Today, Dream Labs is coming out of stealth with Dream-1, a robot foundation model that controls bimanual robots it has never seen, on tasks it was never shown, in places it has never been, from $3/hour. Dream API is live for everyone, today. 🧵

Tech Twitter lately: seed rounds framed as mega-fund call options. Portfolio founders dissed in public. Pro-rata fights going legal. UGH! There are wartime investors and peacetime investors. Our take: founders should find their wartime investors early ↓

New @ThePeelPod with @bznotes! Bilal's spent 18 years investing hard tech, most of it as a GP at Lux ($7B AUM), before starting @RedGlassVC in 2025 to bet on AI that touches the physical world. Bilal was making defense tech investments 15 years ago, and he led the Series A in Applied Intuition, now a $15 billion company. We hung out for a few hours and he shared everything he's learned investing in hardware, why hardware is still hard, why the deployment layer is the most attractive area of physical AI, taking a 90% pay cut to start Red Glass, "AUM as a moat", how to become a billionaire off management fees, and why consensus investing is mostly a circle jerk. Full episode here + links below Timestamps 0:00 Leaving $7B to bet on physical AI 0:51 What most people get wrong in physical AI 4:17 Why a startup has to become a business 11:31 Hardware is still hard, so what changed? 15:40 How government went from SBIR mills to real customer 17:48 Robotics goes from hard-coded to probabilistic 23:34 Where value accrues: full stack, then wedges 25:24 How Applied Intuition became a $15B company 28:03 The deployment layer is the real prize 32:22 OpenAI and Anthropic hint what will happen in physical AI 36:14 Why customers have all the downside 37:56 The best founders paint a picture and execute 40:53 Why it's called Red Glass (founders chew glass) 42:17 Every founder hates fundraising, even Travis 44:29 Red Glass's four core values 50:49 Why founders really take your call (the $250k effect) 53:28 What Bilal looks for in a founder 58:44 Why technical founders underrate fundraising 1:00:26 The Tiger Global era 1:04:02 From $10B to $100B, the goalposts moved 1:06:21 Is this a bubble? Portfolio construction 1:08:55 Building a firm that outlasts him 1:12:13 From Nobel-lab chemist to Zero-to-One investor 1:13:45 Are small funds just outsourced associates? 1:16:35 When AUM becomes the moat 1:18:34 How big funds choke off SPV money 1:21:24 The risk of VCs backing competitors 1:23:31 Why fund economics push everyone bigger 1:26:20 Consensus investing is a circle jerk 1:29:51 You can be a billionaire on fees alone 1:33:09 Equal carry and the right to win 1:35:55 Why good ethics in VC matter

Foundry has acquired @AveaRobotics. @aryind_ and @vikramvadrevu are deeply technical and customer-obsessed and working together (as an early customer) was easy. Data collection and inference at scale power our factories and that makes Avea a natural fit. Welcome aboard!

NVIDIA plays a very unique and central role in the AI ecosystem, including in PhysicalAI. I am grateful for how their executives spend time with startups, investors, and others to share useful insights, research results, and provide feedback on ideas. It’s good leadership.

The machines are becoming better at understanding and responding to the physical world. That is the technology story. But technology alone will not determine where robotics spreads first or fastest.

Speed and rigor are not trade offs for systems that can’t afford to fail. @aerovironment is proving that at scale. nominal.io/blog/aerovironmen…

At dinner tonight with several prominent GPs. While one end of the VC universe is rapidly morphing into large asset managers, getting paid 2/20 economics for buying call options, the other is still doing painstaking work of true early stage venture. Working closely with founders in unproven areas, and developing trusted relationships to help them build. I know where @redglassvc belongs.

Coming soon. 1,000,000 🤖’s.

Saying goodbye to two of our summer interns. Angela and Jason. 🤗