Khosla Ventures Targets Record $5.5B Fund, Doubling Down on Early AI
TL;DR
Khosla Ventures is reportedly in talks to raise up to $5.5 billion across a new family of funds, which would be the largest capital haul in the firm's 20-year history. The plan, first reported by Bloomberg on July 23, splits into roughly $1 billion for seed, $2 billion for early-stage venture, and a $2.5 billion opportunity fund for later-stage bets. More than half the money is pointed at the earliest, riskiest end of company-building, an unusually early-heavy posture at a moment when most megafunds are chasing late-stage AI. The numbers are a target, not a close, and could still move.
Key Takeaways
This is a target, not a final close. Every reliable account traces back to a Bloomberg report citing people familiar with the talks, all of whom flagged that the figures could shift as fundraising continues. Treat $5.5 billion as an aspiration Khosla is marketing to LPs, not a wired-in number. The distinction matters for anyone modeling deployment pace off it.
The early-stage tilt is the real story. Roughly $3 billion of the $5.5 billion target is earmarked for seed and early-stage, versus $2.5 billion for the opportunity fund. In a cycle where Founders Fund, Sequoia, and others are raising multibillion-dollar vehicles largely to buy more of OpenAI and Anthropic, Khosla is signaling it would rather fund the next OpenAI than average up on the current one.
It is a sharp step up on the last raise. The firm was reportedly targeting $3.5 billion across three funds in early 2025 and closed around $4 billion in that vintage. A $5.5 billion target is close to a 40% increase, arriving only about 17 months later, which tells you how quickly AI has re-rated the price of staying top-tier.
The deal flow already justifies a bigger early-stage pool. Khosla's recent pace, including leading Norm Ai's $120 million round, General Intuition's $320 million Series A, and Factory's $150 million Series C, shows a firm writing larger early cheques into applied and physical AI. The fund size is catching up to a check-writing cadence that is already in motion.
Fund Overview
Fund Name: New Khosla Ventures fund family (seed, early-stage/venture, and opportunity funds)
Fund Size: Up to $5.5 billion targeted (reported): ~$1B seed, ~$2B early-stage, ~$2.5B opportunity
Stage: Seed through growth/opportunity
Check Size: Sub-$1M seed cheques up to nine-figure opportunity investments
Geography: Primarily United States, global mandate
Focus: AI across foundation, applied, and physical layers, plus deep tech, health, climate, defense, and robotics
Key LPs: Not disclosed
Why This Fund Matters
Khosla Ventures is not raising in a vacuum. The last 18 months have produced a run of enormous AI-era vehicles: Founders Fund closed a $6 billion growth fund, Sequoia raised roughly $7 billion for an expansion fund, Kleiner Perkins closed $3.5 billion across two AI-focused funds, and Andreessen Horowitz gathered more than $15 billion across six funds. Against that backdrop, a $5.5 billion target reads less like an outlier and more like the current going rate for a firm that wants to keep competing for the best AI deals.
What separates Khosla's plan is where the money is pointed. Most of the megafund capital raised in this cycle has been late-stage by design, built to concentrate into a short list of already-anointed foundation model companies. Khosla's proposed split does the opposite: it front-loads seed and early-stage. That is a deliberate bet that the returns that made the firm's name, most famously an early cheque into OpenAI written before ChatGPT existed, come from being early and contrarian rather than from paying up for consensus winners.
There is a defensible thesis underneath the posture. If AI's value is going to spread from a handful of model labs into applied software, regulated industries, robotics, and the physical world, then the next generation of category-defining companies has not been funded yet, and a lot of it will be seeded rather than bought at Series D. A $3 billion early-stage war chest is a way to buy a very large number of lottery tickets in exactly that scenario.
The risk is equally clear. Deploying $3 billion at seed and early-stage without inflating your own entry valuations, or quietly drifting upmarket to put the money to work, is genuinely hard. Seed is a small-cheque business, and scaling it usually means either doing many more deals, writing bigger seed rounds at higher prices, or both. Every large early-stage fund in history has had to fight the gravitational pull toward later, safer, larger cheques. That tension is the thing to watch as this fund actually deploys.
The Team
The firm was founded by Vinod Khosla, the Sun Microsystems co-founder turned investor whose willingness to underwrite technical and scientific risk, from clean energy to fusion to AI, has defined the Khosla Ventures brand for two decades. Khosla's early conviction in OpenAI, as its first outside investor, is the reference point the current raise leans on most heavily.
The pitch to LPs is backed by an active partnership deploying across AI, bio, climate, and hard tech. Managing director Samir Kaul, one of the firm's longtime leaders, has framed the current thesis around trust: AI will not be adopted inside regulated industries until the institutions that oversee those industries trust it, a logic that runs through Khosla's legal-AI, defense, and robotics bets. That framing helps explain why the firm is leaning into applied and physical AI rather than only backing model labs.
Early Portfolio
Khosla's recent cheque-writing is the clearest evidence for why it wants a larger early-stage pool. In July 2026 the firm led a $120 million round for legal-AI startup Norm Ai at a reported $1.2 billion valuation. In June it led General Intuition's $320 million Series A, a company training AI on gameplay data to teach machines to act in the physical world, roughly three months after leading its seed. In April it led Factory's $150 million Series C for AI coding agents, and it has backed defense startup Mach Industries alongside Sequoia. On the AI portfolio more broadly, the firm counts OpenAI, Cognition, Sakana AI, and Physical Intelligence among its bets.
What This Means for Founders
If the raise lands as described, Khosla becomes one of the deepest-pocketed early-stage buyers in AI, with the added advantage of an in-house $2.5 billion opportunity fund to keep following its winners. For technical founders in applied AI, robotics, defense, health, and infrastructure, that means a single firm can plausibly lead your seed and still be writing meaningful cheques at Series C. That continuity is valuable, and it is exactly what a barbell of early plus opportunity capital is built to deliver.
The flip side is founder selection. Khosla has always been a high-conviction, opinionated investor that leans into hard science and long development cycles, and it is comfortable being a demanding board partner. Founders who want a hands-off check should calibrate accordingly. The firms that fit Khosla best are the ones tackling genuinely difficult technical problems where patient, contrarian capital, and a partner willing to fund the unproven, is the actual value-add.
Fund Momentum Take
Our read: this is the most interesting megafund of the cycle precisely because it refuses to behave like one. Anyone can raise billions to buy more OpenAI and Anthropic. Committing more than half of a record fund to seed and early-stage, in a market that is paying record prices for late-stage AI certainty, is a real point of view. If Khosla is right that the next generation of winners is still un-funded, this is the vehicle best positioned to catch them.
The bet is not without hair. Deploying $3 billion at the early stage is an operational challenge as much as an investing one, and the history of large early-stage funds is a history of style drift toward bigger, later cheques. There is also the simple fact that a $5.5 billion target inflates the returns bar: even elite seed performance has to clear a very large denominator. And because this is still a target rather than a close, the final shape could look different from the pitch.
Still, if you are grading on conviction and differentiation rather than raw size, Khosla's plan is the one worth studying. Our bet is that the number moves before it closes, but the early-stage tilt holds, because that tilt is the whole reason the firm thinks it can beat a field of look-alike megafunds.
Frequently Asked Questions
Has Khosla Ventures actually closed a $5.5 billion fund?
No. As of late July 2026 the firm is reportedly in talks to raise up to $5.5 billion, per a Bloomberg report, and sources cautioned the figures could change. It is a target, not a completed close.
How is the $5.5 billion split?
The reported plan is roughly $1 billion for seed, $2 billion for early-stage venture, and $2.5 billion for an opportunity fund focused on more mature companies, meaning about $3 billion is aimed at seed and early-stage.
How does this compare to Khosla's previous fundraise?
The firm reportedly targeted $3.5 billion across three funds in early 2025 and closed around $4 billion in that vintage. A $5.5 billion target is close to a 40% increase, roughly 17 months later.
Why does the early-stage focus stand out?
Most large AI-era funds this cycle have concentrated on late-stage bets in a few foundation model companies. Khosla is directing the majority of its target at seed and early-stage, a more contrarian, early-first posture.
What has Khosla been investing in recently?
Recent lead deals include Norm Ai ($120M), General Intuition ($320M Series A), and Factory ($150M Series C), alongside earlier AI positions such as OpenAI, Cognition, Sakana AI, and Physical Intelligence.
Have a fund closing to announce? Submit your fund here.
Need help raising capital? Check out our Fundraising Advisory services.