Podcast Digest
Weekly summaries from the podcasts I follow — macro, markets, tech, and geopolitics.
Week of 16 March 2026
12 Podcasts CoveredThis Week’s Dominant Themes
The Iran War Shock
Dominates macro discourse this week. Energy price spike, missile stockpile depletion, cyberwar, oil at $200 tail risk, and Gulf/China secondary effects discussed across Eisman, Odd Lots (4 episodes), All-In, and Dwarkesh.
Private Credit Stress
Eisman dedicates two episodes to emerging cracks in private credit and the PE/life insurance nexus (Apollo, KKR, Brookfield captive reinsurance scandal). The Synopsis flags Constellation’s sell-off as AI-panic-related rather than fundamental.
AI as Disruptor vs. Enabler
The software sell-off is the battleground: Figma (down 80%), Duolingo (down 80%), and SaaS broadly are being indiscriminately de-rated. BG2 and All-In push back — Anthropic/OpenAI revenue data suggests AI is creating value, not just destroying it.
National Defence Tech & Palantir
ILTB, a16z, and All-In all converge on defence AI as a structural investment theme. Palantir’s forward-deployed model, AI-enabled precision warfare, and Silicon Valley’s obligation to national security are recurring threads.
AI Compute Bottlenecks
Dwarkesh/Dylan Patel provides the most rigorous technical analysis: logic, memory, and power are the three binding constraints. ASML identified as the 2030 chokepoint; NVIDIA’s TSMC allocation moat is underappreciated. Critical for semiconductor positioning.
Energy Transition Reality Check
Cembalest’s ‘Fighting Words’ energy paper and Odd Lots on fertiliser/oil shock force a reckoning with the pace of energy transition. AI power demand, SMR timeline hype, and the unfavourable EV industry economics create a nuanced picture for energy investors.
Episode-by-Episode Summaries
The Real Eisman Playbook
- Helima Croft (RBC/ex-CIA) unpacks Iran’s likely retaliation options and the energy market shock vector — upside oil risk is material if Strait of Hormuz remains functionally blocked.
- Eisman and Croft debate what ‘winning’ looks like for the US — unclear exit ramp given Iranian domestic politics post-strike.
- European exposure to the energy shock is Croft’s key underappreciated risk; LNG supply chain disruption could revive European energy stress seen in 2022.
- Eisman notes the war overlaps with private credit stress building beneath the surface — not a coincidence in terms of sentiment spillover.
Also Recent: Ep.49: Private Credit Cracks, The AI ‘Boogeyman’ & Why Crypto is for Boomers (9 Mar, 52 min) · Weekly Wrap: Iran War Volatility & Solar Crash (6 Mar, 22 min)
The Intrinsic Value Podcast
- Figma (FIG) has crashed 80% from IPO — the podcast dissects whether IPO-related SBC accounting distorted 2024–2025 financials and made the stock optically expensive.
- AI tailwind is central to the thesis: Figma is integrating AI into collaborative design workflows, which could expand its TAM well beyond its current base.
- Bull case hinges on Figma becoming a diversified enterprise design platform vs. remaining a ‘single hit’ product vulnerable to Adobe or new entrants.
- Shawn and Daniel assess whether FIG earns portfolio inclusion — valuation has compressed substantially, making risk/reward more interesting at current levels.
Also Recent: TIVP062: Duolingo (DUOL) — AI Threat or Generational Opportunity? (8 Mar) · TIVP061: Lyft (LYFT) — The Key to Winning the AV Wars? (1 Mar)
Odd Lots
- Tom Karako (CSIS) quantifies the ‘brutal math’ of missile attrition: the Iran war combined with Ukraine drawdown has severely depleted US and allied stockpiles.
- Supply chain constraints for manufacturing replacement missiles are significant — the industrial base was never sized for two concurrent high-intensity conflicts.
- Pentagon is accelerating production ramp, but lead times are multi-year; near-term the US is burning inventory faster than it can replenish.
- Key market implication: sustained upward pressure on US defence procurement budgets and a structural multi-year tailwind for missile/defence primes (RTX, LMT, NOC).
Also Recent: Iran War & China’s Teapot Oil Refineries (13 Mar) · Cyberwar in the Age of AI — Matt Suiche (12 Mar) · Oil at $200/barrel? Rory Johnston on Hormuz tail risk (10 Mar)
Nanalyze
- Scrutinises the AI data-centre power demand thesis as the primary driver of nuclear stock re-rating in 2026 — OKLO and CEG are focal tickers.
- Key concern raised: the timeline for small modular reactors (SMRs) remains uncertain; near-term power supply from nuclear is limited given construction lead times.
- Also covers AI in drug development — the broken economics of traditional pharma (10–15 years, $2bn per drug, 90% failure rate) as the disruption opportunity.
- LiDAR stocks also covered — remains focused on disruptive technology verticals with specific investability assessment.
Also Recent: AI Disruption of Drug Development — the Broken R&D Model (5 Mar) · LiDAR Stocks Update (7 Mar)
Acquired
- Deep dive into Formula 1’s business history — from near-bankruptcy to a $20bn+ franchise business dominated by Liberty Media.
- Traces how Bernie Ecclestone built the commercial rights empire, and how Liberty Media’s acquisition transformed F1 into a global media and events powerhouse.
- Investment angle: F1 is a live event / rights monetisation case study — comparable analysis relevant to other sports media franchises and the rights market.
- Partners this season include JP Morgan Payments and ServiceNow, reflecting F1’s premium demographic audience and continued sponsor premium.
Also Recent: The NFL — Special Episode (26 Jan 2026) · Coca-Cola — Fall 2025, Ep.3 (Nov 2025)
The Synopsis
- Speedwell dissects the recent software sector sell-off — arguing AI disruption fears have created indiscriminate de-rating, with quality compounders like Constellation Software caught in the crossfire.
- Introduces the ‘Point of Monetisation’ framework: dating apps fail because monetisation occurs at the moment when the core product (finding a match) succeeds, destroying the revenue relationship.
- Constellation Software (CSU) bull case remains intact — vertical market software has structural moats AI cannot easily commoditise due to customer lock-in and implementation complexity.
- Prior episode on Shift4 (Feb 19) is a high-quality deep-dive for anyone tracking payment processing vertical moats.
Also Recent: Shift4 (SFO) — Deep Dive from 85-Page Research Report (19 Feb 2026)
Invest Like the Best
- Shyam Sankar explains Palantir’s forward-deployed engineering model — AI agents embedded with clients to operationalise data rather than just analyse it; a key differentiator vs. generic AI platforms.
- Sankar argues the US must reindustrialise urgently: decades of offshoring have left critical supply chains (incl. defence) dangerously exposed to adversarial disruption.
- On AI value accrual: Sankar is sceptical that value flows primarily to foundational model providers — the bulk will accrue to entities that control deployment, data, and operational integration.
- Historical framing around ‘heretics’ (Rickover/nuclear Navy, Kelly Johnson/Skunk Works) is a framework for identifying unconventional talent — directly applicable to identifying exceptional management in investable companies.
Also Recent: John Arnold — China, Energy Markets & Fixing America’s Systems [EP.461] (4 Mar) · Dan Sundheim (D1 Capital) — The Art of Public & Private Market Investing [EP.460] (24 Feb)
All In Podcast
- Brad Gerstner joins to assess the Iran war economic fallout — escalation scenarios modelled against midterm election dynamics and Gulf state positioning; oil price is the key transmission mechanism.
- Notable data point: Anthropic and OpenAI are scaling revenue faster than any companies in history — Gerstner frames this as a structural shift, not hype.
- China angle discussed: the Iran war is creating secondary pressure on China’s oil supply (teapot refineries) while also accelerating the US-China technology competition narrative.
- Prior episode with SEC’s Paul Atkins & CFTC’s Michael Selig (Mar 11) is actionable: fixes to the IPO drought, crypto regulation framework, and the autonomous hedge fund question are all market structure events to monitor.
Also Recent: SEC’s Paul Atkins & CFTC’s Michael Selig — Crypto, IPO Reform, AI Trading Bots (11 Mar) · Harvard Prof. Graham Allison on Iran Strategy & Netanyahu’s Role (9 Mar)
Dwarkesh Podcast
- Dylan Patel identifies the three bottlenecks to AI scaling: logic (TSMC fab capacity / ASML EUV), memory (HBM supply crunch incoming), and power (US grid constraints the hardest near-term ceiling).
- ASML identified as potentially the #1 constraint for AI compute scaling by 2030 — implications for ASML valuation and the broader semiconductor equipment value chain.
- Google is being squeezed on chip supply despite massive capex — NVIDIA’s early TSMC allocation lock-in is a durable competitive advantage that is underappreciated by the market.
- Patel’s view on the China timeline: they could outscale the West in semiconductors if the US fails to maintain EUV export controls — geopolitical risk to AI infrastructure is material and underpriced.
- Space GPU manufacturing dismissed as ‘not this decade’ — useful corrective to periodic hype cycles around orbital compute.
Also Recent: Ada Palmer (Historian/Novelist) — Renaissance History & Long-Term Thinking (6 Mar) · The Department of War is Making a Huge Mistake (11 Mar)
Eye on the Market
- Cembalest frames the energy debate as a three-way standoff (renewables vs. fossil fuels vs. nuclear) — ‘Fighting Words’ is his most combative edition yet, tackling AI data-centre power demand head on.
- Key argument: the ‘primary energy fallacy’ overstates renewable progress by ignoring waste heat — the actual energy system transition is materially slower than headline figures suggest.
- Data centre AI buildout is creating real, durable demand for all power sources — Cembalest is notably sceptical on near-term SMR timelines but constructive on gas-bridging strategies.
- EV economics under scrutiny: highlights the persistently loss-making EV industry and the unfavourable economics of hybrid charging as warning signs for the sector.
- 2026 Outlook (‘Smothering Heights’) remains essential reading: AI-related names account for 65–75% of S&P 500 returns, profits, and capex since ChatGPT launch — extreme concentration risk.
Also Recent: 2026 Outlook: Smothering Heights — AI Concentration, Power Constraints, Taiwan Risk (Jan 2026)
The a16z Podcast
- Karp makes a pointed argument that Silicon Valley has a moral and commercial obligation to support national defence — Palantir’s forward-deployed AI in active theatres is the embodiment of this thesis.
- A companion piece to the ILTB/Shyam Sankar episode — together they present the most complete current picture of how Palantir operates and where AI value accrues in government contexts.
- Karp discusses America’s ability to cultivate unconventional talent as a strategic advantage — echoes Sankar’s ‘heretics’ framing and is relevant to talent-density analysis for tech-heavy mandates.
- a16z maintains active coverage across defence tech, AI infrastructure, and fintech — useful systematic signal for where institutional venture capital is flowing.
Also Recent: Ben Horowitz (a16z co-founder) — Backing America’s Future, AI Inequality & VC Strategy [ILTB EP.457] (3 Feb)
BG2 Pod
- Gerstner and Gurley assess ChatGPT’s evolution into a ‘super assistant’ — argues we are crossing a threshold where AI becomes genuinely productivity-transformative for knowledge workers at scale.
- Key market implication: if ChatGPT/OpenAI achieves super-assistant status at 500m+ users, the competitive moat widens significantly — reinforces the Anthropic/OpenAI revenue surge data from the All-In episode.
- Discussion touches on the China innovation challenge — China’s open-source AI models and rapid deployment are creating real US competitive risk beyond just chips.
- Prior episode with Jensen Huang (NVIDIA) remains one of the most important of the year for understanding AI infrastructure strategy and NVIDIA’s dominant position.
Also Recent: China’s Innovation & Global Tech Competition — Open Source AI and US Competitiveness (Feb 2026) · Jensen Huang (NVIDIA CEO) — AI Factories, Sovereign AI & the American Dream (Feb 2026)
Disclaimer: The summaries on this page reflect my personal notes and takeaways from publicly available podcasts. They are not transcriptions, and this site is not affiliated with or endorsed by any of the podcast creators listed. All intellectual property belongs to the respective creators.
If you enjoy these shows, please subscribe and support them directly.