Sourcery
Sourcery

Nov 16, 2026 · 1h 6m

AI Competition Moves From Model Scores to Sticky Products

Dylan Patel on NVIDIA's Future, SemiAnalysis & His $500M VC Fund

The episode connects AI’s next phase to physical infrastructure, product design, regulation, and the credibility challenges facing research and investment firms.

3 key takeaways
  1. 1AI companies increasingly compete through product quality, persistent applications, enterprise contracts, and user experience rather than isolated model benchmarks.
  2. 2Data-center expansion faces permitting, water, power, labor, noise, and community opposition that can delay infrastructure needed for AI growth.
  3. 3Dylan Patel is building a combined research, consulting, media, and investment platform anchored by a $500 million hardware-focused fund.

Don't miss

Patel lays out why his $500 million fund focuses on hardware and infrastructure, where specialized supply-chain knowledge can shape investment decisions.

The brief

Dylan Patel maps AI’s infrastructure economy from NVIDIA and Chinese chip scrutiny to data-center permitting, power, water, and community backlash.

The conversation’s central shift is from model-release scorekeeping to products that retain users, support persistent agents, and win durable enterprise contracts.

Patel says regulation and local opposition can push data-center projects elsewhere, even as communities weigh tax revenue against water use, grid strain, and noise.

His $500 million early-stage fund targets hardware and infrastructure, including chips, robotics, optics, manufacturing, networking, power, and packaging.

The harder institutional question is whether SemiAnalysis and Citrini Research can combine attention, consulting, media, and investing without compromising independent analysis.

Weekly email

State of the Pod, every Monday

The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.

Listen to the full episode and explore every guest, topic, and moment on PodLume.