The Twenty Minute VCThe Impact of H1B Visas on Startups in the US & NVIDIA Invests $100BN Into OpenAI
CHAPTERS
Frothy AI market vibes and what the show will tackle
The episode opens with banter that quickly sets the tone: markets feel exuberant, diligence is slipping, and the panel wants to pressure-test what’s real versus hype. They tee up NVIDIA/OpenAI as the anchor story and broaden into venture/public market behavior.
NVIDIA’s reported $100B OpenAI investment: money loop or scaling bet?
The group unpacks the strategic logic of NVIDIA financing OpenAI: it enables OpenAI to keep pushing scaling laws without a capital “timeout.” They debate whether this resembles vendor financing dynamics from past cycles and what has to be true for returns to justify the spend.
Do scaling laws still work—and why Sam Altman’s messaging matters
Harry challenges the assumption that scaling continues, citing a perceived shift toward efficiency. Jason argues that Altman’s public claims (e.g., needing 1,000x compute) tend to be deliberate and predictive, meaning the industry may be in an extended “double down until it stops” phase.
Anthropic vs OpenAI: chips, capital, and monopoly risk management
They assess whether Anthropic is strategically disadvantaged by OpenAI’s tighter NVIDIA relationship. Beyond capital, the real edge may be preferential GPU access and the ability to finance massive compute buildouts—while OpenAI also tries to avoid being seen as an outright monopolist.
NVIDIA’s concentration: a $4T company with a handful of buyers
Rory highlights the fragility of NVIDIA’s revenue concentration—six customers representing the bulk of revenue—contrasting it with Apple/Microsoft’s broader bases. The bull case is that those few buyers are committed to spending aggressively to win the AI platform prize.
Echoes of 1999, but with unprecedented capital and guarantees
The panel compares today’s mood to 1999’s “unlimited possibility,” noting that this time incumbents have immense free cash flow and can backstop the ecosystem. They discuss NVIDIA’s cash generation and buybacks, plus the perennial mistake of corporate buybacks at peaks.
Personal positioning and portfolio risk: cash, valuation, and sleep-at-night decisions
They pivot to investing behavior: Jason jokes he’s at 0% cash, while Rory emphasizes asset allocation over short-term market timing. The group discusses how valuation predicts 10-year returns better than 1-year returns, and how risk tolerance changes with wealth and fund dynamics.
Is “triple, triple, double, double” dead—and why funding feels harder
They debate the meme that classic SaaS growth profiles no longer get funded. Rory argues the core early-stage market is similar; what changed is an additional “ultra-late-stage private-public” layer that is highly concentrated. Jason notes the ‘S-tier’ is obvious, but the next tier down is unpredictable and meeting access is harder.
Exit math reality check: DPI, fund-returners, and why ‘boring’ $5–8B exits still matter
The panel reframes what “counts” in 2025: big-name AI winners dominate mindshare, but many strong outcomes still produce great multiples. Rory argues a 7–10x is a 7–10x regardless of narrative stature, while Jason worries about being an “asterisk” in modern DPI tables where mega-winners dwarf everything else.
Navan files to go public: going first, comps, and IPO window strategy
Navan’s S-1 prompts a discussion of why a company might IPO before profitability: novelty, market attention, and avoiding being the “last comp out.” They explore whether Navan is truly comparable to Brex/Ramp or more distinct due to travel booking revenue, and why being early can be strategically valuable.
How liquidity really works post-IPO: lockups, secondaries, and slow exits
Rory explains the mechanics behind headline “paper wealth”: investors face lockups, trading windows, and often need structured secondaries to exit meaningfully. The realistic timeline for substantial liquidity can extend 12–24 months, and board roles create both constraints and informational advantages.
H-1B fee changes: practical impact on startups and talent strategies
They discuss the announced $100k fee/payment for new H‑1B visas and its likely chilling effect on startups hiring skilled immigrants. While acknowledging immigration’s net-positive impact on tech and GDP, Jason argues companies will route around constraints via O‑1 visas or big tech simply paying up—making near-term effects modest but directionally harmful.
Notion at $500M ARR and the hangover from 2021 valuations
Notion’s re-acceleration sparks a broader point: at hundreds of millions in ARR, even 30–40% growth is impressive and IPO-relevant. They argue legacy productivity companies must lean into AI while being valued on fundamentals—then call for finally ‘moving on’ from 2021 peak valuation anchors.
‘Founder-friendly’ is broken: no-diligence deals, term sheet pulls, and what real support looks like
The panel critiques modern “founder-friendly” posturing amid competitive AI rounds, where deals are rushed and diligence may happen after term sheets—sometimes leading to pulls. They argue true founder support only shows in hard moments: bridging crises, doing real recruiting work, and being candid rather than performatively positive.
Kalshi quick-fire: TikTok timing, Meta smart glasses, and Atlassian M&A
They close with rapid predictions: Rory jokes TikTok may drag on; Jason bets a near-term resolution tied to broader deal-making. Jason is skeptical Meta smart glasses will become mainstream, while Rory thinks Atlassian’s acquisitions help defensively but won’t make it an AI category leader.
