How Menlo Ventures closed a $3 billion fund after its bet on Anthropic paid off

Menlo Ventures, one of Silicon Valley's veteran venture capital firms, has announced the closing of a new $3 billion fund driven by the return on its early bet on Anthropic. According to TechCrunch on Tuesday, the fund size is roughly twice that of Menlo's previous fund and the largest single fund close since the firm's founding in 1976.
Menlo's initial investment in Anthropic was made in early 2023, when the company was still at a pre-Series C stage. The firm put $50 million into that round, securing a significant stake in Anthropic's then balance-sheet-based valuation. Recent private-market valuation estimates for Anthropic have hovered around $100 billion; that implies a roughly 100x value appreciation on Menlo's initial investment.
Matt Murphy, head of Menlo Ventures, told TechCrunch "we saw that being a day-one partner with an AI frontier company like Anthropic meant we needed to re-architect our fund strategy." Murphy added that the new fund will amplify the firm's active AI thesis.
The fund is backed by institutional investors including US university endowments, insurance companies, sovereign wealth funds, and family offices. TechCrunch reported that the investor list is led by the Princeton University endowment, Singapore's GIC, and the Dutch PGGM pension fund.
The size of the new fund is particularly notable when placed alongside other large Silicon Valley funds. Andreessen Horowitz closed a $7 billion flagship in 2022; Founders Fund raised $7 billion in 2024. Three billion dollars is striking for a Series A/B-focused firm like Menlo, which is why it attracted industry attention.
The fund's strategy was pitched to investors on three pillars: (1) AI infrastructure companies (from training chips to data center software), (2) AI agent and application-layer firms, and (3) the looming post-CUDA race in the semiconductor sector. The Anthropic partnership gives Menlo a direct strategic link in the third category.
Pitchbook venture analyst Kyle Stanford told TechCrunch that "the value appreciation in AI through 2024 to 2025 produced extremely concentrated returns for a limited number of early investors; the new fund sizes reflect that logical consequence." According to Stanford, the speed of fund-size growth signals a "meta-regime change" in the industry.
Menlo's investment decision is also at the center of a contested industry debate. Some limited partners (LPs) have raised questions about the sustainability of private-market AI valuations since 2024. Coatue Management, followed by Tiger Global, has signaled it is throttling AI exposures. Menlo says its fund strategy is structured to account for those warnings.
The fund's first investment will be in Aprhinox AI, which is working on practical interpretability of machine-learning models. The company is building smaller models capable of explaining AI agent decisions to humans, a capability that is becoming critical for both regulated sectors (finance, health care) and internal audit.
Menlo's new $3 billion fund is a sign that the firm intends to remain a dynamic player in the AI market in Silicon Valley. Over the next 18 to 24 months in which the fund is expected to be deployed, it is expected to remain a major driver of large investment rounds in the AI sector.
Read next

Hollywood and OpenAI: how artificial intelligence is reshaping the cinema economy
Italian director Luca Guadagnino's new artificial-intelligence-themed film "Artificial" reflects a growing alignment between Hollywood's major studios and AI companies such as OpenAI. A report from The Verge sets out what the alignment means for the film industry.

Why vulnerability reports are no longer 'special': the false-alarm problem in the bug-bounty ecosystem
Former Go security lead Filippo Valsorda argues in a blog post that genuine security findings are becoming harder to distinguish in the flood of reports arriving at bug-bounty programs. The post, discussed at the top of Hacker News, exposes how open-source projects are affected.

Global cargo delivery from orbit: how SpaceX's Starfall is designed to work
SpaceX has revealed the first details of Starfall, a suborbital cargo-drop variant of Starship. According to Ars Technica's report on the project, the goal is to deliver material to any point on Earth in less than an hour.

What is post-quantum cryptography, and why the US has accelerated its migration deadline
The White House has significantly shortened the deadline by which federal systems must move off quantum-vulnerable cryptography. The executive order, described by Ars Technica, explains what threat post-quantum cryptography (PQC) protects against and why the migration is technically difficult.

What is YOLO26? A practical introduction to the real-time object-detection model
YOLO26, the latest version in the YOLO family, has set a new reference point for the real-time computer-vision community. A comprehensive introduction from Roboflow covers the new architecture, ease of training and real-world applications in depth.
