Khosla Bets $10M on Founder Whose Last Company Imploded

Khosla Ventures has committed $10 million to Ian Crosby's new venture, Synthetic, a move that requires either extraordinary faith in second acts or a truly alarming case of amnesia. Crosby previously founded Bench, a bookkeeping automation platform that imploded—a detail that somehow did not disqualify him from another ten-figure bet on essentially the same problem. The fact that Khosla is doubling down on this particular founder and this particular category suggests either a profound belief in redemption arcs or a portfolio strategy that relies heavily on hope as a risk mitigation tool.

Synthetic positions itself as a 'fully autonomous AI bookkeeping service for startups,' which is a polite way of saying it exists in a category so commoditized that even the word 'autonomous' cannot make it sound interesting. Bookkeeping automation has been a solved problem since the late 2000s—QuickBooks, Xero, Wave, and a dozen other platforms have been handling this workflow for over a decade with varying degrees of competence. The addition of 'AI' to an existing category is the venture capital equivalent of adding 'blockchain' to a spreadsheet: technically defensible, functionally unnecessary, and structurally designed to justify a higher valuation.

The irony is almost too neat to mention. Bench was supposed to solve bookkeeping for startups. It did not. Now Crosby is back with Synthetic, also supposedly solving bookkeeping for startups, but this time with the magic ingredient that venture capitalists cannot resist. If Bench's failure teaches us anything, it is that the problem was not a lack of automation or intelligence—it was that startups do not want to pay for bookkeeping services, and accountants do not want to be replaced by software. Neither of these truths has changed in the intervening years.

Khosla's investment memo likely reads something like: 'Founder demonstrates resilience and pattern recognition from first attempt,' which is VC-speak for 'we are funding the founder despite the data.' The press release probably features language about 'reimagining financial operations' and 'enterprise-grade autonomous workflows'—terms that have been applied to every bookkeeping tool ever built, including the ones that quietly shut down after burning through their Series A.

History suggests that second-time founder dynamics in failed categories tend to produce second-rate results. When the market rejects a business model once, throwing fresh capital and a rebranded pitch at it rarely changes the underlying economics. Synthetic will either spend years slowly gaining adoption among the subset of startups that tolerate imperfect bookkeeping automation, or it will discover that the original market rejection was not about technology at all—it was about demand that was never there to begin with.

This deal reflects a broader VC compulsion: the belief that founder pedigree and fresh buzzwords can overcome category saturation. Khosla has elected to ignore one of the clearest market signals available—a previous founder's product failure in the exact same space—in favor of the narrative that this time will be different. In venture capital, 'second chance' often means 'we have convinced ourselves to ignore obvious precedent.'

At least when Synthetic inevitably underperforms, there will be a perfectly good explanation: the market was not ready for autonomous bookkeeping. It turns out that was Bench's problem too.