Menlo Ventures Raises $3B for AI-Focused Funds in Its Biggest Fundraise Ever

Menlo Ventures Raises $3B and Rearranges Its AI Strategy

Menlo Ventures announced in June that it had raised $3 billion across two new funds, the largest capital raise in the 50-year history of the Silicon Valley firm. The money will be split between Menlo Ventures XVII, which will focus on seed and Series A companies, and Menlo Inflection IV, which will provide growth capital to startups at Series B and later.

The new funds are aimed squarely at the AI market, spanning foundational models and infrastructure through enterprise, healthcare and consumer applications. The structure gives Menlo the flexibility to back companies from their earliest days and stay with them through later rounds that can require hundreds of millions of dollars.

The firm's most prominent AI investment is Anthropic, the model developer backed by Menlo since its Series C in 2023. Menlo later led Anthropic's Series D, which partner Matt Murphy described as the largest investment the firm had ever made at the time. Other AI bets in Menlo's portfolio include app-building platform Lovable, music startup Suno, model marketplace OpenRouter, AI infrastructure companies Fireworks AI and Modal, robotics developer Skild AI and research lab Goodfire.

In an interview with Crunchbase News, Murphy said the scale of the raise reflects a structural shift in how AI companies consume capital — they stay private longer, grow faster once they break out, and need money to sustain hypergrowth. He described the firm's new approach as a 'barbell' strategy: patient but smaller early-stage bets, followed by aggressive concentration in proven winners.

Why AI Is Pushing Menlo Toward Bigger, More Concentrated Bets

Why AI Changed Menlo's Investment Playbook

The size of the raise is not just a bigger fund — it is a response to how AI businesses differ from earlier software companies. Murphy's explanation is that AI winners separate from the pack quickly and need capital to sustain that growth, which pushes venture firms toward concentration. The template is Anthropic: Menlo first invested in the C round to get close to the team, then led the D round in what was then the firm's largest-ever investment. That pattern — small entry, conviction-building, then a large follow-on — has now been formalized, with the inflection fund designed to write the later-stage checks.

That logic already shows up in practice. Murphy confirmed that Menlo has recently invested $100 million in companies including Lovable and Suno, and said the firm will keep the bar for such concentrated bets high.

From Model Selection to Model Optimisation

Murphy frames the market as moving from 'Phase 1' to 'Phase 2.' In the first phase, developers simply chose a model and started building. In the second, companies are at scale and shifting their focus to optimising spending, infrastructure choices and multi-model workflows. He argues the market will be multi-model rather than winner-take-all, creating tailwinds for companies like OpenRouter, Fireworks AI, Modal and Gimlet that sit alongside model providers such as Claude and Claude Code.

Menlo is also backing more specialised models, including Chai Discovery in life sciences and Skild in robotics, suggesting the firm sees value beyond general-purpose frontier models.

The Bottleneck Is Getting AI Code Into Production

Murphy identified the single biggest constraint across the AI ecosystem: taking newly written code and shipping it to production faster, safely and securely. That points demand toward software delivery platforms like Harness, security tools like Semgrep and code review and testing systems like Greptile. The rise of open-weight and custom models adds another set of bottlenecks around compute, training and sandbox environments, where Modal and Fireworks are aggregating capacity from providers including Nebius and CoreWeave.

These are interpretations of market demand, not guarantees. The companies Murphy names are Menlo portfolio companies, and their tailwinds are described from the firm's vantage point.

The Risks Inside a Land Grab

Murphy is candid about the market's exuberance. He says many AI categories are overfunded and there is a huge amount of speculation, with most companies optimising for market share rather than gross margin. He estimates there are roughly 60 model companies, many claiming differentiated techniques, and says Menlo has invested in more than five of what it considers the best — expecting to lean into one or two as they ramp. For a firm raising $3 billion, the risk is deploying that capital at peak valuations in categories that consolidate faster than expected.

It is worth separating what is verified from what is opinion here. The fund size, the two-fund structure, the named investments and Murphy's biographical background are facts. The phase-shift narrative, the barbell strategy and the list of bottlenecks are Murphy's analysis of the market, reported as such.

What Founders and Investors Should Take From Menlo's New War Chest

For founders, investors and operators watching Menlo's move, the concrete implications are:

• Seed and Series A AI founders should treat early rounds as relationship-building with multi-stage firms. Menlo's model — a small first check, a window to observe execution, then a much larger follow-on — means the practical path to a nine-figure investment starts long before the later round is raised.
• Infrastructure and developer-tool startups have the clearest wind at their backs. Menlo's portfolio points to demand in software delivery (Harness), code security (Semgrep), code review and testing (Greptile) and compute orchestration (Modal, Fireworks).
• Late-stage AI companies should expect investor focus to shift from growth at any cost toward margin improvement. Murphy himself notes most companies are optimising for market share now, which implies the next phase of scrutiny will be unit economics.
• Founders in crowded AI categories should be cautious about relying on differentiation claims alone. Murphy estimates roughly 60 model companies exist and expects to concentrate on only one or two of the five-plus Menlo has backed — a signal that most AI labs will not get follow-on capital from the firm.