. In Bilal's vision, for the first time, the caliber of founder who used to default to enterprise software is looking at the physical world and making a different choice: If I'm going to take this kind of risk anyway, I might as well build something that outlasts me. Not a Star Wars future. Not an iRobot future. Something a customer would actually pay for.

If this were just data ops challenge with no open science, then I feel like we should already be further along. Scale AI is a very good at robotics data. Generalist showed 100s of thousands of hours are not showing open world generality. PI0.7 is impressive as is DreamGen for their early hints of generality, but feels like still more science to be done.

nobody has built the @scale_AI for physical AI. eight years after @CovariantAI proved the data flywheel is the business. eight years. the reason is obvious once you see it: simulation looks like it should work. reduce 500 real demonstrations to 50, close the gap with compute, ship. @nvidia has more sim-to-real resources than anyone. their own documentation says the reality gap is still the core problem. physics doesn’t care about your render quality. force signals, contact dynamics, failure modes: simulation generates plausible versions of all three. a deployed robot finds the difference in the first hour. the companies that got it right stopped trying to solve simulation fidelity and built data operations instead. Covariant. @physical_int . the architecture is the research contribution. the data operation is the actual business. what needs to exist: dedicated manipulation data infrastructure. teleoperation rigs, tactile sensors, failure annotation, sold as a service to the foundation model labs who need contact-rich data but won’t build this themselves. operationally harder than language data. structurally more defensible. almost entirely unbuilt. the moat compounds with every deployed robot hour. the window is open.

@kvablack The physical ai space has been showing “hints of generalization” for the past few months. DreamGen, Generalist GEN-1, SKILD… but true human-level generalization still feels elusive. Suspect the world model questions continue until someone shows that. But congrats regardless!

can VCs stop asking us about world models now?

@taiyasaki @theworldlabs Fascinating time. The foundation model companies are looking at going vertical, but so is Jensen. There’s a good argument as the marginal cost of software and AI goes to zero, the value will be in hardware and services.

Now that this is open source… what will the 1B$+ of @theworldlabs be used for?

@Alfred_Lin I grew up in rural upstate NY… when we moved there in 1980 the small hamlet had 28 small to medium family farms. When I left for college there were <5 very large farms, driven by productivity and market pressures. The world changes. We adapt.

@BradPorter_ My father got the calculator. We got the internet. Our kids get AI. The question is always the same: are you the farmer who adapts, or the one who doesn't?

Deployed Episode 2 is live. Watch it now @Alfred_Lin's father was destined to be a farmer. The abacus got him a job at a bank. Then calculators replaced the abacus. Then computers replaced calculators. He learned each one. Eventually he ran the bank. That's how @Alfred_Lin thinks about Physical AI. Deployed EP2 drops tomorrow.

@hthieblot Cobots!

pitch me your company in 1 word.