Bernstein's $10 trillion prediction-market forecast pegs financial assets' half on a 0.5% capture rate

The 70% compounding assumption averages two very different years. Volume went from about $50 billion in 2025 to roughly $300 billion in the first eight months of 2026 ^[4], and Bernstein's own $410 billion full-year estimate makes 2026 an 8.2-fold year ^[20]. From there, $410 billion to $10 trillion across nine years needs about 43% a year ^[21]. The estimate also puts a slowdown into the back of this year: $300 billion over eight months is $37.5 billion a month, while the remaining $110 billion spread over four months is $27.5 billion ^[22].

Nearly half of the 2035 number comes out of one line in the model. Take 49% of $10 trillion and you get $4.9 trillion; Bernstein's bottom-up route to the same leg is 0.5% of a financial-asset pool it sizes at roughly $700 trillion today and $900 trillion in 2035, which produces about $4.7 trillion ^[7][23]. The published summaries of the note do not say how that pool was measured.

Bernstein names the products it expects to carry that leg. "We expect new products such as KPI markets, which allow users to trade a single corporate metric, such as production, deliveries, or subscriber growth, rather than the stock price itself," the analysts wrote ^[8]. "Further, perp futures are expanding from crypto to commodities and single stock perps," they added ^[9].

Sports falling from 61% of volume to 38% is a statement about share. In dollars, 61% of $50 billion is about $30.5 billion and 38% of $10 trillion is $3.8 trillion, so the forecast has sports volume rising roughly 125-fold over the decade ^[24].

The rotation is uneven across venues. On Polymarket, sports accounted for about 52% of global volume in 2026 year to date against 39% in 2025, politics fell to 22% from 32%, and crypto was about 21% ^[13]. Kalshi's commodity growth is concentrated in one month: of roughly $590 million year to date, $410 million arrived in August, leaving about $180 million for everything before it ^[26]. Bernstein also lists 15-minute bitcoin contracts among the short-duration products behind this year's increase ^[15]. A contract that expires four times an hour recycles the same collateral many times a day, so notional volume can climb without new money arriving.

Robinhood's chief executive is talking the same way. "We're already seeing other categories like crypto taking a disproportionate share," Vlad Tenev told CNBC ^[16], and said: "I think within a few years, sports will actually be in the minority, similar to active trading at large" ^[17].

In my view the mix call is the better-supported half. Crypto's share of Kalshi volume went from under 5% in January to about 20% in August, which is a measured change ^[10], whereas the $10 trillion depends on a capture rate against a notional pool that nobody can observe yet ^[7]. The counter-case sits inside Bernstein's own data: Polymarket's sports share rose this year, not fell ^[13]. If financial-asset contracts are still near 12% of industry volume at the end of 2027, the 49% is a forecast about products that have yet to trade at size ^[5].