Token Rounds, Vanishing VC Brands and the Zepto Story
Key points
Key takeaways from the How I Invest with David Weisburd interview with Will Robbins, founder of Robbins Venture Capital (October 2026):
VC brands are losing their edge. With AI collapsing the cost of building software, Will argues the binding constraints on a startup are no longer capital access and venture-brand prestige but distribution and real-world trust networks. He cites a portfolio company where one adviser — a former senior executive in the US Medicaid programme — has been more helpful than all the VCs combined, because healthcare buyers recognise government leaders, not fund brands. He now backs companies from formation at Robbins Venture Capital, having previously joined Contrary as its first employee and helped it grow past $250M in AUM.
The 'token round' is reshaping seed economics. The old de-risking ladder — raise a seed round, hire engineers, prove each step, then raise $20M — collapses when coding models write the software, leaving '50 billion tokens' rather than a year of engineer time between a founder and a fully featured product. Founders can validate in the market and go straight to a $20–30–50M round, or raise as little as $500K–$1M to spend on tokens, dilute very little, and only take a scaling round from a multi-stage firm once validation exists.
An $800,000 product. A friend who had already taken a company public raised only $4M for his next venture and, six months in, had spent $800,000 of it on a fully featured product ready to launch — keeping any gaps self-funded. Will says fund managers should size funds to play both the capital-light token round and the capital-intensive, ops-heavy round that AI cannot shrink, seen in deep tech and neobanks.
Markets in everything. Capital, talent and information each have their own cycles that are not always correlated: an unfavourable capital market can coincide with a favourable talent market as candidates hunger for quality roles. Tech is unusual in running both hot at once, which he cites as part of why seed rounds stay large even as build costs plunge.
Short-run LPs, long-run founders. Quoting the Buffett aphorism that markets are voting machines in the short run and weighing machines in the long run, Will says a fund lives and dies by its LPs fund by fund, but a 20-year career rests on founder relationships and the flywheel they sustain. Incentives diverge on exits too: a VC wants the 100x deal that returns the fund, while a founder may rationally take a life-changing 10x offer — which is why he rejects blanket rules about how much anyone should raise.
Zepto, the Walmart of modern India. Will tells the story of two 17-year-olds who came to him at Contrary planning to skip college, and built Zepto into a dark-store network covering roughly a mile radius per store, served by large motherhub warehouses. Around five or six years in, it handles about two million orders a day and some $4B in annualised sales, with the CEO now 24 — and the team showed its rigour by redesigning its packaging to save a single rupee, about a penny, per order.
Retail and the accumulating advantage. The essence of great software companies is some accumulating advantage — network effects are one category, economies of scale another, and the latter made retailers such as Amazon, Walmart, Target and Home Depot the great outcomes of recent decades. Instacart shows the pattern: it started advertising in 2019 and ads are now about a third of its revenue and nearly all of its profit margin.
Back the person, not the pitch. Ranked by market cap, the defining venture outcomes skew heavily toward college and graduate-school dropouts — Facebook, Google, Airbnb and Stripe — which is why Will looks for something in the founder beyond the right product and market: the X factor that identifies an exceptional founder before everyone else does.
Read more: How I Invest with David Weisburd (YouTube)
