Skip to content
Uncapped with Jack AltmanUncapped with Jack Altman

From BoxGroup to Benchmark and Back | Greg Rosen, Partner at BoxGroup | Ep. 22

(If you enjoyed this, please like and subscribe!) Greg Rosen is a Partner at BoxGroup. Greg was the first hire at BoxGroup outside of the founders, David Tisch and Adam Rothenberg. After moving to the West Coast to work with Benchmark and Bedrock, Greg rejoined BoxGroup and currently invests out of their San Francisco office. An engineer by training, Greg built iOS games in high school before dropping out of college at 19 to join Jim Pallotta's venture fund in New York City. BoxGroup is an NYC-based seed stage venture capital firm that has invested in over 500 seed-stage startups over the last 15 years, including Plaid, Ro, Ramp, Clay, Scopely, Warp, Cursor, PillPack, Amplitude, Flatiron Health, Stripe, Warby Parker, Harry’s, Oscar, Flexport, Classpass, Vine, GroupMe, Airtable and more. We covered: - Being collaborative at scale - Avoiding adverse selection - Getting to a yes instead of no - Venture calendar audits - Running the right strategy Timestamps: (0:00) Intro (0:47) The collaborative venture model (5:47) Adverse selection vs coverage (11:59) How to see a ton of companies (14:44) Getting to founders early (21:10) Helping teammates get to a yes (23:25) Why there aren’t more BoxGroups (27:57) What’s learnable about picking (31:46) Calendar auditing (34:39) Focusing on where you’re outlier (37:25) Depth vs breadth of network (41:03) The future of code (44:00) Brain computers More on Greg: https://www.boxgroup.com/ https://x.com/grosen More on Jack: https://www.altcap.com/ https://x.com/jaltma https://linktr.ee/uncappedpod Email: friends@uncappedpod.com

Greg RosenguestJack Altmanhost
Aug 19, 202545mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

BoxGroup’s Switzerland venture strategy: win by seeing more, earlier

  1. Rosen argues collaborative venture is inherently anti-scale unless you refuse to lead rounds, because leading breaks the trust required to be “Switzerland” in syndicates.
  2. BoxGroup scales by increasing deal velocity (70–80 investments/year from ~5,000–6,000 qualified opportunities), trading ownership for better access and lower adverse selection.
  3. He contends venture has become far more competitive; since most investors can’t realistically get 2–10x better at picking, the practical response is to see more opportunities earlier and make it easy for founders to say yes.
  4. The conversation also covers operating mechanics (single-trigger decisions, pods, ‘help people get to yes,’ calendar auditing) and closes with views on AI-native developer tools and speculative brain-computer interfaces.

IDEAS WORTH REMEMBERING

5 ideas

Collaborative venture collapses the moment you start leading rounds.

Rosen says even leading a small number of seed rounds creates immediate trust issues: other investors assume you’ll withhold the best deals. BoxGroup preserves sharing by never leading and keeping a consistent ‘Switzerland’ posture.

BoxGroup trades ownership for access to better companies (adverse selection avoidance).

Instead of maximizing percentage ownership, they optimize to be in the right outlier companies. The bet is that lower adverse selection plus more entries into top outcomes beats a model that owns more of weaker picks.

If venture got harder and your process didn’t change, you implicitly accepted lower returns.

Rosen frames venture as see→pick→win→value; with competition up, unchanged inputs mean worse outputs. BoxGroup’s ‘humility’ response is to improve the variables they can—primarily ‘seeing’ more and earlier.

“If there’s a deck, it’s too late” is a sourcing strategy, not a slogan.

They try to meet ‘the human before the company,’ often before someone leaves their job. This requires lots of meetings that won’t convert, but earns the right to be the first yes and help assemble the lead.

Single-trigger + ‘get to yes’ culture is a structural advantage at seed.

Consensus processes push risk down and slow decisions, which breaks early-stage velocity. BoxGroup uses small pods and a norm of pulling threads to support conviction rather than default hole-poking—reducing costly errors of omission.

WORDS WORTH SAVING

5 quotes

“Collaborative venture is hard… it has to be anti-scale in certain dimensions.”

— Greg Rosen

“The second you start leading those rounds, you just can’t be Switzerland.”

— Greg Rosen

“We see on average about five to six thousand qualified opportunities a year… and make about seventy to eighty investments a year.”

— Greg Rosen

“If there’s a deck, it’s too late.”

— Greg Rosen

“Consensus-driven decision-making at Seed doesn’t work, in my opinion.”

— Greg Rosen

Collaborative venture as “Switzerland”Adverse selection vs ownership maximizationScaling via deal velocity (seeing more)Outbound sourcing before a deck existsSingle-trigger decisions and pod structureCalendar auditing and “empty calories”AI-native future of coding; brain-computer speculation

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome