Nikhil KamathInside Silicon Valley’s VC Playbook | WTF is Venture Capital? - 2025 Edition | Ep. 24
CHAPTERS
Setting the agenda: where are the tailwinds for the next decade?
Nikhil frames the conversation as a practical exercise for investors and young professionals: which sectors will have the strongest tailwinds over the next 10 years. The guests agree to go beyond buzzwords and talk through concrete sector calls and second-order effects.
- •Goal: identify durable tailwinds for careers and capital allocation over a decade
- •Move from venture vs. PE labels to broader ‘where would you put $100?’ thinking
- •Plan: introductions, then sectors to avoid, hottest sectors, and future predictions
- •Emphasis on rate of change accelerating and its implications
Deedy Das’ path: from India to Big Tech to Menlo, plus the Anthropic backstory
Deedy explains his journey from Cornell to Meta/Google and then Glean, where he ran AI product lines, before joining Menlo Ventures. He also gives a concise take on Anthropic’s origin and positioning: enterprise-focused, text-first, and a reasoning/coding route to AGI.
- •Deedy’s background: Cornell CS, Meta/Google, founding team at Glean, then Menlo Ventures
- •Glean evolved from enterprise search to enterprise assistant/knowledge finding
- •Menlo: long-standing fund investing across AI, SaaS, infra; notable stake in Anthropic
- •Anthropic story: ex-OpenAI founders, emphasis on reasoning/coding, text-to-text focus, enterprise GTM
Nikunj Kothari’s operator-to-investor story: LinkedIn, startups, Opendoor, Meter, and FPV Ventures
Nikunj outlines a decade-plus as an operator across multiple startups, with formative lessons at Opendoor and later at Meter. He explains why he moved into investing: to help multiple founders at once, and why he prefers concentrated seed/Series A generalist investing.
- •Career arc: LinkedIn → multiple startups → Opendoor → Meter → Khosla Ventures stint → FPV Ventures
- •Opendoor explained: reducing 6% transaction friction via instant cash offers, inventory holding, refurb/resale
- •Meter model: networking hardware as a subscription (no upfront capex), end-to-end install/maintenance
- •FPV thesis: generalist, founder-first; fund size and stage focus
Niko Bonatsos’ investing philosophy: bet founders, avoid the hottest overcrowded spaces
Niko describes his focus on technical, young, first-time founders and argues that early-stage investors should avoid areas with heavy copycat activity. He highlights how inbox patterns reveal overheated themes and why warm intros and unconventional resumes matter.
- •Background: Greece/UK → Silicon Valley since 2009; long tenure at General Catalyst
- •Thesis: founders matter more than sector; pre-seed/seed is about trajectory, not static ideas
- •Early-stage warning: avoid spaces with 10+ near-identical startups and heavy hype
- •Gen Z founders may be hard to evaluate via traditional signals (e.g., not on LinkedIn)
What to avoid: sector “headwinds,” CapEx bias, and the risk of being too late
The group struggles to name universal ‘avoid’ sectors because every industry can be interesting with the right approach. They converge on two practical heuristics: be wary of venture-style return mismatch (often high CapEx) and avoid ultra-hot crowded categories where differentiation is weak.
- •Deedy: high CapEx can be great businesses but often mismatched for venture returns
- •Nikunj: hard to separate from tech lens; many ‘uninvestable’ areas become investable with AI-driven efficiency
- •Niko: for early-stage, the real ‘avoid’ is overcrowded, pattern-matching startup waves
- •Heuristic: break the pattern—ideas that “sound the same” are usually too late
Today’s hot sectors: codegen, RL environments, voice apps, and healthcare scribes
They list what’s currently overheated—areas flooded with similar startups—and discuss why some hot markets can still support multiple winners. Healthcare AI scribes stand out as a rare example where many new entrants are achieving significant revenue growth simultaneously.
- •Overheated themes: AI codegen tools, RL environments/RL-as-a-service, voice/receptionist agents, app builders
- •YC pattern-matching joke: “YC = Y Cursor” (Cursor-for-everything)
- •Healthcare AI scribes: multiple companies with strong ARR and fast growth despite crowding
- •Lesson: ‘hot’ isn’t always bad, but entry timing and differentiation become critical
Emerging AI trends: data limits, reinforcement learning, evals, and long-horizon reasoning
The guests outline underappreciated AI shifts: public data exhaustion, reinforcement learning as a path to improvement, and the growing importance of evaluation frameworks and expert-in-the-loop refinement. Nikunj adds that long-running, autonomous reasoning (hours-long) is a step change that unlocks viable agents.
- •Deedy: public data is running out; RL-based improvement becomes more central
- •RL reward framing (RLVR and LLM-as-judge) as a practical lever for capability gains
- •Niko: ‘evals’ and exception-handling create large markets for expert-guided model refinement
- •Nikunj: models running for ~10 hours reasoning/planning indicates new agent viability and research potential
Demographics meets AI: declining birth rates, phones, and changing human behavior
The conversation pivots to demographic decline and its economic implications, then ties it to technology’s impact on relationships, sex, and meaning. They argue the smartphone era correlates with declining fertility and speculate AI may intensify digital substitution for real-world intimacy and community.
- •Deedy: birth-rate decline threatens consumption and growth; slow-burn but massive impact
- •Nikunj: iPhone era correlates with decline; digital media satisfies emotional needs, reducing real-world pairing
- •Niko: post-COVID social skill erosion; dopamine from digital experiences affects fertility and vice industries
- •Second-order effects framing: autonomy, AVs, and tech cascades reshape entire sub-industries
Abundance vs. capitalism: productivity gains, inequality, and the search for meaning
They debate whether AI leads to an abundant future (more leisure, art, relationships) or amplifies inequality through capital concentration. The discussion touches on Piketty-style inequality dynamics, UBI uncertainty, and how status and meaning might dominate even in ‘abundance.’
- •Nikunj’s optimistic view: efficiency historically creates new jobs and more leisure time
- •Counterpoint: growth often increases inequality; abundance may be unevenly distributed
- •Meaning and status: Maslow’s hierarchy, social comparison, and perceived scarcity persist
- •UBI is discussed as uncertain; structural incentives remain misaligned
Raising kids in a public world: surveillance, identity, and digital detox as a business
They argue we should assume life is effectively public: cameras everywhere, self-driving sensors, and persistent digital records starting in childhood. This fuels a growing market for detox tools, retreats, and ‘dumb-smart’ devices that keep core utilities while limiting addictive feeds.
- •“Live in public” thesis: sensors, phones, building cameras, and data retention
- •Parents create children’s digital identities early (Instagram, content footprints)
- •Detox trend: silent retreats, Vipassana analogs, apps like Opal/Clearspace, and minimalist phones
- •Speculation: future hardware may reduce screen addiction via voice/gesture/AI-first interfaces
The future of dating apps: digital twins, AI matchmaking, and offline-first experiences
They predict dating shifts away from swipe mechanics toward AI-curated intros and potentially AI-assisted early conversations. At the same time, they expect premium offline social formats (classes, curated dinners) to grow as scarce, high-status experiences.
- •Swipe-based dating seen as dopamine/entertainment rather than relationship formation
- •Digital twins and rich personal data could enable higher-quality matching and conversation scaffolding
- •Algorithmic inequality on apps (top % getting most dates) suggests demand for new formats
- •Offline communities/events (e.g., curated dinners, activities) become part of ‘dating infrastructure’
Predictions and a scoring game: education, private secondary markets, health, and senior living
Nikhil turns the discussion into a structured scoring exercise to rank decade-long tailwinds. The panel strongly backs outcome-based education (especially in India), expects healthcare spend to rise, and sees senior living/community models as a major opportunity as societies age.
- •Outcome-focused education in rising-income, smaller-family cultures gets high scores
- •Secondary markets for private companies debated: bullish vs. cyclical/regulatory constraints
- •Healthcare spend expected to rise; business attractiveness depends on bending cost curves
- •Senior living framed as community/longevity enhancer; cultural stigma likely to fade over time
Brand building in the post-search era: content vs. product, and the return of ‘David vs. Goliath’
They argue that as search and performance marketing shift, content becomes a dominant discovery channel—especially for consumer brands—though product quality still determines retention. They also discuss how consumers increasingly value individuality, artisanal experiences, and smaller brands, while incumbents lose youth resonance.
- •Thesis: declining reliance on search ads increases the value of attention and storytelling
- •Counterbalance: too much ‘attention on attention’ risks empty-calorie products with weak retention
- •David vs. Goliath dynamic: big companies lose focus; small brands win via craft and narrative
- •Trust and taste cycles: each generation seeks distinct brands and offline experiences
2035 sector calls: EVs/mobility, beauty/luxury, live events, and climate/energy
They forecast that EV adoption continues but varies by infrastructure, politics, and cost; China’s manufacturing lead is a major factor. Beauty/luxury earns unanimous enthusiasm, live events split the room, and climate is reframed as an energy problem—especially as AI becomes a major energy consumer.
- •EVs: adoption driven by total cost of ownership + charging convenience; likely majority of new cars by 2035
- •Beauty/luxury: strong tailwinds (male grooming, anti-aging, Ozempic framed as ‘beauty product’)
- •Live events: scarcity premium and superstar economics vs. ‘screen time wins’ skepticism
- •Climate tech: definition debated; energy (nuclear, solar, geothermal, fusion) seen as core tailwind due to AI demand
Infrastructure and vices: data centers, prediction markets, and the business of speculation
They discuss data centers as a potentially attractive but operator-dependent infrastructure bet, with energy and regulatory data-localization as key drivers. They also endorse prediction markets/gambling as durable, high-demand ‘vice’ businesses—assuming regulation is handled.
- •Data centers: demand rises with token/compute growth, but margins depend on energy, real estate, and hardware cycles
- •Localization laws (e.g., data sovereignty) create country-specific demand pockets
- •Compute as a geopolitical asset; nations want strategic control over it
- •Prediction markets (Kalshi/Polymarket): framed as entertainment + speculation; strong demand tailwinds if legal
Closing synthesis: unfair advantage vs. macro tailwinds, plus India’s role in the AI race
Nikhil summarizes the top-scoring categories (beauty/luxury, speculation, education, aging-related care, content). Deedy pushes back: macro tailwinds don’t replace ‘unfair advantage,’ especially distribution and execution. They end with a candid discussion on India’s position—strong on applications and demographics, weaker on foundational AI without long-term investment in research, talent, compute, and incentives.
- •Takeaway rankings: beauty/luxury and vices score highest; education and aging themes follow
- •Deedy’s critique: sector tailwinds ≠ founder right-to-win; distribution advantage often decisive
- •India in AI: limited local advantage in foundational models; requires deep investment in research, talent retention, GPU/compute ecosystem
- •Optimism on India’s application layer and demographic advantage; challenge is building/retaining critical mass and culture of risk-taking
