Startup brief
AI Enterprise Agents and Foundation Models Drive Capital Inflows
Mega-rounds for Perplexity and Higgsfield alongside Harvey's open-weights model launch underscore rapid venture momentum in enterprise AI.
Startup
Nvidia is recycling chip profits directly into top software startups — cementing its role as both the primary hardware engine and lead investor across the AI stack.
BackgroundNvidia routinely invests balance sheet capital into prominent AI software companies that consume high volumes of graphics processing units. Perplexity previously completed a funding round at a $20 billion valuation in late 2025.
- Perplexity's annualized revenue expanded from under $250 million in early 2026 to more than $750 million by August, signaling rapid enterprise monetization.
- Growth was catalyzed by Perplexity Computer, an autonomous software agent designed for multi-step Web task execution across corporate workflows.
- Nvidia previously joined Perplexity's Series B round, tightening hardware-software integration as hardware giants double down on application-layer winners.
Startup
Enterprise adoption of synthetic video is scaling rapidly — turning media generation startups into multi-billion-dollar enterprise software companies.
BackgroundGenerative video tools have seen accelerated commercial adoption among advertising agencies and media companies seeking automated content creation. Higgsfield was launched by former Snap executive Alex Mashrabov to automate social video production.
- The $400 million Series B round drew backing from Goldman Sachs Alternatives, Intel Capital, Tribe Capital, Accel, and Menlo Ventures.
- Higgsfield reported growing annualized revenue from $20 million to $700 million within 12 months, driven by corporate marketing software deals.
- Over 30 million global users and marketing teams across 390 Fortune 500 companies actively generate social video content on the platform.
Startup
Top US enterprise AI vendors are adopting open-weight global foundation models — choosing raw reasoning benchmark performance over geographic software origin.
BackgroundHarvey provides specialized generative AI workflow tools to major corporate legal departments and international law firms. Open-weight foundation models permit software developers to train domain-specific reasoning capabilities on private infrastructure without proprietary API vendor lock-in.
- Harvey Tenet supports complex contract analysis, M&A due diligence, and structured document extraction for over 2,400 enterprise clients.
- Integrating Moonshot AI's open-weight Kimi K3 highlights growing adoption of international open-source architectures among US enterprise software startups.
- The startup is currently closing a funding round at a $15.5 billion valuation following rapid enterprise expansion across global law firms.