a16z and top VCs increase seed round investments by 2.37x in the AI era

Author: Pavel Prata

Compiled by: DeepWave TechFlow

DeepSight summary: Mega funds managing over $10 billion are entering seed rounds at an unprecedented pace. Murph Capital pulled data from Harmonic to analyze the early-stage investment behavior of the top 20 mega funds across three eras: the SaaS era, the zero-interest era, and the AI era. The conclusion is not straightforward: while mega funds’ seed conversion rates are indeed 3.7 to 4.2 times the market average, this advantage is rapidly diluted when they scale up deployment. For emerging managers, room for survival still exists—but they must choose the right赛道.

A month ago, I posted a tweet asking a simple question: Are giant funds really taking over seed rounds, or is it just an impression? After 65,000 views and hundreds of direct messages, it’s clear this question struck a nerve.

Emerging Managers (EMs) have written in saying they feel pressure but can’t quantify it; LPs ask: If a16z and Sequoia are already in the game, is there still value in investing in seed funds? Even the GPs of giant funds wonder just how aggressively their competitors are deploying capital at the early stage.

@pavelprata tweeted: Are giant funds really taking over seed rounds? I decided to study the behavior of the world’s largest VC funds (over $10B AUM) in early-stage investing to answer a simple question: Should EM be worried about losing its structural advantage?

A broad consensus quickly formed, and I largely agree:

• Large funds have indeed significantly increased their seed-stage allocations, growing by approximately three times over the past decade.
• The market is large and sufficiently decentralized, but their shares remain relatively small, concentrated primarily in the top quartile.
• Their core motivation is not immediate capital returns, but rather early access to talent and high-signal-to-noise data, minimizing the risk of missing the next generational opportunity.

But consensus is only the starting point. Behind the broad direction lies a more interesting and uneven picture that cannot be seen without data.

So we pulled data from Harmonic, collected the performance of 20 mega funds across three eras (SaaS, zero interest rates, AI), and tried to honestly answer: What exactly happened in the seed market? Where are the mega funds really heading? What impact has this had on pricing? And does EM have real cause for concern?

First, let’s discuss the research framework.

We rely on public intelligence combined with real-time data provided by Harmonic, covering over 30 million companies and 190 million individuals. Over the timeline, we analyzed the past decade, divided into three eras:

• The SaaS era (2015–2019): A five-year normal market cycle. Cloud, SaaS, trading platforms, and fintech were the dominant narratives, with normal interest rates and disciplined markets.
• The zero-interest-rate era (2020–2022): Three years of zero-interest-rate policy. Capital was nearly free, drawing investors from all corners to seek returns in early-stage ventures, with Tiger Global and SoftBank appearing in nearly every meaningful funding round. The seed-stage market became severely overheated, but in a disorganized manner, lacking structural logic.
• The AI Era (2023–2026): From the release of ChatGPT to today. A massive technological shift that has given rise to a new class of companies for which ultra-large seed rounds have become the norm.

Technically, we focus on the seed round, but in practice, we include Pre-Seed and Seed Extension. The reason is simple: the boundaries between these early stages are often blurred or shifting; insisting on precise segmentation would be dishonest.

Let’s get to the point. To be honest, even before I began my research, I had a strong intuition: large funds are appearing more frequently on the radar in early-stage deals. This intuition largely came from social media, where the logos of a16z, General Catalyst, and Sequoia appeared more and more often in seed round announcements, each accompanied by a high-profile media campaign. Data confirms this:

• In the first six months of 2026, a16z participated in approximately 48 seed-stage deals, leading 46% of them. This is a systematic seed strategy, not sporadic bets.
• Most notably, check sizes: the median check size in the a16z-led round is $10.5 million, more akin to a classic Series A than a traditional seed round.
• Including General Catalyst and Sequoia, these three giants completed 87 seed deals in just 5.5 months, averaging one early investment every 1.5 business days.

@a16z tweeted: We’re honored to lead Westmag’s seed round. One underappreciated advantage of investing in the entire hardware stack is gaining firsthand exposure to the supply chain challenges plaguing industrial infrastructure…

Meanwhile, Carta’s latest data shows that seed-stage valuations are rapidly inflating. While some may view this as the result of only a few aggressive players, the fund math for most EMs still forces them to operate at or below the median to secure sufficient initial ownership and maintain a viable return path.

The logic of mega funds is entirely different. With accumulated AUM, brand premium, and access to high-quality project pipelines, price discipline is no longer a real constraint. This divide is fracturing the market into two distinct tiers, which we broadly refer to as “classic seed rounds” and “super seed rounds”:

• The 90th percentile of seed round valuations surged to $93.7 million in Q1 2026, nearly doubling from four years prior.
• Over the past year, valuations above the median have increased by at least 53%.
• The bottom barely moved: the 25th percentile slowly rose from $18 million to $22.7 million.

@PeterJ_Walker tweeted: The top 5% of seed round valuations are now frequently surpassing $175 million, tripling over the past 12 months. There’s just a hint of 2021 absurdity to it (even as an AI believer, I feel this way).

But all of this is still circumstantial evidence, pointing to a general direction without providing definitive answers about what was actually happening in the early market or how systemic the presence of giant funds truly was.

That’s why we decided to dig deeper. We analyzed the individual dynamics of each fund across three eras, breaking down their behavioral patterns and what this shift ultimately means for EM.

Caption: Comparison of early trading volumes among 20 major funds across three eras

On average, a typical giant fund in the SaaS era completed 10.6 early-stage deals per year. In the AI era, this increased to 23.9 deals, representing an average growth of 2.37 times across the entire cohort.

The most interesting thing is what happened after zero interest rates ended. If this growth was merely a byproduct of free money, we would expect it to reverse after rate hikes. But in our dataset of 20 funds, the average annual number of trades in the AI era was nearly identical to that during the zero-rate era: 23.9 versus 24.3. In fact, only three funds slowed their early-stage investment pace. This suggests the shift is structural, even though a few outliers pulled up the overall numbers:

• a16z: 16.6 → 49.7 → 76.8 transactions per year
• General Catalyst: 15.2 → 33.0 → 62.1 deals/year
• Khosla Ventures: 14.6 → 21.0 → 30.9 deals/year

At least three fundamental driving factors are behind this:

Companies in the AI era are inherently more expensive to build. GPU infrastructure, data pipelines, and research scientists earning $300,000 to $500,000 per year have created a completely different baseline cost. What could be accomplished for $500,000 in the SaaS era (two engineers plus AWS) now requires $2 million to $5 million in the AI era. The increased median check size reflects real R&D expenses, not just valuation inflation. Moreover, early-stage SaaS was inherently exploratory (allowing founders to iterate, pivot, and spend years finding PMF), while the first-mover advantage window in AI is far shorter. If your model works, you’ll quickly outpace competitors—and that window closes faster.

The competition for founders has shifted pricing power. In the early stages of a revolutionary technology cycle, top talent with high capability is invaluable. The best AI founders can choose between a16z, Sequoia, and Lightspeed at the seed stage, assembling an investor base that helps them raise larger subsequent rounds more quickly. Often, pricing power shifts from investors to founders: rounds grow larger not because companies objectively need more capital, but because founders can demand and receive it.

The math behind fund sizes is telling. The combined AUM of the top 5 funds in our queue grew from approximately $34 billion to $249 billion over a decade, increasing roughly sevenfold. Meanwhile, the number of their seed deals only increased two- to fourfold. AUM has expanded much faster than seed activity, causing seed checks to represent a smaller share of these funds’ portfolios.

Take a16z as an example: in 2015, it managed around $4 billion; today, it manages $90 billion (including its latest $1.5 billion fundraise, the largest single fund in VC history). A $6 million seed check represents just 0.01% of a $90 billion AUM. Mathematically, the fund has no incentive to negotiate over million-dollar valuations. Conversely, in an increasingly concentrated market, the risk of missing a generational opportunity is catastrophic.

Therefore, we can say with high confidence: the influx of mega-funds into seed rounds in the AI era is not speculative behavior from a free-money era, but a strategic mission. The massive capital flowing into mega-funds, alongside the emergence of a new wave of companies and talent worth competing for at the earliest stages, is driving this transformation together.

Caption: 20 funds grouped by growth trajectory

In the era of zero interest rates, all 20 mega funds in the dataset increased their early-stage investments without exception. Following the pandemic, the Fed lowered interest rates to near zero, causing massive LP capital to flood into VC pockets, with total U.S. VC fundraising reaching a staggering $169.5 billion in 2021.

Holding massive dry powder, some giant funds have moved downstream to seed stage experiments, while others have actively exited late-stage rounds (where valuations were extremely inflated) and similarly shifted downstream.

But in the AI era, interest rates have stabilized above 5%, and the market has become highly differentiated. Macro divergences have divided funds into three behavioral paths:

Accelerator

Trading volume in the AI era has even surpassed that of the zero-interest-rate period:

• a16z (75.3 deals/year)
• General Catalyst (61.5 deals/year)
• Khosla Ventures (31.5 deals/year)

These funds did not merely stay in the seed stage after cheap capital disappeared; instead, they doubled down and aggressively expanded their presence.

Stablecoin

Trading volume in the AI era is slightly below the peak of the zero-interest rate period, but still far exceeds that of the SaaS era:

• Sequoia (19.6 → 49.3 → 50.6)
• Accel (15.2 → 43.3 → 34.7)
• Lightspeed (11.6 → 41.7 → 32.1)

The surge in zero interest rates has peaked and begun to decline, but baseline activity has permanently risen to two to three times historical levels. There's no going back.

Disciplined

Three eras of steady growth:

• Bessemer (9.4 → 23.0 → 20.9)
• Lux (7.2 → 14.3 → 14.7)
• Index Ventures (10.0 → 23.3 → 17.6)

They avoided the surge of zero interest rates and the AI boom, but the baseline has permanently shifted upward. In the SaaS era, it was 10 transactions per year; now it’s steadily at 15–21.

The only exceptions are three funds: Founders Fund, NEA, and Greylock, which either reduced or maintained their early-stage activity from the SaaS era to the AI era.

Founders Fund may be the only institution to have made a philosophical, deliberate choice. Peter Thiel’s contrarian framework, deeply influenced by Girard’s mimetic theory, treats crowded market consensus as a clear signal to seek opportunities elsewhere. So while 17 other giant funds rushed into seed stage investments, Founders Fund did the opposite, shifting toward large, concentrated late-stage bets, channeling capital into generational outliers like OpenAI, Databricks, and Anduril.

Greylock remains deeply committed to the tradition of the "first check," but chooses to play a high-conviction game. Instead of operating like a transaction factory, it focuses on fewer, higher-conviction bets, sometimes even incubating companies directly within its own office.

NEA's large, multi-phase mission makes it harder to analyze seed fluctuations in isolation, and without hard data, we do not speculate.

Caption: Change in the proportion of early trades to total investment across funds

Absolute numbers cannot answer a key question: For these giants, is the seed round a side project or a core strategy?

A fund makes 30 seed investments in a year, but if it simultaneously makes 200 A- to D-round investments, seed rounds account for only 15%. Conversely, if 30 seed investments come from a total of 60 investments, seed rounds account for 50%.

15% means scouting projects, individual partners' pet projects, and cheap options. 50% means a strategic mission: dedicated teams, institutionalized processes, and large-scale deployment of machines.

This is why our third (and perhaps most revealing) perspective tracks the exact proportion each giant fund invests in early-stage ecosystems:

Of the 20 funds, 16 have set a record-high early allocation ratio in the AI era. During the SaaS era, a typical mega fund directed 20–30% of its trading volume to seed stages; in the AI era, this baseline has surged to 35–50%.

Three cases are particularly compelling:

Sequoia: A complete transformation. This is the most dramatic strategic shift in our entire dataset. During the SaaS era, Sequoia’s early-stage investments accounted for less than one-fifth of its portfolio; it was primarily a dominant player in Series A/B+ rounds, making only occasional tactical seed bets. By the AI era, nearly half of its deals were at the early stage—an increase of 30 percentage points.

General Catalyst: V-shaped curve. During the SaaS era, GC shifted toward early-stage investments, accounting for 38%. In the zero-interest era, this dropped to 30%, aligning with peers in chasing growth-stage returns fueled by free capital. But the AI era triggered a sharp reversal, rising to 47%. This is a deliberate, aggressive return to early-stage investing, reaching a higher peak than ever before.

a16z: After stabilizing its baseline, AI surged. a16z’s uniqueness lies in its perfect balance of 31.2% allocation to SaaS and the zero-interest-rate era. While other funds sank into chaos during the zero-interest-rate period, a16z maintained structural equilibrium. Then came the AI era, sharply jumping to 42.5%.

This breakdown is important because LPs often hear the same narrative from large funds: “We write seed checks when we encounter exceptional founding teams.” Data proves this pitch is dead.

Sequoia holds a 49% seed share, GC holds 47%, and a16z holds 42%. Giant funds have shifted their core focus to the seed stage and have weaponized this shift with dedicated teams, customized internal pipelines, and proprietary accelerator programs such as a16z Speedrun and Sequoia Arc.

For EM, this provides a crucial but sobering context: your daily competition has long surpassed the $500 million boutique fund next door. Today, when you’re vying for allocations, your rivals are giants with $10–90 billion in AUM, already directing 40–50% of their institutional trading machinery toward your space.

To truly understand the mechanism of this pressure, you must also factor in a key metric: check and round sizes.

Caption: Median seed round participation by mega funds vs. U.S. seed market median

A core theme we previously emphasized is the fragmentation at the seed stage. The best way to view this divide is by examining the median round size of each era and benchmarking it against the overall U.S. seed index (the market-wide median).

• In the AI era, the median U.S. seed round amount on shareholder tables is $6.2 million.
• The overall market median is only $1.4 million. The gap is 4.4 times.

Big funds do not participate in 'average' seed rounds; they systematically operate in the top quartile of the market.

More interestingly, this multiple remains stable across three macro cycles: 4.8x in the SaaS era, 4.5x in the zero-interest era, and 4.3x in the AI era. Mega funds have not accelerated inflation relative to other parts of the market—they have simply always existed in a completely different price tier.

From another perspective, the 75th percentile of the market ($4 million) serves as the entry baseline for mega funds. Their median round ($6.2 million) sits firmly above the entire U.S. seed ecosystem’s P75, and by definition, these giants are restricted to transactions within the top 25% by size.

But when we overlay the median and the average, things become more interesting.

Caption: Median vs. average across funds reveals the dual-track strategy

The median reflects a fund's "typical" trade, while the mean is heavily skewed by outliers. The gap between the two clearly indicates how "dual-track" a fund's strategy is: is it aggressively operating a dual-engine model targeting super-seed stages, or does it function uniformly across a single price tier?

From this perspective, the queue clearly splits into two categories.

Dual-track (spread of 3x or more)

• Index (median 8.2 million, average 34.3 million, 4.2x spread)
• Lux (6 million vs. 31.7 million, 5.3x)
• Lightspeed (6.8 million vs. 30.8 million, 4.5x)
• Accel (5 million vs. 26 million, 5.2x)
• a16z (6 million vs. 21.8 million, 3.6x)
• Sequoia (5 million vs. 17.4 million, 3.5x)

These funds play on two tables at once: large classic seed rounds ($5–8 million), plus a highly curated selection of super-seed rounds ($50 million–$500 million+), which pull the statistical average way up. Those “$100 million seed round!” TechCrunch headlines don’t reflect everyday reality—their typical deals are actually 4–5 times smaller.

Homogeneous (spread below 2.5x)

• Greylock (6.9 million vs. 13.3 million, 1.9x)
• Founders Fund (7 million vs. 12 million, 1.7x)
• CRV (7.5 million vs. 10.8 million, 1.4x)
• 8VC (6.6 million vs. 8.7 million, 1.3x)
• NEA (7 million vs. 7.4 million, 1.1x)

The median and mean of these funds closely track each other, with no long tail from oversized rounds. They consistently deploy within the $5 million to $8 million price range, with no significant outliers.

Dual-track funds dominate headlines, creating the illusion that the seed round has become a game above $30 million. But data refutes this: even the most dual-track institutions typically see deals firmly in the $500–800 million range. Super seeds are merely the long tail of the distribution, not the center.

For EM, the real competitive pressure comes from homogeneous firms—GC, Khosla, Bessemer, Greylock. These firms systematically operate in the $5 million to $8 million range and are not distracted by super-seed deals. Dual-track funds may sound more intimidating on paper, but they pose less daily competition; part of their time is spent in the super-seed market, which EM wouldn’t compete in anyway.

The fragmentation of the seed market has little to do with abstract rounds of inflation. We are witnessing two entirely separate ecosystems emerging under the single label of "seed round": super-seed rounds (over $20 million) belong to dual-track platforms, while traditional seed rounds ($3–8 million) remain the space where giant funds and EMs still collide. The only difference is that the number of multi-stage giants crowded into this classic range has doubled.

Caption: Comparison of lead investment rates and number of lead investments by fund

Participating and leading are two fundamentally different things.

A fund invested $500,000 in a $6 million round as a follow-on investor, merely a passenger on the cap table. The lead fund in that round was the one setting the valuation, terms, and deciding who could join the co-investment—it was the lead investor who ultimately determined whether there was still room for EM.

What percentage of all seed transactions executed by giant funds are they actually leading?

I’ve categorized these institutions into four types:

Lead-Investor Driven — High Lead Investor Rate + High Trading Volume

• Khosla (60%, 19 lead investments/year)
• Lightspeed (63%, 21 lead investments/year)
• Accel (54%, 20 lead investments/year)

This is the most dangerous group for EMs. They deploy aggressively and demand the lead seat. Lightspeed leads 21 seed rounds annually, with a 63% lead rate, systematically dominating early-stage pricing. If an EM is competing for the same company, it’s fighting for the lead position.

Large-scale — High trading volume, moderate lead investment rate

• a16z (51%, 40 lead investments/year)
• General Catalyst (53%, 33 lead investments/year)
• Sequoia (36%, 19 lead investments/year)

These giants dominate in absolute number of lead investments, even if their lead investment percentage is lower. They take the lead in the best companies in their pipeline, while others take passive positions. This poses a double threat to EM: even if the mega funds don’t lead, their presence on the cap table significantly impacts signaling and subsequent financing dynamics.

Featured Lead-Investor Type — High Lead Investment Rate, Low Trading Volume

• EQT (82%, 7 transactions/year)
• Craft Ventures (76%, 8 deals/year)
• Index Ventures (67%, 12 deals/year)
• Founders Fund (61%, 10 transactions/year)
• Greylock (58%, 6 transactions/year)

These funds led the majority of trades while maintaining high discipline and low speed. Driven purely by conviction: if they write a check, they almost always take a position in that round. They pose little threat to overall market volume, but in any specific deal they enter, they almost certainly secure the lead investor role.

Network-type — Low investment rate

• 8VC (38%, 9 deals/year)
• Amplify (39%, 4 transactions/year)
• Sequoia (36%, 19 transactions/year)
• Bessemer (44%, 9 transactions/year)

These institutions invest alongside lead investors far more frequently than they lead rounds. Their role at the seed stage centers on network, signaling, and optionality, rather than establishing market pricing. They pose the least threat to EMs, as they rarely displace lead investor positions.

An interesting finding: The two funds with the highest absolute early-stage activity volumes have the lowest lead investment rates in the AI era: a16z at 51% and Sequoia at 36%. Both also have lower lead investment rates than during the SaaS era (a16z dropped from 67%, Sequoia from 52%).

It’s simple to explain: when you make 77 or 51 trades per year, it’s physically impossible to lead every single one. Some transactions naturally shift to scout bets, co-investments, and joint investments led by others. At this scale, trading volume and lead rate are a clear trade-off.

In absolute numbers, they still dominate the battlefield: a16z leads approximately 40 early-stage deals per year, and GC leads about 33—more than the combined total of early-stage deals from half the funds on the list.

Overall, the lead investment rate for most funds has increased in the AI era. Thirteen out of twenty funds have a higher lead investment rate in the AI era than in the SaaS era:

Caption: Change in lead investment rates by funds in the SaaS era versus the AI era

Large funds are leading rounds more frequently. For example, Greylock led one in every four seed rounds during the SaaS era, but now leads more than half in the AI era. They have completely shifted from a passive “invited to participate” stance to an active “I’ll lead this round” approach.

LPs must keep this reality in mind when conducting due diligence on funds. Of course, GPs love to fill their fundraising pitch decks with logos of giant funds and the phrase “We co-invest with xxx.” But this dynamic can actually serve as a key signal defining what type of venture capital product LPs are truly investing in.

If an LP asks: “How many rounds did you lead last year, and how many of those had a mega-fund as the co-lead?” and the answer is: “We often co-invest with a16z or GC,” that’s not a structural advantage—it’s a heavy reliance on the deal flow of mega-funds. This isn’t necessarily a bad strategy, but once you factor in larger round sizes, inflated valuations, and diluted ownership targets due to lack of pricing power and lead-investment capability, the underlying fund mathematics changes dramatically.

Conversely, if the answer is “We led rounds that mega funds avoid, or we arrived long before they noticed”—that’s where EM’s true, defensible advantage lies.

Caption: Early activity of mega funds distributed by sector

The above analysis of trading dynamics, round inflation, and lead rates describes the overall situation of mega funds. But in reality, an EM rarely invests in the “entire seed stage”; instead, they invest in specific sectors, and sector selection is often their core strength. So the next logical question is: Where exactly have the mega funds gone?

From this perspective, their footprint is much more concentrated than the overall statistics suggest.

As expected, the enterprise AI and automation, and AI infrastructure and developer tools sectors led in both lead investment rates and total number of transactions. Together, these two sectors encompass 538 companies, accounting for 42% of all early-stage activity in the dataset. All 20 mega funds are actively involved in both sectors. Three core drivers underlie this trend:

Market size. Enterprise spending on generative AI has surged from $1.7 billion in 2023 to $37 billion in 2025, increasing more than 20-fold in just two years. Enterprise AI now accounts for 6% of the global SaaS market, growing faster than any software category in history.

Speed. The pace of time in the AI era is unprecedented. The growth model of the SaaS era was T2D3 (triple, triple, double, double, double); the growth framework for top AI-native companies is Q2T3 (quadruple, quadruple, triple, triple, triple). For funds, the window to enter at the seed stage is closing faster. Hesitating for 12–18 months could mean missing an entire software category.

Performance outliers. Lovable reached $100 million in ARR within 8 months and doubled to $200 million in just 4 months, surpassing OpenAI, Cursor, and all other software companies in history. By May 2026, Sacra estimated Lovable’s annualized revenue had surpassed $5 billion. Cursor raised $2.3 billion at a $29.3 billion valuation. Anthropic’s annualized revenue accelerated from approximately $1 billion at the end of 2024 to $14 billion in February 2026, $30 billion in April, and $47 billion in May, while raising $65 billion at a $965 billion valuation. All of these companies either did not exist or were completely unknown three years ago.

For AI-focused EMs, this means nearly every mega fund is hunting in your backyard. Armed with unlimited capital, these giants are unconstrained by round pricing and can aggressively lead rounds to maximize their stake targets. The survival of new fund managers depends on deep domain expertise, exclusive access to dense founder networks, and the ability to invest even before founders have a pitch deck.

Another key detail: The fastest-growing AI companies (so-called "AI supernovae") have an average gross margin of only about 25%, deliberately sacrificing unit economics to capture market share. Even the more traditional "comets" average only around 60% gross margin, still significantly below the classic SaaS benchmark of 70–85%.

This means that enterprise AI is currently a sector where revenue growth far outpaces profitability. Investors are essentially betting on future economics rather than current margins. Mega funds, with their deep pockets and long time horizons, can easily absorb this structural bet. But an EM managing a $25–75 million fund would be fundamentally vulnerable if the realization of future unit economics takes longer than the market expects.

Caption: Median vs. average round sizes across each赛道

AI infrastructure and developer tools warrant special attention in terms of round structure. The dual-track behavior observed at the fund level is most pronounced in this sector: a median round of $6.8 million, with the average soaring to $48 million—a sevenfold difference.

This massive discrepancy indicates that the space is flooded with super seed rounds exceeding $100 million, inflating the statistical average. This is precisely the breeding ground for headlines like “$50 million seed round,” creating a severely distorted impression of typical transactions for observers.

In contrast, the spread for Commerce & GTM is only 1.4x, and for Healthcare it is 2x. The further away from AI core, the more homogeneous the round landscape becomes.

The behavior of two sectors is disproportionate to their actual size:

Cybersecurity: Only 76 companies, but with a lead investor rate of 62%, the highest among all major sectors. Combined with a median round size of $7 million—one of the highest in the dataset—mega funds dominated pricing in nearly two-thirds of transactions.

Defense and Aerospace: Smaller footprint (34 companies), but lead investment rate hit a record 66%. However, only 12 of the 20 mega-funds were active, indicating concentrated bets by a few high-conviction players rather than systemic, platform-level pressure.

Other sectors are relatively less crowded: climate and energy (26 companies, 12 active funds), logistics (24 companies, 13 active funds), as well as traditional sectors such as PropTech, EdTech, Legal, and HR.

EMs with deep domain expertise in these niches completely escaped platform domination. The competition isn’t 20 large platforms, but 8–12 institutions, pricing 2–3 deals per year—a completely different game.

This is an important practical insight for LPs: due diligence on EMs must shift to the specific sectors they participate in, as sector choice determines the nature of competition and the type of differentiation needed to win.

Caption: Giant fund backing vs. seed to Series B conversion rate for the overall market

So far in this study, we have only shown one side of the coin: giant funds entering seed rounds, making more trades, leading rounds more frequently, and operating within EM price ranges.

But there’s one question we’ve been pushing off until now—it may be the most critical issue in the entire study: Does this strategy actually work?

Yes, mega funds write larger checks, participate in rounds 4.4 times larger than the market median, and direct 40-50% of transaction activity toward the early stage, leading more than half of all seed deals. But if the companies they invest in at the seed stage do not have a higher survival rate than the market average, then everything we’ve described is merely valuation inflation without real value.

Conversely, if startups backed by mega funds reach Series B at a significantly higher rate than the market, the narrative flips. In that scenario, mega funds aren't just "taking over" seed rounds—they're making seed rounds better. LPs should ask: "Why not concentrate capital in mega funds that cover the seed stage, then double down in subsequent rounds to capture the entire market lifecycle within a single institution?"

To this end, we calculated a straightforward metric: what percentage of companies that raised seed funding in a specific era later reached Series B? We compared two groups: the overall market versus seed-stage companies with at least one mega fund on their cap table.

We focus on the SaaS era and the zero-interest-rate era (companies in the AI era are too young). The outcomes are clear, but with nuances.

• The SaaS era: Among 60,110 seed-funded companies, 9.8% reached Series B. Among the 940 with participation from major funds, this figure jumps to 36.7%, a 3.7x increase.
• The era of zero interest rates: Consistent trend: Market at 3.9%, mega funds at 16.5%, gap widening to 4.2 times.

Megafunds are 3.7 to 4.2 times more capable than the market average at converting seed rounds into Series B rounds. More importantly, this gap is widening. In an overheated environment where overall market conversion rates have plummeted during the zero-interest era, the quality screening of megafunds has become even more valuable.

But before drawing conclusions, let’s break down why the conversion rate is so high. Several structural drivers can be collectively referred to as a strong signal effect:

• Elite Series A pipeline: Top-tier Series A investors proactively seek to co-invest with institutional-grade, heavyweight seed lead investors.
• Internal co-investment capability: Large funds have deep pockets and can lead the Series A or Series B rounds for their own portfolio companies.
• Brand-driven talent acquisition: Top engineers see labels like “Invested by Sequoia” or “Invested by a16z,” and hiring friction significantly decreases.
• Media distribution advantage: Greater PR leverage leads to more proactive outreach from potential enterprise clients

Therefore, it must be recognized that a significant portion of the conversion rate is not the result of mega funds "picking the right" companies, but rather mega funds helping companies become the right choices. For LPs, this is a clear signal: the value-added role of mega funds at the seed stage is not just "stock picking," but true "platform as a product."

Caption: Seed-to-Series B conversion rates for funds in the SaaS era versus the zero-interest-rate era

But there’s another side to the coin. When we move beyond aggregate data and examine each institution individually, a troubling pattern emerges: of the 15 funds with sufficient sample sizes (over 10 seed deals per era), 14 experienced a sharp decline in conversion rates from the SaaS era to the zero-interest era—drops ranging from 10 to 25 percentage points:

• Lux: 51% → 19%
• Sequoia: 46% → 14%
• a16z: 42% → 16%
• Index: 45% → 25%

Direct correlation: The funds that increased deal volume the most during the zero-interest era saw the steepest drop in conversion rates. Sequoia’s deal volume tripled (from 20 to ~50 deals/year), while its conversion rate plummeted from 46% to 14%. Lightspeed’s volume quadrupled (from 12 to 42 deals/year), and its conversion rate fell from 31% to 11%.

The only exception is Greylock, whose conversion rate jumped from 29% to 44%. This is no accident: Greylock is the only firm that maintained roughly stable trading volume during the zero-interest era (11.0 to 11.3 trades per year). Fewer trades yielded a higher hit rate. Trading volume discipline directly equals portfolio quality.

This set of conversion data both validates and complicates our entire narrative.

On one hand, it demonstrates that mega funds have indeed delivered tangible results at the seed stage. A 3.7x conversion rate premium is neither coincidental nor a data artifact—companies that secured backing from mega funds early on truly survived and grew better. This is a compelling argument for LPs: brand, network, and platform resources deliver measurable value.

On the other hand, trading volume and quality are always in tension. In today’s AI era, seed trading volumes of giant funds are breaking records. If the zero-interest rate model repeats itself, conversion rates will inevitably be eroded. The only question is by how much. Will the platform effects and signal advantages of these giants in the AI era be sufficient to offset the dilution caused by their large-scale deployment pace?

There will be definitive answers in 3 to 5 years. But historical data offers a sobering warning: mega funds have proven they can pick winners at low trading volumes. They have yet to prove they can do so at scale.

It is precisely in this gap—the space between the verified past and the unverified present—that the real opportunity for EM to do less, but better, lies.

Caption: Risk Index Ranking of 20 Major Funds for Emerging Managers

In the closing section, we did something controversial: created a Danger Index.

This is a data-driven ranking that measures which major funds pose a genuine competitive threat to EM. We anchor on three pillars:

Trading volume: The absolute number of early trades per year in the AI era. The higher, the more frequently EM encounters them in practice.

Strategic commitment: The percentage of the fund’s total investment activity allocated to early-stage ventures. 45% indicates a core strategy, supported by dedicated teams and institutionalized processes. 20% indicates a sideline effort, with the fund potentially scaling back and refocusing on later stages at any time.

Price overlap: The median size of the funding rounds in which funds participate. This may be the most critical factor. Mega-funds participating in $8–10 million rounds primarily compete with other multi-stage giants. But mega-funds operating in the $4–5 million range directly compete with EMs—this is precisely the sweet spot where $50–100 million seed funds deploy capital.

Score each factor from 0 to 10; the final risk score is the sum of all three, with a maximum of 30.

The results were unexpected. Four institutions fell into the first tier (greatest threat): General Catalyst, a16z, Sequoia, and Accel.

• These four firms each complete 37–83 early-stage deals annually, allocate 39%–50% of their total investment activity to seed stage, and operate within the $4.4–5.4 million round range—directly targeting EM’s domain.
• Counterintuitively, GC ranks ahead of a16z despite a16z’s higher absolute trading volume (83 vs. 65). The difference lies in GC’s perfect alignment with three risk vectors: high speed, the highest early allocation ratio in this tier (48%), and a median round size of $5 million—right at the center of the EM pricing sweet spot. a16z’s pricing is slightly higher (median of $5.4M) with slightly lower early concentration (43%). The gap is subtle but statistically significant.
• Sequoia ranking third was also surprising. It has the lowest lead rate among the top five funds (36%), investing as a follower far more often than as a lead. Yet its median round size is only $4.6 million—the lowest among large platforms. It is systematically buying rounds that are cheaper (by the standards of giant funds).
• In contrast, Index Ventures surprisingly landed in the third tier, despite maintaining an annual rate of 19 deals and a 66% lead investor rate. Why? A median round size of $8.4 million. Index operates entirely above the traditional EM range.
• The same structural logic applies to Founders Fund (median of $7.8M) and Greylock (median of $7M), both firmly positioned in the third tier. They have a clear early-stage footprint but have not entered the price ecosystem most EMs are fighting to survive in.

The danger index is not a death sentence for EM. Think of it as a minefield map.

It distills the entire macro research into a practical, high-stakes question: “Which top-tier platforms are hunting in your exact price range and sector?”

If the answer is “GC and a16z—both invest in AI software, both entered at the $4–6 million round,” then EM must clearly articulate to LPs: What specific advantages enable you to outperform two firms that together make 150 seed investments annually right in your backyard?

If the answer were, "There are no first-tier giants; I'm leading a $2-3 million investment in climate tech," that would be an entirely different conversation. The risk index indicates that institutional pressure in that space is structurally lower, and deep domain expertise itself can serve as a highly defensible advantage.

• The average number of early-stage deals by mega funds increased from 10.6 per year in the SaaS era to 23.9 in the AI era. Only 3 out of 20 reduced their activity. This is a structural shift, not a cyclical one.
• Seed round valuations have sharply diverged. The 90th percentile reached $93.7 million in Q1 2026, nearly doubling over four years. The 25th percentile rose only from $18 million to $22.7 million over the same period.
• In the AI era, the median seed round for mega funds is $6.2 million, compared to $1.4 million for the overall market—a 4.4x gap that has remained stable across three eras.
• 16 out of 20 funds have set record-high early allocation ratios for the AI era. Typical mega funds have increased their allocations from 20-30% during the SaaS era to 35-50% today.
• 13 out of 20 funds now lead more seed rounds than they did in the SaaS era. Greylock increased from 24% to 58%. A passive co-investment stance is being replaced by structured lead-investment strategies.
• 42% of the giant funds' early activities were concentrated in two sectors: enterprise AI and automation, and AI infrastructure and developer tools. All 20 funds were active in both sectors simultaneously.
• Startups backed by giant funds are 3.7 to 4.2 times more likely to reach Series B than the market average. However, among the 15 funds with sufficient sample sizes, 14 saw their conversion rates drop sharply from the SaaS era to the zero-interest era—the largest declines occurring precisely among those that had grown the most.
• Greylock maintained trading volume flat in a zero-interest-rate environment and was the only fund with an actual increase in conversion rate. Trading volume discipline equals portfolio quality.
• The risk index ranks GC, a16z, Sequoia, and Accel in the top tier—the only four funds that simultaneously meet the criteria of high speed, an early allocation ratio of 39–50%, and a median round size below $5.5 million, directly falling within EM’s pricing sweet spot.
• Traditional sectors such as climate and energy, logistics, PropTech, and EdTech remain structurally less crowded—with only 8 to 13 mega funds active (compared to 20 in the AI sector) and lead investment rates significantly below the category average.

The influx of giant funds into early-stage markets is not a temporary anomaly of a technological cycle, but a permanent recalibration of the fundamental workings of venture capital.

As multi-stage giants continue to absorb the top quartile of seed ecosystems with hundreds of billions of dollars, trying to outcompete them in their high-speed, deep-pockets game is a mathematical dead end. But data reveals a critical crack in their seemingly impenetrable armor—the unavoidable tension between scale of deployment and the quality of portfolio conversion.

In the AI era, EM’s true advantage is no longer about scaling up institutional trading machines or blindly chasing popular assets whose pricing rules are set by top-tier platforms. Instead, it lies in the disciplined selection of niches, the patience to invest in complex future unit economics often overlooked by giant funds, and the courage to remain small, highly focused, and deeply aligned with founders before multi-stage platforms even notice their existence.

In a venture capital ecosystem increasingly崇尚 pure scale, EM’s ultimate counterstrategy is to command a premium for absolute discipline, rather than matching the transaction volume of giants.