By Gloria Ogbonna.
Revolutionizing Early-Stage Investment: Proprietary AI Model ‘Moneyball for VC’ Identifies Top Startup Contenders”
Traditional early-stage startup investments have often relied on subjective factors like intuition, founder experience, and personal connections. However, TRAC, a San Francisco-based venture firm established in 2020 by Fred Campbell, Joseph Aaron, Scott Pyne, Steve Marek, and Dick Fredericks, aims to transform this approach.
The firm has developed a proprietary AI model to predict which early-stage startups have the potential to become unicorns—companies valued at over a billion dollars. TRAC has disclosed 31 startups identified by their model exclusively to Insider and has unveiled insights into their methodology.
In a notable departure from conventional strategies, TRAC’s AI emphasizes pinpointing startups that are unlikely to succeed rather than exclusively focusing on potential winners. This methodology involves utilizing over 30 sources of both public and private data.
Aaron humorously notes, “This is the first time we have shared the ‘family jewels’ or our secret sauce.” At 74 years old, Aaron jests about being the oldest person to initiate a venture fund.
TRAC’s model presents surprising elements, including a unique approach that values investors over founders for predictive insights. Rather than founders, the model values the 255,559 investors in its database. Among these, around 300 “SuperForecasters,” including notable figures like Sam Altman, have exhibited a remarkable ability to identify successful startups.
These SuperForecasters have proven track records, profiting from two-thirds of their positions, and attaining over 10X returns in one out of five investments. TRAC identifies less than 2% of all startups as potential SuperForecaster candidates.
TRAC’s model usually identifies about 150 companies annually, yet this year’s count is lower due to reduced deal activity. The firm acknowledges the challenge of limited funds and the competitive nature of securing positions on startups’ cap tables.
Remarkably, TRAC is 4 times more likely to invest in women-led startups compared to most VCs. Campbell attributes this to the higher percentage of Angel SuperForecasters’ capital invested in such startups. TRAC’s formula has identified three future unicorns led by at least one female founder.
While the accuracy of TRAC’s formula takes time to assess, the firm notes that the companies it identifies have a 1 in 5 probability of becoming unicorns. TRAC’s success lies in its ability to eliminate false positives, investments that fail.
Aaron underscores their achievement, stating, “TRAC has made 53 Deal #1 investments, and none have lost money. That is the only stat we have with bragging rights.” Just as with early-stage investing, true evaluation often requires a decade or more due to the nature of venture investments.