Agrawal maps venture return benchmarks by fund stage using three core metrics, IRR measuring speed of returns, TVPI combining realized and unrealized value, and DPI measuring cash actually returned, arguing IRR alone misleads early in a fund's life due to J-curve effects and should be paired with DPI to know whether gains are actually realized. The piece gives stage-by-stage benchmark ranges, roughly 20-40%+ net IRR and 4-6x+ MOIC for top seed funds down to 10-15% IRR and 2-3x MOIC for growth equity, and walks through a worked example: a $100 million seed fund's $3M check at a $12M post-money dilutes from 25% to roughly 16% by exit across two rounds, producing a 10.7x single-position return but only an 11.6% blended net IRR. Citing PitchBook-NVCA data, it notes that excluding just the five largest deals and exits from a recent quarter collapses reported deal and exit value by 73.2% and 86.6% respectively, and that roughly 32,000 private companies worth $3.8 trillion remain unsold per Bain, meaning many current fund IRRs rest on unrealized marks rather than cash actually distributed.

Founders negotiating dilution across rounds get a genuinely actionable reframe: rather than asking whether an investor's fund performs well in the abstract, the piece argues the useful question is which IRR tier that specific investor sits in and what their fund needs from your round to hit target returns, since a fund tracking 8% IRR in year six faces very different pressure than one tracking above 20%. The worked seed-to-exit dilution model, moving a single position from 25% ownership down to 16% across two rounds while producing a strong 10.7x deal return but a modest 11.6% blended fund IRR, gives founders a concrete numerical anchor for understanding why VCs push hard on ownership percentage even on deals that will individually outperform, since fund-level math dilutes the impact of any single win. The specific list of common IRR calculation mistakes, using committed rather than called capital, comparing gross to net figures, or benchmarking across mismatched vintage years, is also a practical due diligence checklist for LPs or founders trying to sanity check a GP's stated track record rather than taking a headline return figure at face value.