Bessemer's $5.75bn close puts more than two thirds of the money behind growth rounds
One of the oldest names in venture has raised $4bn for growth and $1.75bn for early stage, a bet that the biggest AI returns will be earned late and in fewer companies.
Bessemer Venture Partners said on Wednesday, September 23, that it had raised $5.75bn of new capital in a single close. The money is split unevenly: $1.75bn is set aside for seed and early-stage investing and $4bn for growth rounds. For a firm whose reputation rests on finding companies before they have revenue, that is a striking allocation, and it says a great deal about where large limited partners now believe venture returns will be made.
The raise also matters because of who is making it. Bessemer traces its roots to 1911 and reports more than 155 IPOs, with a portfolio that has included LinkedIn, Pinterest, Shopify, Twilio and Canva. When a manager with that record decides to put more than two dollars into growth for every dollar it puts into early stage, the rest of the industry tends to take note.
A deliberate shift in the mix
The comparison with Bessemer's previous cycle is instructive. In September 2022 the firm closed $4.6bn across two vehicles, of which $3.85bn went to its flagship early-stage fund, BVP XII, and about $780m to a later-stage vehicle called Forge. Four years on, the early-stage pool is less than half of what it was and the growth pool is roughly five times larger. The structures are not identical, and Bessemer has not given the new vehicles formal names in its announcement, so the comparison is directional rather than exact. The direction, however, is unmistakable.
Bessemer insists that its early-stage habits are unchanged. According to the firm, about 70% of its investments still begin at the early stage, often before a company has revenue or even a defined category. That figure counts deals, not dollars, and the gap between the two measures is the heart of the story. The firm will keep making many small, early bets, but most of its capital will be concentrated in a handful of large cheques written later.
Talia Goldberg, a partner, put it plainly in an interview on the day of the announcement: "Returns are concentrating in fewer larger winners." Her conclusion was that "the right move for us is to be a meaningful investor." Bessemer says its growth platform is now a dedicated unit that will lead concentrated, high-conviction rounds, both in existing portfolio companies and in new ones.
The AI arithmetic behind the decision
The case for this mix rests on artificial intelligence. Since 2022, Bessemer says, it has backed more than 260 AI-native companies and invested over $3bn across what it calls the AI stack, from compute and infrastructure to foundation models, developer platforms, applications and agents. Its growth portfolio already includes Anthropic, Cognition, Ramp, Legora, ClickHouse, fal, EliseAI, EvenUp, ShopMy, DriveNets and Waymo. Earlier-stage examples it cites include Perplexity, Abridge and Wonderful.
Byron Deeter, a partner, argued in the announcement that "AI-native companies are scaling faster than any category of technology we've backed before." Jeremy Levine, another partner, framed the change as continuity: "While the playbook has evolved, our conviction has not."
The logic is easy to follow. AI companies need very large sums once they find product market fit, chiefly to pay for compute and talent, and they are staying private for longer. A firm that only writes seed and Series A cheques sees its ownership diluted away in exactly the years when the value is created. A firm with a large growth pool can defend its stake, or build one, in its best companies. Bessemer's own description of companies remaining private longer is an acknowledgment that much of the compounding that once happened on public markets now happens in late private rounds.
Part of a wider pattern among large managers
Bessemer is not alone. Kleiner Perkins announced $3.5bn of new funds in March, with $1bn for its 22nd venture fund and $2.5bn for growth, a ratio almost identical to Bessemer's. Accel has closed $3.5bn across four funds this year, and Bain Capital Ventures closed its eleventh fund at $1.6bn earlier this month, aimed at early AI infrastructure and physical AI.
The pattern is a bifurcation of the venture market. Established firms with long records can raise multibillion dollar pools and are using them to move up the capital stack. Smaller and newer managers, by contrast, have found fundraising much harder in a year when, as one account of the Bessemer close noted, many firms struggled to close funds at all. The limited partners who back the large franchises are, in effect, choosing concentration twice: first by picking a small number of managers, then by funding those managers to pick a small number of companies.
The risks in writing bigger cheques
Concentration cuts both ways. Growth investing in AI today means paying valuations that assume continued rapid revenue growth and a long runway of access to capital. If those assumptions hold, a few very large positions can return an entire fund. If they do not, a $4bn growth pool has far less room for error than a seed portfolio of hundreds of small bets, where a single breakout can cover many failures.
There is also a question of discipline. Leading growth rounds means setting prices, not merely accepting them, and the competition for allocations in the most sought-after AI companies is intense. Large crossover funds, sovereign investors and corporate strategics are all bidding for the same stakes. A venture firm entering that contest with more money than before must resist the temptation to deploy quickly simply because the capital is there.
Finally, the early-stage pool, while smaller than in 2022, still carries the firm's identity. If the best founders come to see Bessemer chiefly as a late-stage cheque writer, its seed pipeline could suffer, and with it the proprietary access to future growth deals that justifies the larger pool in the first place. The firm's insistence that most of its deals still start early is partly a message to those founders.
Our view
Bessemer has made an honest bet: it believes the returns of this cycle will be earned by holding large stakes in a few AI winners through their late private years, and it has sized its funds accordingly. For limited partners, that is a clearer proposition than a firm pretending its strategy has not changed. For founders, the message is mixed. Companies already in Bessemer's portfolio gain a backer able to lead big rounds, which is valuable when capital needs run into hundreds of millions. Earlier-stage founders should note that a smaller share of the firm's capital is now reserved for them, even if the number of seed deals holds up.
What to watch is pricing. The test of this fund will not be how fast it deploys $4bn, but whether Bessemer can lead rounds at prices that leave room for a return, in a market where the same handful of companies is courted by everyone with a large balance sheet. If AI valuations cool, the firms that tilted hardest toward growth will feel it first.
Sources
- Innovation Open Lab (Business Wire): Bessemer Venture Partners Closes $5.75 Billion to Back Founders From Seed Through Growth
- TBPN Digest: Bessemer Venture Partners closes $5.75B fund, with $1.75B for early stage and $4B for growth
- Tech Funding News: Bessemer Venture Partners closes $5.75B to back founders from seed to growth
- Technologies.org: Bessemer Venture Partners Closes $5.75 Billion, With $4 Billion Reserved for Growth-Stage AI Deals
- Fund Momentum: Bessemer Raises $5.75B for AI Bets
- Calcalist Ctech: Bessemer Venture Partners closes $3.85 billion for early-stage investments
- Gunderson Dettmer: Kleiner Perkins Announces New Funds Totaling $3.5 Billion
Filed under Venture Capital. Written by InvestTimes Editorial under our editorial standards. Commentary and analysis for information only, not investment advice.
