Are Mega-Funds Taking Over Seed?
We analyzed 20+ mega-funds transforming the early-stage landscape and unpacked how their AI-era allocations, lead rates, and pricing power really work
A month ago I posted a tweet with a pretty simple question: are mega-funds actually taking over Seed or does it just feel that way? 65,000+ impressions and a few hundred DMs later, it was clear the question had hit a nerve.
Emerging managers wrote saying they felt the pressure but couldn't measure it, LPs asked whether it still made sense to allocate to seed funds if a16z and Sequoia had moved in, GPs at the mega-funds themselves wanted to know where their firm stood relative to competitors and how aggressively those competitors were actually deploying at early stages.
Many people commented and quoted this tweet and broad consensus emerged quickly, and basically I agree with it:
Mega-funds have meaningfully increased their Seed allocation – roughly 3x over the past decade
The market remains large and fragmented enough that their share is still relatively small, concentrated primarily in the top quartile
And their core motivation isn't immediate capital return, but early access to talent and high-signal data to minimize the risk of missing out on the next generational opportunity
But consensus is just a starting point. Behind the broad strokes is a far more interesting and uneven picture that you can’t see without data.
So we pulled Harmonic, gathered data on 20 mega-funds across three eras (SaaS, ZIRP, and AI) and tried to answer honestly: what’s actually happening to the Seed market, where exactly are mega-funds going, how is it affecting pricing, and ultimately – do emerging managers have a real reason to worry?
In Partnership With
Legion is a high-signal marketplace matching emerging GPs with relevant LPs, featuring seamless fund admin, built-in secondary liquidity, and over $300M in investor capital flows.
Harmonic aggregates real-time data on 30M+ companies and 190M+ people to surface the signals that actually matter – all through AI-powered workflow.
Rings AI is an intelligent CRM built for VCs and LPs with auto-enriched data on 100M+ contacts, and a visual map of who-knows-who across your network.
Cura is a VC portfolio intelligence on autopilot that connects your Granola notes, iMessage, LinkedIn data and more to automate support and intelligence.
The Gut Feeling vs. The Data
Before diving into the data, it is worth laying out the baseline assumptions and parameters of our research.
First, we relied on open-source intelligence and paired it with data from our partner, Harmonic, which aggregates real-time insights on over 30M+ companies and 190M+ people.
As for the timeline, we analyzed the past decade and divided it into three distinct eras:
The SaaS era (2015-2019): 5 years of a normal market regime. Cloud, SaaS, marketplaces, and fintech were the dominant theses, interest rates remained standard, and the market operated with general discipline.
The ZIRP era (2020-2022): 3 years of zero interest rate policy. Capital was practically free, investors of all stripes moved down-funnel in search of yield, and Tiger Global and SoftBank seemed to participate in almost every meaningful round. The Seed market became severely overheated though in a chaotic way, lacking any structural logic.
The AI era (2023-2026): The period from the launch of ChatGPT to the present day. It is defined by a massive technological shock and the emergence of a new class of companies for whom mega-seeds have become the norm rather than the exception.
Technically, our analysis focuses on Seed rounds. In practice, however, we included Pre-Seed and Seed Extensions. The primary reason for this choice is that the boundaries between these early stages are often blurred or shifting, so drawing them with surgical precision would be intellectually dishonest.
But let’s start with the big picture. To be frank, even before starting this research, I had a strong gut feeling that mega-funds were appearing on the early-stage radar much more frequently. This intuition was largely shaped by social media, where announcements of Seed rounds featuring logos like a16z, General Catalyst, and Sequoia flashed with increasing frequency, always accompanied by high-profile media campaigns. The data backs this up:
For instance, in the first 6 months of 2026, a16z participated in roughly 48 Seed deals, leading 46% of them – clear evidence of a systematic Seed strategy rather than a series of sporadic bets.
While every such round was heavily marketed, what stood out the most was the check size: the median round led by a16z clocked in at $10.5M, which inherently feels much closer to a classic Series A than a traditional Seed round.
If you add General Catalyst and Sequoia into the mix, these 3 giants collectively closed 87 Seed deals in just 5.5 months. That translates to an aggressive deployment pace of one early-stage round every 1.5 business days.
In parallel, recent data from Carta reveals that when looking at the market through the lens of valuations, they are inflating rapidly. While one might assume this is simply the result of aggressive behavior by a few outlier players, the fund math for most emerging managers still forces them to operate near or below the median to secure sufficient initial ownership and maintain a viable path to capital return.
Mega-funds, however, operate on a fundamentally different logic. Given their cumulative AUM, brand equity, and premium deal flow, price discipline ceases to be a real constraint. This disparity is fracturing the market into two distinct classes, which we loosely call classic Seed and mega-Seed:
The 90th percentile of Seed valuations skyrocketed to $93.7M in Q1 2026, representing a nearly 2x increase compared to four years prior.
Over the past year alone, valuations above the median surged by at least 53%.
The lower percentiles are barely moving: the 25th percentile crept up from just $18M to $22.7M over the same four-year period.
Yet, all of this remains circumstantial evidence – pointing in a general direction rather than providing a definitive answer about what is actually happening in the early-stage market and how systematic the mega-fund presence truly is.
That is precisely why we decided to look deeper. We analyzed the individual dynamics of each fund across all three eras to uncover their behavioral patterns and unpack what this shift ultimately means for emerging managers.
Deconstructing the Deal Machine
Looking at the averages, during the SaaS era, a typical mega-fund in our dataset completed 10.6 early-stage deals per year. In the AI era, that number jumped to 23.9, representing a 2.37x average increase across the entire cohort.
But the most interesting part is what happened after ZIRP. If this surge were purely a byproduct of free capital, it should have reversed after the rate hikes of 2022–2023. Yet, among the 20 funds in our dataset, the average annual deal count remained virtually unchanged after the ZIRP era ended, clocking in at 23.9 in the AI era versus 24.3 during ZIRP. In fact, only 3 out of the 20 funds scaled back their early-stage investment pace compared to the SaaS era. This proves that the shift is genuinely structural, even though a few outliers are undeniably pulling the aggregate metrics upward:
a16z: 16.6 → 49.7 → 76.8 deals/year
General Catalyst: 15.2 → 33.0 → 62.1 deals/year
Khosla Ventures: 14.6 → 21.0 → 30.9 deals/year
But the underlying drivers of this shift remain fundamental, and there are at least three of them:
First, the AI era inherently birthed a class of companies with a significantly higher cost of production. GPU infrastructure, data pipelines, and research scientists commanding $300K-$500K in compensation create a completely different baseline cost. What cost $500K in the SaaS era (two engineers and AWS) now requires $2M-$5M in the AI era. Therefore, the expanding median check size partially reflects real R&D expenses rather than just valuation inflation. Furthermore, while the early stage in the SaaS era was exploratory by nature (allowing founders to iterate, pivot, and search for product-market fit over several years) the AI landscape grants a much stronger first-mover advantage. If your model works, you break away from the competition rapidly, and that window closes much faster.
Second, the fierce competition for founders shifted the pricing power. At the dawn of a revolutionary tech cycle, high competency paired with elite talent is worth its weight in gold. The best AI founders can choose between a16z, Sequoia, and Lightspeed right at the Seed stage, building a cap table designed to help them raise an even larger subsequent round in an even shorter timeframe. In many cases, this shifted pricing power from the investor to the founder: rounds grew larger not because companies objectively needed more capital, but because founders could demand it and get it.
Third, we must look at fund sizes, because the math here is deeply telling. The combined AUM of the top 5 funds in our cohort grew from roughly $34B to $249B – roughly a 7x increase over 10 years. Meanwhile, their Seed deal count grew by only 2-4x. This means AUM expanded significantly faster than seed activity, making Seed checks a proportionally smaller part of these mega-funds’ portfolios.
Take a16z: in 2015, they managed around $4B – today, they manage $90B, factoring in their latest $15B raise (the largest in VC history). A $6M Seed check out of a $90B AUM represents just 0.01% of the fund. Mathematically, the fund has zero incentive to fight over every million in valuation. Conversely, the risk of missing out on a generational opportunity in an increasingly concentrated market is catastrophic.
Thus, we can state with a high degree of confidence that the influx of mega-funds into Seed during the AI era is not the opportunism of the free-money epoch, but a strategic mandate fueled simultaneously by the massive capital inflows into mega-funds and the emergence of a new class of companies and talent worth competing for at the absolute earliest stage.
Growth-Based Cohort Analysis
While I briefly touched upon the shifting average deal counts across different eras, this section takes a deeper, more granular look at their specific behavioral trajectories.
During the ZIRP era, every single one of the 20 mega-funds in our dataset ramped up their early-stage deals per year compared to the SaaS era. Post-COVID monetary easing pushed the Federal Reserve to cut interest rates to near-zero, unleashing a massive wave of LP capital into VC pockets, which drove total US VC fundraising to a staggering $169.5B in 2021.
Armed with massive dry powder, one cohort of mega-funds stepped down-funnel to test the waters at the Seed stage and experiment with new frameworks. Another cohort actively shifted away from late-stage rounds (which were battling hyper-inflated valuations at the time) and similarly moved down-market.
However, in the AI era with interest rates holding steady above 5% the market highly fragmented. This macro divergence split the funds into three distinct behavioral lines:
The Accelerators
Their deal volume in the AI era is even higher than it was during ZIRP:
a16z (75.3/yr)
General Catalyst (61.5/yr)
Khosla Ventures (31.5/yr)
These funds didn’t just stay active in Seed after cheap money disappeared – they doubled down and aggressively ramped up their presence.
The Stabilizers
Their AI-era deal count is slightly lower than their ZIRP peak, but still significantly higher than during the SaaS era:
Sequoia (19.6 → 49.3 → 50.6)
Accel (15.2 → 43.3 → 34.7)
Lightspeed (11.6 → 41.7 → 32.1)
While their ZIRP-era spike has leveled off, their baseline activity remains permanently elevated at 2-3x their historical norm. There is no going back to the old status quo.
The Disciplined
Steady, incremental growth across all three eras:
Bessemer (9.4 → 23.0 → 20.9)
Lux (7.2 → 14.3 → 14.7)
Index Ventures (10.0 → 23.3 → 17.6)
They avoided ZIRP spikes and AI explosions, yet their baseline shifted upward permanently. What used to be 10 deals per year in the SaaS era is now consistently 15–21.
The only exceptions in our dataset are three funds – Founders Fund, NEA, and Greylock). Each of them reduced or held flat their early-stage activity from the SaaS era to AI.
While we lack granular internal data for these three, we can make a few high-probability assumptions based on their market moves.
Founders Fund is arguably the only firm that made a deliberate philosophical choice to step back. Peter Thiel’s contrarian framework (deeply rooted in René Girard’s mimetic theory) treats crowded market consensus as a clear signal to look elsewhere. So, while 17 other mega-funds rushed to down-market into Seed, Founders Fund did the exact opposite. They pivoted toward massive, concentrated, late-stage bets, pumping capital into generational outliers like OpenAI, Databricks, and Anduril.
Greylock, on the other hand, remains deeply committed to its “first-check” heritage – but they choose to play a high-concentration game. Instead of building an assembly-line deal machine, they focus on fewer, higher-conviction bets, sometimes going as far as incubating companies “directly within their own offices”.
As for NEA, their massive multi-stage mandate makes their seed fluctuations harder to isolate, and we prefer not to speculate without hard data.
Core Allocation vs. Side Projects
We have already looked at the absolute numbers: how many early-stage deals these mega-funds close per year and how that velocity shifted from era to era. However, absolute numbers fail to answer one critical question: is Seed a mere side activity for these giants, or is it their core strategy?
Here is the thing: a fund can close 30 Seed deals a year, but if it simultaneously executes 200 deals across Series A through D, Seed represents a mere 15% of its business. Conversely, if those same 30 deals come out of 60 total investments, Seed captures 50% of the entire firm’s deal architecture.
And this difference is fundamental. A 15% share points to a scout program, pet projects of individual partners, or cheap optionality and a 50% share, however, signals a strategic mandate: dedicated teams, a institutionalized process, and a massive deployment machine.
This is why our third (and perhaps most revealing) lens tracks the exact percentage of total deal activity each mega-fund directs into the early-stage ecosystem:
For 16 out of the 20 funds, the AI era represents an all-time high in early-stage allocation. For the vast majority of our cohort, the AI era captured the maximum percentage of early-stage portfolio activity across all three timelines. In the SaaS era, a typical mega-fund allocated 20-30% of its deal volume to Seed. In the AI era, that baseline has spiked to 35-50%.
Three cases stand out as particularly compelling:
Sequoia: A complete overhaul. This is the most dramatic strategic pivot in our entire dataset. In the SaaS era, less than a fifth of Sequoia’s portfolio touched the early stage – it was predominantly a Series A/B+ powerhouse making sporadic, tactical Seed bets. In the AI era, nearly half of all their deals are at the early stage with a 30-percentage-point surge.
General Catalyst: the V-shaped curve. During the SaaS era, GC was already relatively early-heavy at 38%. In the ZIRP era, that share dipped to 30% as GC, like many of its peers, chased growth-stage yield fueled by free capital. But the AI era triggered a sharp reversal to 47%. This is a deliberate, aggressive return to early-stage investing, peaking higher than ever before.
a16z: a stable baseline followed by an AI surge. a16z is unique because its early-stage allocation remained perfectly flat across both the SaaS and ZIRP eras at 31.2%. While other funds moved down-market erratically during ZIRP, a16z maintained its structural balance. Then came the AI era, forcing a massive jump to 42.5%.
This breakdown matters because LPs routinely hear a familiar narrative from mega-funds: “We occasionally write a Seed check when we encounter an extraordinary founding team.” The data proves this narrative is dead.
For Sequoia, Seed represents 49% of total deals. For GC, it’s 47%. For a16z, it’s 42%. It proves that mega-funds have re-oriented their core engines toward Seed and weaponized this shift with dedicated teams, bespoke internal tracks, and proprietary accelerator programs like a16z Speedrun and Sequoia Arc.
For an emerging seed manager, this provides vital, sobering context. Your daily competition has expanded far beyond neighboring $50M boutique funds. Today, you are fighting for allocation against $10B-$90B AUM behemoths that have redirected 40-50% of their institutional deal machines into your exact stage.
However, to truly understand the mechanics of this pressure, we must layer in one more critical metric: check and round sizes. We break that down below.
Traditional Seed vs. Mega-Seed
One of the defining themes we highlighted earlier is the fragmentation of the Seed stage, which has effectively split into two entirely new sub-classes. The best way to visualize this fracture is to look at the median round size across each era and benchmark it against the overall US “Seed Index” (the market-wide median).
The data is telling: in the AI era, the median US Seed round with a mega-fund on the cap table sits at $6.2M.
Meanwhile, the broader market median is just $1.4M. That is a massive 4.4x gap.
Mega-funds simply don’t participate in the “average” seed round – they systematically operate in the market’s top quartile.
What’s even more striking is how stable this delta has remained across all three macro cycles: 4.8x in the SaaS era, 4.5x during ZIRP, and 4.3x in the AI era. This proves that mega-funds aren’t actually accelerating inflation relative to the rest of the market – they’ve just always existed in a completely different price tier.
To put it in perspective, the market’s 75th percentile ($4.0M) serves as the baseline entry point for mega-funds. Because their median round ($6.2M) sits comfortably above the P75 of the entire US Seed ecosystem, these giants are, by definition, constrained to the top 25% of deals by size.
But things get significantly more interesting when we stack these medians against the averages.
While the median reflects a fund’s “typical” deal, the average is heavily skewed by outliers. The spread between the two serves as a clear proxy for how bimodal a fund’s strategy actually is: does it run a dual-engine model that aggressively dips into mega-seeds, or does it uniformly operate within a single price tier?
When viewed through this lens, the cohort cleanly fractures into two distinct classes.
Class 1: “Bimodal” (3x+ spread)
Index (Median: $8.2M, Average: $34.3M – 4.2x spread)
Lux ($6.0M vs $31.7M – 5.3x)
Lightspeed ($6.8M vs $30.8M – 4.5x)
Accel ($5.0M vs $26.0M – 5.2x)
a16z ($6.0M vs $21.8M – 3.6x)
Sequoia ($5.0M vs $17.4M – 3.5x)
These funds play on two tables simultaneously: volume-based classic Seed rounds at $5-8M, and highly selective mega-seeds at $50-500M+ that pull their statistical averages into the stratosphere. Their TechCrunch headlines (“$100M Seed Round!”) don’t reflect daily reality because their typical deal is actually 4 to 5 times smaller.
Class 2: “Homogeneous” (<2.5x spread)
Greylock ($6.9M vs $13.3M – 1.9x)
Founders Fund ($7.0M vs $12.0M – 1.7x)
CRV ($7.5M vs $10.8M – 1.4x)
8VC ($6.6M vs $8.7M – 1.3x)
NEA ($7.0M vs $7.4M – 1.1x)
For this cohort, the median and average track closely, indicating the absence of a long tail of mega-rounds. They consistently deploy within a uniform
$5-8M price bracket without major outliers.
Bimodal funds capture the headlines, fueling the illusion that Seed has mutated into a $30M+ game. But the data refutes this: a typical deal even for the most bimodal firms remains firmly in the $5-8M range. Mega-seeds are merely the long tail of the distribution and not the center.
For emerging managers, the real competitive pressure stems from the homogeneous cohort – GC, Khosla, Bessemer, Greylock. These firms systematically execute in the $5-8M segment without being distracted by mega-seeds. Bimodal funds are more formidable in headlines but less dangerous in day-to-day competition. They spend part of their time in the mega-seed market, where an emerging manager wouldn’t compete anyway.
Ultimately, the fragmentation of the Seed market has very little to do with abstract round inflation. We are witnessing the birth of two entirely separate ecosystems masquerading under a single label: mega-seed ($20M+) for the bimodal platforms, and traditional seed ($3-8M), where mega-funds and emerging managers still collide. The only difference today is that the sheer volume of multi-stage giants crowding this classic segment has multiplied.
Who Shapes the Terms and Who Rides Along?
Yes, we have established that mega-funds maintain a systematic presence in Seed rounds – both in the upper quartile and within the $5-8M bracket, the exact territory where emerging managers operate. However, presence and leadership are two fundamentally different things.
A fund participating in a $6M round with a minor $500K check is merely a co-investor, a passenger on the cap table. The fund that leads that exact same round is the one driving the valuation, shaping the term sheet, and determining who else gets into the syndicate. It is the lead investor who ultimately decides whether there is room left for an emerging manager.
Hence, the next logical question arises: out of all these Seed deals executed by mega-funds, what percentage are they actually leading?
To unpack this, I would break these firms down into four distinct clusters:
"Conviction Leaders" – High Lead Rate + High Volume
Khosla (60%, 19 leads/yr)
Lightspeed (63%, 21 leads/yr)
Accel (54%, 20 leads/yr)
This is the most dangerous cohort for emerging managers. They deploy aggressively and they demand the driver's seat. When Lightspeed leads 21 Seed rounds a year with a 63% lead rate, they are systematically dictating early-stage pricing. If an emerging manager is competing for the same company – they are fighting for the right to lead.
“Volume Players” – High Volume, Moderate Lead Rate
a16z (51%, 40 leads/year)
General Catalyst (53%, 33 leads/year)
Sequoia (36%, 19 leads/year)
These giants command a massive absolute volume of leads, even if their percentage-wise lead rate is lower. They lead the absolute best companies in their pipeline and take passive positions in the rest. For an emerging manager, this presents a double threat: even when a mega-fund doesn’t lead, their presence on the cap table heavily impacts signaling and subsequent follow-on dynamics.
“Selective Leaders” – High Lead Rate, Low Volume
EQT (82%, 7 leads/year)
Craft Ventures (76%, 8 leads/year)
Index Ventures (67%, 12 leads/year)
Founders Fund (61%, 10 leads/year)
Greylock (58%, 6 leads/year)
These funds lead the vast majority of their deals, but they maintain a highly disciplined, low-velocity pace. This is a pure conviction-driven approach: if they write a check, they almost always want to run the round. While they pose less of a threat in terms of pure market volume, they will almost certainly secure the lead position in any specific deal they enter.
“Network Players” – Low Lead Rate
8VC (38%, 9 leads/year)
Amplify (39%, 4 leads/year)
Sequoia (36%, 19 leads/year)
Bessemer (44%, 9 leads/year)
These firms choose to participate far more often than they lead. Their role at the Seed stage is centered around network, signaling, and buying optionality, rather than establishing market pricing. For an emerging manager, this is the least threatening type of mega-fund, as they rarely crowd out the lead position.
What’s equally fascinating is that two of the largest funds by absolute early-stage activity post the lowest lead rates in the AI era: a16z at 51% and Sequoia at 36%. Furthermore, both have seen their lead rates drop since the SaaS era (a16z down from 67%, Sequoia from 52%).
While this might seem counterintuitive at first glance, the explanation is simple: when you are executing 77 deals a year or 51 deals a year, it becomes physically impossible to lead every single one of them. Consequently, a portion of these deals naturally shifts toward scout bets, co-investments, and syndicate positions led by someone else. In this tier, volume and lead rate represent a clear trade-off.
Yet, in absolute numbers, they still dominate the field: a16z leads ~40 early-stage deals per year, while General Catalyst leads ~33. That is more than the total early-stage deal flow of half the funds on our list combined.
On balance, lead rates have trended upward for the majority of players in the AI era. In fact, 13 out of the 20 funds analyzed show a higher lead rate in the AI era than they did during the SaaS era:
This means that mega-funds are leading more often. For example, during the SaaS era, Greylock led only one out of every four Seed deals, but in the AI era, they lead more than half. They have fundamentally pivoted from a passive “participate when invited” approach to an active “we structure the round” playbook.
LPs must keep this reality in mind during fund due diligence. Of course, emerging managers love to splash mega-fund logos across their fundraising decks alongside phrases like “we co-invest with,” but this dynamic can actually serve as a vital signal defining the exact venture product an LP is underwriting.
If an LP asks, “How many rounds did you lead last year, and in how many of those was the other lead a mega-fund?” and the answer is, “we frequently co-invest alongside a16z or General Catalyst,” then this isn’t a structural edge. Instead, it signals a heavy dependence on mega-fund deal flow to build a diversified portfolio. While this is not inherently a bad strategy, the underlying fund math changes drastically once you factor in larger round sizes, inflated valuations, and diluted ownership targets due to a lack of pricing power and ability to lead.
Conversely, if the answer is, “We lead the exact rounds mega-funds don’t touch or we get there long before they do,” – that is where a true, defensible edge for emerging managers actually lives.
Where the Pressure is Highest
Everything we’ve established so far regarding deal dynamics, round inflation, and lead rates describes mega-funds in the aggregate. In reality, however, an emerging manager rarely invests in “Seed as a whole.” Instead, they back specific verticals, which often serves as their primary edge. This is why our next logical question is where exactly they are going.
When viewed through this lens, the footprint is significantly more concentrated than aggregate statistics imply.
Unsurprisingly, Enterprise AI & Automation and AI Infra & DevTools dominate both lead rates and total deal counts. Together, these two sectors account for 538 companies or 42% of all early-stage activity across our entire dataset. Remarkably, all 20 mega-funds are simultaneously active in both verticals. There are three core drivers behind this concentration:
First: market scale. Enterprise spending on Generative AI skyrocketed from $1.7B in 2023 to $37B in 2025 – a staggering 20x-plus surge in just two years. Enterprise AI has already captured 6% of the global SaaS market, scaling faster than any software category in history. For mega-funds, this breakneck growth is a fundamental signal of generational value creation.
Second: Velocity. The temporal dynamics of the AI era are entirely unprecedented. While the SaaS era was defined by the T2D3 playbook (triple, triple, double, double, double), top AI-native companies are growing under a Q2T3 framework (quadruple, quadruple, triple, triple, triple). For a fund, this means the entry window at the Seed stage slams shut much faster. Hesitating for 12 to 18 months can mean missing out on an entire software category.
Third: performance outliers. Lovable reached $100M ARR in just 8 months and doubled that figure to $200M a mere four months later outpacing OpenAI, Cursor, and every other software company in history. By May 2026, Sacra estimated Lovable’s annualized revenue run-rate had already breached $500M. Meanwhile, Cursor raised $2.3B at a $29.3B valuation, and Anthropic’s run-rate revenue accelerated from about $1B at the end of 2024 to $14B in February 2026, $30B in April, and $47B by May 2026, while the company raised $65B at a $965B valuation. All of these companies were either non-existent or completely obscure just three years ago.
For an emerging manager investing in AI, this creates a landscape where practically every mega-fund is hunting in their exact backyard. Armed with infinite capital, these giants are unconstrained by round pricing, allowing them to aggressively lead rounds and maximize ownership targets. So the survival of new fund managers depends on deep domain expertise, proprietary access to high-density founder networks, and the ability to back founders at the pre-pitch-deck stage.
There is another critical nuance: the fastest-growing AI companies (the so-called “AI Supernovas”) operate at an average of just 25% gross margins, deliberately sacrificing unit economics to capture market share. Even the more traditional “Shooting Stars” average around 60% gross margins, which remains well below the classic SaaS benchmark of 70-85%.
This indicates that Enterprise AI currently functions as a sector where top-line revenue scales far ahead of profitability and investors are effectively underwriting future economics rather than current margins. Mega-funds, with their deep pockets and long time horizons, can easily absorb this structural bet, so an emerging manager running a $25M-$75M vehicle finds themselves in a fundamentally vulnerable position if those future unit economics take longer to materialize than the market anticipates.
AI Infra & DevTools deserves special attention regarding round structures. This is where the bimodal behavior we observed at the fund level manifests most acutely: the median round stands at $6.8M, while the average skyrockets to $48M – a massive 7x spread.
This massive delta indicates that the sector is heavily populated by $100M+ mega-seeds that drag the statistical average upward. This is precisely the breeding ground for the “$50M Seed round” headlines that give casual observers a highly distorted perception of a typical deal in the space.
By comparison, the spread drops to just 1.4x in Commerce & GTM and 2.0x in Healthcare. The further you move from the AI core, the more homogeneous the round landscape becomes.
Two sectors stand out for behaving atypically relative to their actual size:
Cybersecurity: Despite a relatively small footprint of 76 companies, the sector posts a 62% lead rate – the highest among major verticals. Paired with a $7M median round (one of the highest in our dataset), it’s clear that mega-funds that they dominate here, running the show in nearly two-thirds of all deals.
Defense & Aerospace: With an even smaller footprint of 34 companies, this sector boasts a record-breaking 66% lead rate. However, only 12 out of the 20 mega-funds are active here, signaling highly concentrated, conviction-driven bets by a handful of specific players rather than systematic, platform-wide pressure.
Finally, there are sectors that remain relatively uncrowded: Climate & Energy (26 companies, 12 active funds), Logistics (24 companies, 13 active funds), and other traditional verticals like PropTech, EdTech, Legal, and HR.
An emerging manager with deep domain expertise in these verticals completely escapes the platform crunch. Instead of wrestling with 20 massive platforms, they are going up against 8 to 12 firms pricing maybe 2 or 3 deals a year, which is an entirely different game.
All of this drives an important practical takeaway for LPs: the correct due diligence question for an emerging manager must pivot toward the specific verticals they play in, as sector selection dictates the very nature of the competition they face and, consequently, the exact type of differentiation required to win.
Are Mega-Fund Seeds Worth the Premium?
Throughout this entire research, we’ve been showcasing only one side of the coin: mega-funds have invaded Seed, they are closing more deals, leading more often, and operating right in the price territory of emerging managers.
But there is a question we have intentionally pushed aside until now and it is arguably the most critical question of this entire study: does it actually work?
Yes, mega-funds write larger checks, participate in rounds that are 4.4x larger than the market median, direct 40-50% of their deal activity toward the early stages, and lead more than half of their Seed deals. However, if the companies they back at Seed don’t boast higher survival rates than the market average, then everything we’ve mapped out is just valuation inflation stripped of actual value.
Conversely, if mega-fund-backed Seed companies convert to Series B at a significantly higher rate than the broader market, it completely flips the narrative. In that scenario, mega-funds aren’t merely “taking over Seed” – they are actually making Seed better. This then forces a logical question for LPs: “Why not consolidate capital into mega-funds that cover Seed and then double down on later stages, effectively capturing the entire market lifecycle within one firm?”
To answer this, we calculated one straightforward metric: what percentage of companies raising a Seed round in a specific era subsequently reached Series B? We benchmarked two distinct groups: the broader Seed market versus Seed companies that had at least one of our 20 mega-funds on the cap table.
We focused on two macro cycles that have had enough maturity for companies to scale to Series B: the SaaS era (2015-2019) and the ZIRP era (2020-2022). The AI era is still too nascent to measure, as most companies are simply too young. The results turned out to be definitive, yet nuanced.
In the SaaS era: Out of 60,110 companies that raised a Seed round, 9.8% reached Series B. For the 940 companies with mega-fund participation at Seed, that number jumped to 36.7% – a 3.7x delta.
In the ZIRP era: The trend remained identical: the broader market sat at 3.9%, while mega-funds hit 16.5% – widening the gap to 4.2x.
Mega-funds convert Seed rounds into Series B 3.7x to 4.2x better than the broader market. More importantly, this gap is expanding. In a hyper-inflated ZIRP market where conversion rates plummeted across the board, the quality filter of a mega-fund proved to be more valuable, not less. At first glance, this is compelling evidence of superiority.
But before jumping to a definitive conclusion, we need to unpack why this conversion is so much higher. There are several structural drivers at play, which can collectively be attributed to a powerful signaling effect:
Elite Series A deal flow: Top-tier Series A investors actively seek to co-invest alongside an institutional, heavy-hitting Seed lead.
Internal follow-on capacity: A mega-fund has the deep pockets to internally lead the Series A or B for its own breakout Seed portfolio companies.
Brand-driven talent acquisition: Elite engineers see a “backed by Sequoia” or “a16z” badge, which significantly lowers the friction for hiring.
Outsized media distribution: More PR leverage leads to increased inbound interest from potential enterprise clients.
Therefore, it is vital to recognize that a significant portion of this conversion rate is not a result of a mega-fund “picking correctly,” but rather the mega-fund helping the company become the correct choice. For LPs, this is a clear signal that a mega-fund’s value-add at Seed isn’t limited to mere “stock picking” – it functions as a true “platform as a product.”
However, there is a flip side. When we look past the aggregate metrics and examine each firm individually, a troubling pattern emerges: out of the 15 funds with a statistically significant sample size (10+ Seed deals per era), 14 saw their conversion rates plummet from the SaaS era to ZIRP. We are talking about sharp drops ranging from 10 to 25 percentage points:
The correlation is direct: the funds that scaled their deal volume the most during ZIRP suffered the steepest declines in conversion. For instance, Sequoia tripled its deal count (from 20 to ~50/year), and its conversion crashed from 46% to 14%. Lightspeed quadrupled its volume (from 12 to 42/year), and its conversion sank from 31% to 11%.
The sole exception to this rule is Greylock, whose conversion actually jumped from 29% to 44%. This was no fluke; Greylock was the only firm that kept its deal volume virtually flat during ZIRP (moving slightly from 11.0 to 11.3 deals/year). Fewer deals yielded a higher hit rate. Volume discipline directly equals portfolio quality.
This conversion data simultaneously validates and complicates our entire narrative.
On one hand, it proves that mega-funds do move the needle at the Seed stage. A 3.7x conversion premium is neither a fluke nor a data artifact. Companies backed by mega-funds early on genuinely survive and scale better than the broader market. For LPs, this is a powerful argument: brand, network, and platform resources drive measurable value.
On the other hand, volume and quality exist in constant tension. Today, in the AI era, mega-funds are hitting record-breaking Seed deal volumes. If the ZIRP pattern repeats itself, their conversion rates will inevitably erode. The only question is by how much. Will the AI-era platform effects and signaling advantages of these giants be enough to offset the dilution caused by their massive deployment pace?
We will have a definitive answer in 3 to 5 years. But historical data offers a sobering warning: mega-funds have proven they can pick winners at low volumes. They have yet to prove they can do it at scale.
And it is precisely within this gap (the space between a proven past and an unverified present) where the real opportunity lies for emerging managers who are ready to do less, but do it better.
Danger Index
To wrap things up, we decided to do something that is admittedly controversial and built on a set of highly generalized assumptions. Yet, we felt it was critical to share it with the emerging managers and LPs who read this blog: The Danger Index.
Essentially, this is a data-driven ranking of the specific mega-funds that pose a genuine competitive threat to emerging managers based on our findings. We anchored this index on three core pillars:
Volume: The absolute number of early-stage deals a fund executes per year in the AI era. The higher the volume, the more frequently an emerging manager will run into them on the ground.
Strategic Commitment: The percentage of a firm’s total portfolio activity allocated to the early stage. If Seed commands 45% of their deal activity, it is a core strategy backed by dedicated teams and institutionalized processes. If it is sitting at 20%, it is a side project – an optionality bet that the fund could easily scale back to re-focus on later stages.
Price Overlap: The median size of the rounds the fund participates in. This is arguably the most critical factor. A mega-fund playing in $8-10M rounds is mostly competing with other multi-stage behemoths. However, a mega-fund operating within the $4-5M bracket is competing directly with emerging managers, as this is the exact pricing Sweet Spot where a typical $50-100M Seed fund deploys its capital.
We assigned a normalized score from 0 to 10 for each individual factor. The final Danger Score is the sum of these three components, with a maximum possible score of 30.
The results proved counterintuitive. Four powerhouse firms landed in Tier 1 (Maximum Danger): General Catalyst, a16z, Sequoia, and Accel.
All four simultaneously close 37 to 83 early-stage deals per year, allocate 39% to 50% of their total portfolio activity to Seed, and operate within the $4.4M–$5.4M round bracket – landing right in the crosshairs of emerging manager territory.
Counterintuitively, General Catalyst snatched the #1 spot over a16z, even though a16z clocks a higher absolute deal volume (83 vs. 65). The differentiator is that GC perfectly synchronizes all three risk vectors: high velocity, the highest early-stage allocation in the tier (48%), and a median round size of $5.0M – smack in the center of the EM pricing sweet spot. By comparison, a16z plays slightly upstream ($5.4M median) with a slightly lower early-stage concentration (43%). The delta is subtle, but statistically meaningful.
Sequoia’s third-place finish was another surprise. They post the lowest lead rate among the top 5 funds by volume (36%), meaning they act as a co-investor far more often than they run the round. However, their median round size sits at $4.6M – the lowest among the absolute largest platforms. They systematically buying into cheaper rounds by mega-fund standards.
Conversely, Index Ventures landed unexpectedly low in Tier 3, despite maintaining 19 deals per year and a formidable 66% lead rate. The reason? A steep $8.4M median round size. Index plays entirely above the traditional emerging manager zone.
The exact same structural logic applies to Founders Fund ($7.8M median) and Greylock ($7.0M median) – both sit comfortably in Tier 3. They maintain an undeniable early-stage footprint, but they don’t crowd the specific price ecosystem where most emerging managers fight for survival.
Yes, the Danger Index is not a death sentence for emerging managers. We look at it as a map of the minefield.
And it distills this entire macro research down to one practical, high-stakes question: “Which Tier 1 platforms are hunting in your exact price bracket and vertical?”
If your answer is, “GC and a16z, both backing AI software, both entering at $4-6M rounds,” then as an emerging manager, you must clearly articulate to your LPs what specific edge allows you to win against two institutions closing a combined 150 Seed deals a year in your backyard.
But if the answer is, “None of the Tier 1 giants, I lead $2–3M rounds in Climate Tech,” that is an entirely different conversation. The Danger Index reveals that institutional pressure in that segment is structurally lower, meaning deep domain expertise can still serve as a highly defensible edge on its own.
Key Takeways
The average mega-fund went from 10.6 early-stage deals per year in the SaaS era to 23.9 in the AI era. Only 3 out of 20 funds scaled back. The shift is structural, not cyclical.
Seed valuations have bifurcated sharply. The 90th percentile hit $93.7M in Q1 2026 – nearly 2x compared to four years prior. The 25th percentile moved from $18M to $22.7M over the same period.
The median round with a mega-fund on the cap table sits at $6.2M in the AI era, versus $1.4M for the broader market – a 4.4x gap that has held stable across all three eras.
For 16 out of 20 mega-funds, the AI era represents an all-time high in early-stage allocation. The typical mega-fund went from directing 20–30% of its deal volume to Seed in the SaaS era, to 35–50% today.
13 out of 20 mega-funds now lead more Seed rounds than they did in the SaaS era. Greylock went from a 24% to a 58% lead rate. The passive co-investor posture is being replaced by a structured, round-leading playbook.
42% of all mega-fund early-stage activity concentrates in just two sectors: Enterprise AI & Automation and AI Infra & DevTools. All 20 funds are simultaneously active in both verticals.
Mega-fund-backed Seed companies reach Series B at 3.7x to 4.2x the rate of the broader market. But 14 out of 15 funds with sufficient sample size saw their conversion rates drop sharply from SaaS to ZIRP – with the steepest declines correlating directly with the largest volume increases.
Greylock kept its deal count flat during ZIRP and was the only fund whose conversion rate actually improved. Volume discipline equals portfolio quality.
The Danger Index places General Catalyst, a16z, Sequoia, and Accel in Tier 1 – the only four funds that simultaneously combine high velocity, 39–50% early-stage portfolio concentration, and a median round size below $5.5M, landing directly in the emerging manager pricing sweet spot.
Climate & Energy, Logistics, and traditional verticals like PropTech and EdTech remain structurally undercrowded – with only 8 to 13 mega-funds active versus 20 in AI, and lead rates well below the category average.
Conclusion
All of this drives a final, structural perspective for both emerging managers and the LPs backing them: the invasion of mega-funds into the early-stage landscape is not a temporary anomaly of a tech cycle, but a permanent recalibration of how venture capital operates at its foundations.
As multi-stage behemoths continue to weaponize their billions to absorb the top quartile of the Seed ecosystem, trying to beat them at their own high-velocity, deep-pocketed game is a mathematical dead end. However, the data reveals a critical crack in their seemingly flawless armor – the inescapable tension between massive deployment volume and portfolio conversion quality.
Ultimately, this shifts the definition of what a true early-stage edge looks like in the AI era. Winning as an emerging manager no longer means striving to scale an institutional deal machine or blindly chasing hyped categories where Tier 1 platforms dictate the pricing rules. Instead, it requires the rigorous discipline of sector selection, the patience to underwrite complex future unit economics that mega-funds often overlook, and the courage to remain small, focused, and deeply integrated with founders long before the multi-stage platforms even notice they exist.
In a venture ecosystem that increasingly values pure scale, the ultimate counter-strategy for emerging managers is mastering the premium of absolute discipline and not matching the sheer volume of the giants.
I appreciate the effort you put into every piece of content! In my humble opinion, it's absolutely worth sharing.
Great piece, Pavel. And I think a lot of this has to do with the fact that there is so much value to be captured within AI for seed-stage startups specifically, considering how novel a lot of these technologies are.