Fireworks AI

Index Ventures Raises $2B Across Three Funds

Index Ventures announced it has raised $2 billion across three funds: a $400 million seed-focused fund, a $900 million venture fund, and a $700 million addition to an existing growth vehicle. The $700 million tops up a $1.5 billion growth fund raised in 2024, bringing Index’s total available capital to $3.5 billion. The raise follows strong portfolio performance, including Index’s early investment in Wiz, which was sold to Alphabet for $32 billion (Index held roughly 12% and that stake was reported to be worth about $3.8 billion). Index has also been an early backer of Figma and has AI-related investments including Physical Intelligence, Fireworks AI and Anthropic.

China Open-Weight LLM Sparks US AI Policy Clash

Moonshot AI’s new open-weight LLM, Kimi K3, released July 16 and priced far below U.S. frontier models, has triggered a high-stakes policy and industry debate. U.S. vendors (notably OpenAI and Anthropic) and some government actors are pushing regulatory pressure and investigations over alleged model distillation, while startups, investors, and major tech leaders voice support for open weights and cheaper access. The rush around Kimi K3 has disrupted alliances, prompted the formation of a “Little Tech Alliance” of startups opposing bans, and accelerated fundraising and valuation talk for Moonshot. Separately, PitchBook data shows corporate venture arms dominating AI VC dollars this year, with Nvidia leading in invested capital.

Moonshot's Kimi K3: Open Model and Running Costs

Moonshot AI released Kimi K3 on July 16, 2026, and plans to publish the model weights for free download on July 27. K3 is a 2.8-trillion-parameter open-weight model with a 1,048,576-token context window and native image/video understanding; Moonshot also offers interactive access on its site. Early independent benchmarks report leading results in frontend coding, 3D design, agentic task success across 8,000+ sessions, and high CVE detection, often at roughly one-third the per-task cost of comparable closed models via cache and token-efficiency economics. Moonshot says K3 is optimized for long, context-heavy workloads, notes that running it at scale requires substantial data-center hardware (the deployment guide recommends at least 64 high-end AI accelerators), and concedes K3 still trails the strongest proprietary models in some areas.