Greylock launches $1.5bn Fund XVIII for AI startups

About Greylock

  • Founded: 1965

  • Headquarters: Menlo Park, California

  • AUM: $7bn as of 2024

  • Strategy: Early-stage venture capital, pre-seed through Series A, North America

  • Thesis: Backs technical founders building transformative companies at the intersection of infrastructure and application layers, with particular focus on AI, developer tools, and consumer platforms

  • Track record: Prior funds include Fund XVII ($1bn, 2023), Fund XVI ($1bn, 2020); portfolio includes LinkedIn, Airbnb, Discord, Workday, and Coinbase

  • Leadership: Sarah Guo, Reid Hoffman, David Sze

Greylock has launched Fund XVIII at $1.5bn, its largest early-stage vehicle to date and a 50% increase over its predecessor fund closed just two years prior. The fund will target pre-seed, seed, and Series A companies, with emphasis on AI infrastructure and applications. The firm has invested in AI-native companies including Adept, Cohere, and Credo AI across recent vintages.

The 50% step-up in fund size mirrors the broader market recalibration around AI as a platform shift rather than a feature increment. Sequoia's 2024 seed fund closed at $195m, up from $85m in 2022; a16z's American Dynamism fund expanded to $600m in 2023 from $400m a year earlier. LP appetite for early-stage AI exposure is concentrated among firms with defensible technical diligence infrastructure and portfolio company support — attributes that scale unevenly as fund sizes rise. Greylock's entry into the $1bn-plus early-stage cohort raises deployment pressure: at 25-30 investments per fund, check sizes push into the $40m-$60m range for ownership maintenance, compressing the stage definition of early.

The firm's AUM has grown from roughly $3.5bn in 2018 to $7bn today, implying three funds raised in six years. That pace, combined with Fund XVIII's size, suggests Greylock is front-loading capital deployment into a vintage window it views as particularly productive. The counterpoint: if AI infrastructure commoditizes faster than application-layer defensibility materializes, funds deploying at 2024-2025 entry prices face margin compression on exits. The test is whether Greylock's technical founder network and early positioning in model infrastructure translate to durable platform wins — or whether the fund's scale reflects temporary supply-demand imbalances in a sector still searching for repeatable business models.

Source: AltAssets