Kevin Ryan’s AlleyCorp raises new $335 million fund, all in on early-stage bets

AlleyCorp, Kevin Ryan’s venture firm, raised a $335 million second fund (after a 2024 $250 million fund) and continues making early bets, including a recent investment at a $25 million valuation for a company with ~$0.5 million in revenue. Ryan frames current AI-scale mega-deals (e.g., Anthropic/OpenAI and other mega-platform investments) as a distinct category from traditional VC, arguing that removing those outliers could make VC appear down even if underlying venture activity persists. Overall, the piece is constructive on AlleyCorp’s strategy and readiness to adapt in an AI-driven cycle.

Analysis

This reads less like a broad endorsement of venture and more like evidence of capital concentration. When mega-rounds soak up most of the funding pool, the median seed-to-Series B company faces a worse exit matrix: fewer IPO comparables, less acquisition appetite, and a longer bridge to liquidity. That is structurally negative for the smaller VC ecosystem and for any public proxy tied to venture formation, even if headline deployment stays strong.

The cleaner public-market implication is that AI spend keeps narrowing toward a handful of infrastructure suppliers. That favors semis and adjacent picks-and-shovels over application-layer software, because the marginal dollar is flowing into compute, networking, power, and memory rather than into yet another enterprise SaaS layer. NVDA is still the clearest expression of that dynamic; the setup is more about sustained order visibility over the next 1-3 quarters than about a single-day sympathy move.

Contrarian view: consensus is reading AI as a tide that lifts all startup boats, but the article points the other way — it’s a barbell market where a few names get effectively public-company scale while everyone else competes for scraps. The risk is that the narrative looks healthy on aggregate dollar figures while the underlying startup survival rate deteriorates. That would hit unprofitable software and VC-linked assets first, and it would be exposed quickly if hyperscaler capex or NVDA guidance softens over the next 1-2 earnings cycles.

Key Decisions for Investors

• Maintain/accumulate NVDA on pullbacks over the next 4-8 weeks; the trade is not 'venture is hot' but 'AI capex remains concentrated.' Use a 5-7% stop below entry because the setup is valuation-sensitive if capex momentum slows.
• Pair long SMH vs short ARKK or IGV for a 1-3 month horizon. The spread should benefit if capital keeps migrating to infrastructure and away from venture-backed application names; falsifier is a broad re-acceleration in software revenue growth or a sharp drop in rates.