Dinki

I'm hosting a VC/LP evening in SF on Aug 27. no panels, just a private villa above the Golden Gate, appetizers, light cocktails, and the last informal gathering before fall fundraising season kicks off. leica photos throughout the night + some swag to send you into the season with the right aura. sponsored by @Wefunder and @rhobusiness. private. application only. for VCs and LPs: luma.com/6i9sei8v

a few insights after talking to a bunch of VCs and LPs this week: 1/ a large number of EMs are trying to build concentrated portfolios without any real rationale behind it. Mostly they're either chasing the prevailing narrative ("we're in an era of concentration now!"), or bending to LP mandates that demand high-conviction portfolios of just 10-15 positions and won't write a check otherwise, or simply copying the mega-fund playbook – funds that concentrate their own bets in high-capex AI labs as a way to deploy the billions they've raised. But copying the model without understanding the reasoning behind it is destructive, and very few managers can actually execute a concentration strategy well. 2/ even among the best seed investors historically, only a small share of deals ends up returning the fund. One GP pointed to USV as an example, where that figure was around 7-8%. For a strong emerging manager today, his estimate was closer to 4-5%. His core argument: in the AI era, that picking rate at seed will only get worse, since the number of companies is growing and the pace of disruption is accelerating. The implication is that LP appetite for concentration may simply be out of step with what's actually optimal portfolio construction for most seed managers. 3/ many GPs are aggressive about growing fund size from Fund I to Fund II, or Fund II to Fund III, and few can really justify it. If you raise a $25M fund and then want $50M, then $150M – that's a major red flag for LPs unless there's a clear rationale. Scaling up fund size forces one of three tradeoffs: > you dilute focus by moving into later rounds (seed to Series A) > or you increase the number of portfolio companies (without the operational bandwidth to actually manage that many, or the ability to physically write that many checks within a 3-4 year investment period, given brand constraints) > or you push for higher ownership (which isn't always feasible either, since it means either leading rounds or hunting for deals where you can pay less which nudges managers to optimize for price over deal quality) 4/ the market got very excited about the upcoming SpaceX mega-IPO, but DPI from that exit hasn't really hit yet, and the liquidity window probably won't open before year-end. It'll likely take until early next year, along with a few more IPOs, before sentiment fully shifts – though plenty of mega-platforms are already raising billion-dollar funds off the back of it (e.g. Menlo, which raised $3B thanks to Anthropic bet). But most LPs, though, are already planning their budgets around early 2027. 5/ a separate, interesting pattern: some LPs and FoFs are increasingly looking past a manager's "brand" and toward their apprenticeship lineage instead. One LP I spoke with framed it as a bet on elite apprentices – it matters less whether someone formally came out of Sequoia or Benchmark or another tie-1 platform fund, and more whether they worked closely under one specific strong partner and absorbed their taste, judgment, network, and actual underwriting craft. The logic is that VC is still fundamentally an apprenticeship business, and if an EM was genuinely "under the wing" of an exceptional investor, the LP is getting a piece of their whole school of thought. 6/ one more LP insight I heard more than once: builders tend to convert better than institutional allocators. People who've built companies themselves generally have a better feel for what a small VC fund actually offers (access, judgment, network, founder quality) and are more willing to back a manager even without full institutional packaging. Institutional allocators default to "no" more often as their decision-making is more inertia-driven and career-risk averse, so it's simply easier to go with a familiar brand than back a niche or emerging manager, so EMs often find traction faster with builders, operators, and entrepreneurial FOs first, rather than trying to look "institution-ready" to everyone right out of the gate.