VENTURE CAPITAL

Venture capital is a form of private equity financing designed for early-stage, high-growth companies where the risk of failure is high but the potential return justifies the bet.

How VC Works

VC funds are structured as limited partnerships: limited partners (LPs) — endowments, pension funds, family offices — provide capital, while general partners (GPs) manage the fund and make investment decisions. GPs typically charge a 2% annual management fee and take 20% of profits (carried interest) above a hurdle rate.

The power law dominates VC returns. The top 10–20 investments in a fund typically account for more than 100% of total returns — winners more than compensate for all losses combined. This shapes everything about how VCs behave. They optimize for upside potential rather than failure avoidance, which is why VC-backed companies are pushed to grow aggressively. A company that returns 2x capital is nearly worthless to a fund; one that returns 100x saves it.

As a Technology Signal

VC capital flows are an imperfect but useful leading indicator of where technology is heading. These capital allocation decisions, when concentrated in a sector, accelerate development timelines, attract talent, and force competitive dynamics that would otherwise take years to emerge. The current concentration in AI infrastructure, foundation models, and AI-native application layers is the clearest such signal since cloud computing in the 2010s.

The Current Cycle

AI investment has attracted capital at a scale and speed that resembles internet-era dynamics. The structural question is how much value will accrue to infrastructure providers (compute, models), how much to application-layer companies, and how much will be competed away to end users as lower prices. That distribution question — fundamentally about platform economics — is the central analytical challenge of the current investment cycle — and VC allocation patterns are one of the cleaner ways to track how the market is answering it in real time.

The Liquidity Problem

VC as an asset class depends on exits — IPOs or acquisitions that return capital to LPs. The 2021–2024 period saw the IPO window effectively close for most technology companies, creating a backlog of late-stage companies that raised at peak valuations and cannot exit without crystallizing losses. Without exits, LPs don't receive distributions, fund performance stays unrealized, and information about which bets actually worked is delayed by years. The resolution of this backlog — through eventual IPOs, strategic acquisitions, or down-round restructurings — will determine the actual return profile of 2019–2022 vintage funds and recalibrate risk appetite for the next cycle.

The Information Advantage Decay

Early-stage VC returns historically derived from information advantages: VCs knew which founders were exceptional, which technologies were real, which markets were ready to tip before the broader market did. AI tooling is compressing those advantages. Market analysis is faster, founder track records are more transparent, and more capital is chasing fewer genuinely differentiated deals. Firms that sustain strong returns will be those with real founder relationships and deep sector knowledge — durable economic moats rather than pattern-matching heuristics that any competent analyst with good AI tools can now replicate.

Open Questions

• How much of AI application-layer value survives commoditization of the underlying models?
• Will the IPO backlog resolve through recovery or restructuring, and what does each path signal about the 2019–2022 vintage?
• As information advantages compress, does VC alpha increasingly collapse to a few relationship-driven firms, or does the field stratify in a new way?
• How do sovereign wealth funds and corporate strategics entering early-stage investing change GP incentive structures and return expectations?

Part of the knowledge graph at The Best Blog Ever — reference definitions for ideas that matter.