MoneySeptember 10, 2026via SiliconAngle

Chipmaker Positron nabs $875M to speed up inference with consumer-grade memory

Why it matters

A well-funded hardware startup is attacking a core constraint in AI inference: the cost and scarcity of high-bandwidth memory. If consumer-grade memory can sustain inference workloads at scale, it reshapes who can afford to run models and where.

Key signals

  • Positron raised $875M in Series B funding
  • Led by NEA, Atreides Management, Valor Equity Partners, Andra Capital, SemiAnalysis Capital
  • Jim Clark (SGI, Netscape co-founder) participated
  • Company builds inference appliances without high-bandwidth memory
  • Focus on inference optimization using consumer-grade memory
  • Founded to address HBM scarcity and cost as deployment bottleneck

The hook

$875M. Positron just raised a massive round to make AI inference run on cheap memory — changing the economics of deployment.

Positron AI Inc., a developer of inference appliances that don’t include high-bandwidth memory, today announced that it has raised $875 million in funding. NEA, Atreides Management, Valor Equity Partners, Andra Capital, SemiAnalysis Capital and Silicon Graphics Inc. and Netscape Inc. co-founder Jim

The week's key stories, every Friday.

For practitioners and enthusiasts — free, in your inbox.

Free forever. No spam.