8 clips
BBuilding Great Tech
Ian P. Cook argues that AI superiority over China is a critical national security imperative for the United States. He claims achieving this will require building massive data centers spanning entire states, potentially covering areas the size of Manhattan, Rhode Island, or even all of Nevada.
The Twenty Minute VC (20VC)
A discussion of the critical balance AI companies must strike when investing in compute infrastructure. The speaker explains how under-investing can cause you to miss cycles and lose competitive advantage, while over-investing in compute can lead to unsustainable returns and potential business failure.
Equity
A discussion about how companies adjacent to data centers, particularly energy storage businesses, are experiencing rapid growth driven by AI infrastructure needs. The speaker predicts this trend will accelerate in 2026, citing a former Tesla executive's energy division as an example of this opportunity.
VVenture Unlocked
A discussion of how the shift from AI infrastructure to applications creates challenges for seed investors. The speakers explore why capital-intensive infrastructure deals don't fit the typical seed investment model, though some investors like Ed Sim at Boldstar have found ways to make it work.
RRiding Unicorns
Akash Bajwa explains the challenge of evaluating AI infrastructure vs applications, using vector databases as a cautionary tale. He warns against extrapolating short-term growth in hot categories like vector databases into overly bullish long-term assumptions, highlighting how easy it is to get caught up in cycle-driven hype.
EEUVC
The discussion explores how AI models are becoming small enough to run locally on devices like iPhones, potentially shifting compute from centralized data centers to edge devices. The speakers note that developers are increasingly running models locally not just for performance, but because they keep hitting token limits on services like Claude and OpenAI.
Akash Bajwa explains why certain AI primitives like security tools and vector databases aren't building standalone billion-dollar companies, but instead getting bundled into larger incumbent platforms. He points to recent acquisitions like Palo Alto's ProtectAI purchase and Checkpoint's Lakera deal as evidence that these specialized tools work better as features than standalone products.
The speaker explains why AI infrastructure companies like Anthropic and OpenAI create a fundamental mismatch with seed funding, as they require $100M+ capital that seed investors can't provide. This leaves AI application layer companies as the more viable option for early-stage funding.
