Benchmark Capital, the storied Silicon Valley VC agency identified for early investments in eBay, Snap, Uber, and Twitter, is breaking with certainly one of its signature traditions: preserving its funds to about $425 million and backing solely younger startups. After greater than twenty years of limiting its automobiles to that quantity or decrease, the outfit has closed on commitments of $2 billion throughout two new funds, together with a $1.25 billion car devoted to later-stage investments, based on the Wall Avenue Journal.
Whereas the fund sizes of many enterprise capital companies have ballooned into billions of {dollars} over the past decade, Benchmark caught to the technique that helped make it legendary. By being staunchly selective and taking a big—sometimes 20%—stake in each startup the agency backed, it maintained a mannequin designed to maximise outsized returns for its restricted companions.
Nevertheless, Benchmark’s comparatively small fund sizes have doubtless prevented the agency from investing in capital-intensive AI startups, significantly basis mannequin makers, whose spherical sizes typically attain into lots of of hundreds of thousands. Consequently, the agency hasn’t invested in Anthropic, OpenAI, or any of the opposite capital-intensive AI labs, reminiscent of Periodic Labs, Reflection AI, or Recursive Superintelligence.
The place Benchmark has positioned AI bets, the outcomes have been blended. The agency led a $75 million spherical in Manus, a Singapore-based AI agent platform that hit $100 million in annual recurring income inside eight months of launching. When Meta agreed to acquire Manus for roughly $2 billion late final yr, it regarded like one other Benchmark winner within the making. However Chinese language regulators — arguing the corporate, which was based in China earlier than relocating to Singapore, had violated export management legal guidelines — blocked the deal in April, leaving Benchmark’s stake in limbo.
Benchmark’s new $750 million early-stage fund will give the agency extra flexibility to write down checks in an surroundings the place early-stage valuations have skyrocketed. Whereas the agency has historically backed firms on the Sequence A stage, Benchmark has just lately given itself extra flexibility to put money into firms at different early phases of growth.
In current months, Benchmark backed two Sequence B startups: Gumloop, a platform that permits enterprises to create AI brokers with out writing code, and Monaco, an AI-native gross sales and CRM platform.
Benchmark normal associate Everett Randle beforehand advised TechCrunch that the agency seems to construct a “significant and deep relationship with the entrepreneurs, and that may occur comparatively early within the firm’s lifecycle, at seed, [Series] A, at [Series] B.”
The agency dipped its toe into late-stage investing when it raised a $225 million particular function car (SPV) to take part in a $1 billion pre-IPO spherical for Cerebras, as TechCrunch reported earlier. Benchmark first led the chipmaker’s Sequence A in 2016. Cerebras held its IPO final month, returning Benchmark $3.25 billion on the IPO value.
That windfall prompted the agency to lift a devoted progress fund. That new car will make 5 to 6 giant investments in each present portfolio firms and new startups, based on an individual aware of Benchmark’s technique.
The 2 new funds aren’t the one modifications at Benchmark. During the last two years, the agency has undergone a big shift in its normal companions.
In 2024, Miles Grimshaw left the agency to rejoin Thrive Capital. Then, final yr, Sarah Tavel—Benchmark’s first and solely feminine normal associate so far—took on the less-involved position of enterprise associate, whereas Victor Lazarte departed to begin his personal VC agency.
To replenish its ranks, Benchmark — which historically runs with 4 to 6 normal companions — added two new high-profile buyers to its workforce: Randle, poached from Kleiner Perkins, and Jack Altman, the brother of OpenAI CEO Sam Altman. The strikes counsel that even Benchmark, lengthy outlined by its resistance to progress, now sees the AI period as requiring a distinct playbook — extra capital, extra phases, and contemporary blood on the associate desk.
Once you buy by means of hyperlinks in our articles, we may earn a small commission. This doesn’t have an effect on our editorial independence.
