Prioritizing Access and Flexibility in AI Venture Capital

In this conversation, Matt Murphy of Menlo Ventures maps how venture capital is changing in the age of AI. He explains that the traditional focus on high ownership percentages is a relic of a time when outcomes were smaller. By prioritizing access to outlier companies over rigid entry terms, Murphy shows a shift toward a full-stack investment model where flexibility leads to better returns. The implication is that the most durable competitive advantages for both investors and startups are no longer found in defensive moats, but in the ability to navigate rapid, compounding technological shifts. Readers should note that the Series A stage has become a compression zone where traditional signals fail, making early-stage agility and late-stage conviction the only viable paths for long-term outperformance.

The Hidden Cost of Rigid Investment Mandates

The most important insight from Murphy’s experience with Anthropic is that sticking to rigid investment rules, such as minimum ownership targets or specific stage mandates, is a form of self-exclusion. While traditional venture doctrine suggests a firm must own 15 to 20 percent of a company to justify the risk, Murphy argues that in an outlier-driven market, ownership percentage is secondary to being present in the winning company.

"I think other firms can be and not to throw any shade out anybody because I'm such great respect but you kind of getting these situations where we have to own 15 or 20% ownership or we don't do this and don't do that. And I think the new Menlo that I'm part of has shown extreme flexibility to just do what makes sense."

-- Matt Murphy

By bypassing internal resistance to invest in a pre-revenue company at a high valuation, Menlo secured a position in a generational winner. The system-level consequence is that firms that prioritize perfect entry points over market participation eventually find themselves priced out of the entire category.

Why the Series A Middle Ground is Failing

Murphy identifies the Series A stage as the most difficult environment for investors today. The traditional signal, moving from a few POCs to 1 to 3 million dollars in ARR, no longer provides the clarity it once did. Because growth has compressed, the time between seed and Series A has shrunk from years to weeks, causing valuations to balloon before the underlying business has truly matured.

This creates a barbell dynamic. Menlo has responded by shifting resources to both ends of the spectrum: writing smaller seed checks to build proprietary deal flow and taking massive, concentrated positions in companies that have already proven themselves as outliers. The downstream effect of this shift is that boutique seed funds are becoming increasingly vulnerable; they are often too large to be friendly but too small to lead follow-on rounds, leaving them stranded in a market that rewards either extreme early-stage conviction or late-stage scale.

The Evolution of the Full-Stack Firm

The conversation reveals that the era of swim lanes, where firms strictly defined themselves as seed, growth, or late-stage, is effectively over. Firms are being forced to go full-stack to maintain relevance. This is a systemic response to the fact that great companies are compounding faster than ever.

"If you get even a wedge into a company, you're 10x more likely to be able to participate significantly in the next round or lead."

-- Matt Murphy

This strategy creates a feedback loop: by taking small tracker positions early, a firm secures the relationships and information necessary to lead subsequent, more lucrative rounds. The competitive advantage here is delayed; it requires the patience to hold small positions for years while waiting for the right moment to pile in when the company trajectory becomes undeniable.

Systems Thinking: The Shift from Models to Tooling

While the initial wave of AI investment focused on foundation models, the system is now routing around the initial bottleneck of model performance. As companies optimize for cost and latency, they are increasingly adopting a multi-model approach. Murphy notes that the infrastructure layer, including observability, agent frameworks, and routing tools like Open Router, is currently under-invested.

The hidden consequence of this shift is that the value is migrating from the intelligence itself to the orchestration of that intelligence. As firms move from Wave 1, which is getting it running, to Wave 2, which is optimizing for efficiency, the companies that provide the management layer for these diverse model stacks will likely capture significant value, even if they aren't the ones training the models.

Key Action Items

• Adopt a Barbell Strategy: Allocate capital toward either early-stage tracker checks to build relationships or late-stage outlier bets where the winner is already anointed. (Immediate)
• Prioritize Access Over Ownership: In high-growth markets, accept lower initial ownership to secure a seat at the table. The payoff comes in the ability to double down on winners later. (12-18 months)
• Audit Your Swim Lanes: If your investment process is rigid regarding stage or valuation, you are likely missing the most significant outliers. Build mechanisms to bypass standard IC hurdles for high-conviction, category-defining opportunities. (Next quarter)
• Shift Focus to Infrastructure: Move beyond the foundation model hype. Invest in the tooling that allows enterprises to manage, route, and optimize multiple models. This is where the next wave of operational value will be captured. (12-18 months)
• Embrace Collaborative Syndication: Shift from the sole investor mentality to a collaborative approach. Syndication is a structural necessity in a market where rounds are too large for any single firm to dominate. (Immediate)
• Build Hard Mode Relationships: Focus on founders who have demonstrated the grit to build in difficult markets, such as Europe. The difficulty of their starting environment often acts as a filter for superior execution. (Ongoing)