Venture capital has always been a business of concentrated bets, but rarely has one wager loomed so large over a firm’s fortunes. Menlo Ventures has closed a $3 billion fund, its largest ever, after going heavily into Anthropic well before the AI lab became a consensus darling. According to a roundup by Build Fast with AI, citing TechCrunch and Bloomberg reporting, that early conviction has appreciated significantly, and it now underwrites Menlo’s biggest-ever raise. The story is not just about one firm’s good timing. It is a case study in how AI conviction is reshaping the math, and the risk profile, of venture capital itself.
The raise
Menlo’s $3 billion vehicle is a milestone for a firm that has operated for decades without needing headline-grabbing fund sizes. The scale is notable, but the backstory matters more: this fund follows an outsized, early position in Anthropic, a bet placed when the outcome was far from obvious. That position has since appreciated considerably, and the returns, on paper at least, have given Menlo both the credibility and the LP appetite to raise at a level it had never approached before.
What’s striking is the causality. This is not a mega-fund raised on the back of a diversified track record across many sectors; it is a mega-fund raised largely on the strength of AI conviction, and one AI position in particular. As Build Fast with AI framed it, the fund illustrates how venture capital is concentrating around a small set of AI leaders, a dynamic that raises both the potential returns and the concentration risk if those bets falter. Capital, in other words, is flowing toward the firms that got the AI thesis right early, and those firms are deploying it right back into the same thesis.
Why it matters
The Menlo raise is a clean signal of where the venture industry’s center of gravity has moved. Mega-funds are increasingly chasing a small cluster of AI winners, the frontier labs and the infrastructure players positioned around them. That narrows the field. Instead of spreading capital across dozens of promising categories, a meaningful share of new venture dollars is pointed at a handful of names that already command extraordinary valuations.
Part of what drives this is reflexivity. In AI, valuations and returns feed on each other: a lab’s rising valuation attracts more capital, which funds more compute and talent, which improves the product and the narrative, which in turn justifies a higher valuation. For investors who got in early, paper markups become the proof point for the next fund, which is then deployed to defend and extend those same positions. It is a self-reinforcing loop that works beautifully on the way up.
Limited partners are leaning in. Pension funds, endowments, and sovereign allocators want exposure to what many believe is a generational platform shift, and the cleanest way to get it is by backing the firms that already have positions in the leading labs. That appetite is precisely what makes a $3 billion fund possible. When LPs are hungry for AI exposure and a firm can credibly claim to have it, the fundraising math takes care of itself. The demand is real, and it is concentrated in the same direction as the deployment.
The risks
The obvious danger is concentration. When a fund’s success, and increasingly an entire slice of the venture ecosystem, rests on a handful of frontier labs, the whole structure becomes correlated. Anthropic, OpenAI, and a small number of peers now anchor a disproportionate share of AI value. If one or two of those leaders stumble, whether through a technical plateau, a business-model reckoning, regulatory friction, or simply losing the model race, the repricing would not be contained. It would ripple across funds, LPs, and the startups built on top of those labs.
There is also the question of pricing discipline in a hot market. When conviction is high and capital is abundant, the temptation is to pay up, because the fear of missing the defining company of the era outweighs the discipline of entry price. But venture returns are made at the point of entry as much as the point of exit. A great company bought at a punishing valuation can still be a mediocre investment. The reflexivity that lifts valuations on the way up offers no protection on the way down; markups are not liquidity, and paper gains can compress fast when sentiment turns.
None of this means the bet is wrong. Menlo’s Anthropic position may well prove to be one of the great venture calls of the decade. But a fund built on that logic is, by design, exposed to a narrow set of outcomes. The upside is enormous; so is the covariance of everything riding on it.
The India read
For Indian founders and investors, the concentration of global AI capital has real downstream effects. When the largest pools of venture money cluster around a few US labs, less attention and fewer marginal dollars are available for everyone else, including the emerging AI ecosystem in India. That can make it harder for Indian AI startups to raise, particularly those building foundational or capital-intensive systems that compete, however indirectly, with the well-funded frontier.
But the picture is not purely negative. The gravitational pull toward a handful of labs also creates space at the application layer, where India has genuine strengths: solving for local languages, cost-sensitive markets, and specific vertical workflows in finance, commerce, and services. Startups that build on top of the leading models, rather than trying to out-spend them, can turn the concentration of foundational capability into an advantage. The labs everyone is funding become infrastructure that Indian companies can deploy without bearing the cost of building it.
The deeper lesson for Indian investors is the tension between conviction and diversification. Menlo’s fund is a monument to conviction, and conviction is how outsized returns get made. But most funds are not built to survive a single bet going wrong, and India’s venture ecosystem, still maturing, has less cushion to absorb a correlated AI drawdown. The pragmatic path is to hold conviction on the shift while diversifying the specific bets: back multiple approaches, favor capital-efficient models, and stay disciplined on price even when the narrative is loud. The winners of the last cycle rewarded early conviction. The survivors of the next one may be the ones who paired it with restraint.
