The AI Stack Every Founder Is Betting On in 2026 (and Why) | Apoorva Pandhi

Описание: Many founders believe winning in AI means building better models. In reality, the real advantage is being built elsewhere in the stack.

In this episode of Post Money Podcast, Nilanjana Bhowmik sits down with Apoorva Pandhi, early-stage investor at Zetta Ventures, to unpack how the AI stack is actually evolving and where founders and VCs are placing their bets heading into 2026.

Apoorva explains why foundation models are becoming commoditized, where defensibility is forming in data infrastructure and workflows, how agents differ from copilots and traditional SaaS, and when AI companies must go full-stack by building their own models. He also shares how Zetta evaluates AI startups today, how portfolio construction is changing in an agentic world, and what founders consistently get wrong when pitching AI businesses.

CHAPTERS:

00:00 Commercializing AI (not building models)
04:12 The modern AI stack explained
07:42 The 3 layers Zetta invests in
08:48 Why foundation models are becoming commodities
12:29 Data ownership and avoiding vendor lock-in
13:13 Data infrastructure as a long-term moat
22:27 Full-stack AI example
37:20 How VCs are building AI portfolios in 2026

-----------------------------------------------
Listen on Your Platform of Choice:

Apple Podcast: https://podcasts.apple.com/us/podcast...
Spotify: https://open.spotify.com/episode/3E05...
Substack: https://open.substack.com/pub/postmon...

-----------------------------------------------

Connect with Apoorva on LinkedIn:   / apoorva-pandhi  
Zetta Ventures: https://www.zettavc.com/

About Apoorva Pandhi:

Apoorva Pandhi is an early-stage investor at Zetta Ventures, a fund focused on AI, data infrastructure, and AI-native applications. Prior to Zetta, Apoorva invested at Lightspeed and Foundation Capital and previously worked as an operator in NLP-driven companies, leading go-to-market efforts for AI products. His work spans data infrastructure, AI tooling, and full-stack AI applications across biotech, enterprise software, and developer platforms.

#artificialintelligence #venturecapital #founderadvice