Greylock Caps Fund 18 at $1.5B, Betting Against the Megafund
TL;DR
Greylock has closed Greylock 18 at $1.5 billion, a 50% step-up from the $1 billion vehicle it raised in 2023. The headline is not the number — it is the ceiling. Partner Saam Motamedi says the firm could have raised a multiple of that figure and chose not to, capping the fund to preserve a deliberately small portfolio of roughly 25 companies. In a year when US venture deployment has shattered records and megafunds have become the default competitive response, one of the oldest firms in the business just made the opposite bet in public.
Key Takeaways
Capping a fund is now a strategy statement, not modesty. Greylock's ten partners each make one or two new investments a year. That arithmetic — not LP appetite — sets fund size. By working backwards from portfolio construction to dollars raised, Greylock is inverting the industry's default logic, in which the fund is sized to what the market will bear and the strategy is retrofitted afterward.
The incubation franchise is the real moat. Palo Alto Networks was started inside Greylock's offices 21 years ago. Abnormal was incubated in 2018 and last valued at $5.1 billion. This is a company-formation practice, not a check-writing practice, and it is the one thing a $10 billion fund cannot buy more of. Formation capacity scales with partner hours, which is precisely why the fund has a cap.
The 15% later-stage allocation is the honest part of the pitch. Motamedi estimates roughly 15% of Greylock 18 will go to later-stage companies the firm missed early. Fund 17 did exactly this with Anthropic, Revolut, and Wiz — and the Anthropic investment, made at a $183 billion Series F valuation, is the largest in the firm's history. Every early-stage firm now runs a shadow crossover book. Greylock is unusual mainly in disclosing the percentage.
Restraint is a recruiting pitch aimed at LPs and founders simultaneously. "The most important partner to the most important entrepreneurs" is a line that only survives contact with reality if the denominator stays small. Greylock is betting that in a market awash in undifferentiated capital, scarcity of partner attention is the product founders will actually pay for in dilution.
Fund Overview
Fund Name: Greylock 18
Fund Size: $1.5 billion (up 50% from the $1 billion 2023 vehicle)
Stage: Incubation, seed, and Series A, with roughly 15% reserved for later-stage entries
Check Size: Not disclosed; the firm leads seed and Series A rounds
Geography: Primarily US, with the firm's centre of gravity in Silicon Valley
Focus: AI-native companies spanning infrastructure, cybersecurity, applications, deep tech, consumer, and fintech
Key LPs: Not disclosed
Why This Fund Matters
Fund size has quietly become the loudest signal in venture, and almost nobody has been willing to send the contrarian one. The logic of the last three years has been mechanical: AI rounds got bigger, ownership targets required bigger checks, bigger checks required bigger funds, and bigger funds required a management-fee-friendly story about "platform." Greylock, at 61 years old and with nothing to prove, has just declined to run that loop — and said out loud that it could have.
That matters because it reframes the strategic question every GP is currently avoiding. The industry has been treating fund growth as a proxy for franchise strength. Greylock is arguing that beyond a certain point, fund growth is a tax on the thing that generated the returns in the first place. If your edge is starting companies from a blank page, and that edge is bounded by how many blank pages ten people can credibly stare at in a year, then every dollar raised past that bound is dilution of the edge dressed up as growth.
The uncomfortable subtext for LPs is that the two models are not really competing for the same outcome. A $5 billion early-stage fund is, in practice, an index on AI with a venture fee structure. A $1.5 billion fund making 25 investments is a concentrated bet that specific people, identified before their companies exist, will build the winners. LPs have spent three years buying the first thing while telling themselves they were buying the second. Greylock 18 forces the distinction into the open.
There is also a market-timing read here that deserves to be said plainly. Capping a fund at the top of a cycle is a much cheaper decision than capping one at the bottom. Restraint announced in a year of record deployment costs Greylock some AUM and buys it enormous positioning if the AI funding environment normalises. That is not cynicism — it is good sequencing. But it should temper any reading of this as pure principle.
The Team
Greylock's partnership numbers ten investing partners, a figure the firm treats as the binding constraint on portfolio size rather than a headcount to be grown. Saam Motamedi, who led the firm's public framing of this raise, has been the most visible voice on Greylock's AI and security thesis and sits at the centre of both the Abnormal incubation lineage and the firm's newer infrastructure bets, including its Series A in Baseten in 2022 — a company now valued at $13 billion.
The structural detail that says the most about how the firm works: partners meet every Monday to review a pipeline whose agenda, per Motamedi, consists mostly of people's names rather than company names. That is a firm that has organised its entire operating cadence around sourcing founders pre-formation. It also explains why the partner count cannot be inflated on demand — relationships of that kind do not transfer, and they do not parallelise.
Early Portfolio
Greylock 18 is newly closed and its investments have not been disclosed. The lineage it is built on is well documented: Palo Alto Networks and Abnormal Security on the incubation side, Baseten from a 2022 Series A, and Fund 17's later-stage entries into Anthropic, Revolut, and Wiz.
What This Means for Founders
If you are pre-company — a strong operator with a thesis and no entity — Greylock is now unusually well capitalised to fund exactly that, and structurally motivated to spend real partner time on you. The Monday pipeline of names is not marketing. Founders who want a firm that will help recruit the first ten engineers and open the first ten enterprise doors are the intended customer, and a 25-company portfolio means that promise is arithmetically credible in a way it is not at a firm writing 120 checks a fund.
The flip side is brutal selectivity. Roughly 25 slots across the life of a $1.5 billion fund, sourced disproportionately from people the partnership already knows, is a narrow door. If you are not in the network and not being incubated, the realistic path is a later-stage entry into that 15% bucket — which means the fastest way to get Greylock on your cap table may be to become undeniable somewhere else first.
Fund Momentum Take
This is the most interesting fund announcement of the month, and the number is the least interesting part of it. Greylock has correctly identified that the scarce input in venture is not capital — it has not been capital for years — but the ability to be genuinely useful to a small number of exceptional people. Sizing the fund to that constraint rather than to LP demand is the right answer, and it is an answer almost no firm with Greylock's brand has been willing to give, because turning down money is career risk that only a 61-year-old franchise can absorb.
The risk is that discipline and adverse selection look identical from the outside until the returns land. Twenty-five investments is a concentrated book, and concentration cuts both ways: it produces the outlier returns the model is designed for, but it also means two or three misjudged founder bets meaningfully impair the fund. Greylock is also explicitly hedging with a later-stage bucket that includes a record-setting Anthropic position at $183 billion — a mark that requires an extraordinary outcome to be a good venture return. If that 15% bucket quietly grows, the restraint story unravels fast.
Our bet: this ages well, and not primarily because of the returns. Greylock has handed every LP in the market a new question to ask their GPs — "what portfolio construction does your fund size imply, and did you work forwards or backwards?" That question is going to be uncomfortable for a lot of firms that raised on momentum. The megafund era will not end because someone wrote an essay about it. It will end because allocators start doing that arithmetic, and Greylock just showed them how.
Frequently Asked Questions
How big is Greylock 18 and how does it compare to the last fund?
Greylock 18 closed at $1.5 billion, 50% larger than the $1 billion fund the firm raised in 2023. Per TechCrunch's reporting, it roughly matches what Greylock raised across its seed and flagship funds during the pandemic period.
Why did Greylock cap the fund?
Partner Saam Motamedi told TechCrunch the firm could have raised a multiple of $1.5 billion but sized it to its portfolio construction instead. Ten partners making one to two new investments each per year implies roughly 25 companies, and the firm argues it can only be deeply useful at that scale.
What stage does Greylock 18 invest at?
Primarily incubation, seed, and Series A. Motamedi estimates roughly 15% of the fund will go into later-stage companies the firm missed early, following the pattern of Fund 17's entries into Anthropic, Revolut, and Wiz.
What is Greylock's incubation track record?
Palo Alto Networks launched inside Greylock's offices 21 years ago, and the firm incubated Abnormal Security in 2018 — a company last publicly valued at $5.1 billion. Company formation from scratch is the firm's signature practice.
What was Greylock's Anthropic investment?
Greylock first invested in Anthropic when the company raised its Series F at a $183 billion valuation. Motamedi described it as the largest investment in the firm's history. It sat in Fund 17, not Greylock 18.
Have a fund closing to announce? Submit your fund here.
Need help raising capital? Check out our Fundraising Advisory services.