When the Experts Best Placed to Judge Are Also the Ones Best Placed to Benefit

The RDI Fund's Real Test Isn't Whether Anyone Cheated

• It's Whether the Structure Could Ever Prove They Didn't TAN's position is that the RDI Fund's design gets the hard part right and the easy part wrong. It correctly recognised that evaluating deep-tech proposals - quantum computing, AI infrastructure, space technology - requires people who actually understand the sector and traditional bureaucratic evaluators don't. But it then relied on disclosure and recusal, the standard corporate-governance toolkit, to manage a conflict-of-interest risk those tools were never built to fully solve. THE CASE FOR the current structure is genuinely strong. India's private R&D spending sits below 0.7% of GDP and deep-tech is precisely the category ordinary venture capital avoids - long timelines, high failure rates and technical complexity that generalist evaluators can't assess. Pulling in venture capitalists and technologists who are already active in the space isn't a workaround, it's the only realistic way to get evaluators who can distinguish a genuine quantum-computing proposal from a well-dressed pitch deck. And the second-round data - only one of 13 companies linked to a panelist - is real evidence that the first round's concentration wasn't pure favouritism, since a structurally corrupt process wouldn't self-correct that sharply without any changed rules. THE CASE FOR structural conflict-of-interest failure is equally strong. Recusal only prevents a panelist from voting on a company they're invested in - it does nothing about the informational and reputational advantage that sitting on the panel confers in the first place or about the pull that draws ambitious founders toward investors who also happen to control access to public capital. In a sector where credible ventures are genuinely scarce and information is genuinely thin, being seen as someone who has "made it" onto a public panel is itself a market signal that benefits a panelist's existing portfolio companies - a benefit that exists whether or not that panelist ever casts a vote on their own investment. WHERE THIS LEAVES US: TAN's position is that the fund's structure genuinely falls short - not necessarily in outcome, but in provability. The problem with disclosure-and-recusal as a complete answer is that it cannot distinguish "the well-connected founders who happened to reach the panel first" from "a panel that quietly favoured its own network," because both produce exactly the same first-round numbers. A system built on public money needs an architecture that makes that distinction verifiable, not one that asks taxpayers to trust the panel's self-assessment of its own good faith. Rotating the evaluator pool and adding independent academic screening for initial technical assessment wouldn't slow deep-tech funding meaningfully - but it would let the next investigation come back with a structural answer, not just a defensive one.