Two venues hold 84% of the $5.78 billion bet on stocks and commodities through on-chain perps

Divide $5.78 billion ^[1] by the 25.6-fold year-to-date rise in DefiLlama's data, as reported by Cryptopolitan ^[2], and the market at the start of the year comes out at roughly $226 million ^[15]. All of it sat on trade[XYZ], then the only venue listing these contracts ^[7]. trade[XYZ] has since grown its open interest about 17 times, to $3.866 billion ^[16]. Every other venue combined went from zero to about $1.91 billion, starting around April ^[17]. Over the same stretch the leader's share fell from 100% to about 67% ^[20].

The 84% figure ^[5] adds two separate systems together. The top two hold $4.834 billion between them, or 83.6% of the total ^[18], but they are different venues, and the share exposed to any single outage or pricing error is the 67% on trade[XYZ] ^[3]. Below them the field is thin. At least eight other venues ^[13] share about $946 million ^[19], and the largest of them, QFEX, holds $274.61 million ^[8], about 4.8% of the market ^[21].

Venue matters more here than it would for a tokenized share. A trader holding an Nvidia perp owns zero Nvidia shares, and the contract tracks the price and nothing else ^[6]. Tokenized stocks and Treasuries are usually backed by assets an issuer or custodian holds, and the $5.78 billion is counted separately from them ^[9]. These perps have no expiry, usually settle in stablecoins and carry leverage as standard ^[10]. None of the open interest is a claim on a share held anywhere, so each position exists only on the venue that prices and settles it ^[6][9].

The trading hours are the draw. Stock exchanges close at night and on weekends while perp venues keep running, so anyone with stablecoins in hand can open a leveraged gold or tech-stock position on a Sunday at 3 a.m. ^[11]. According to Cryptopolitan, that matters most to users outside the U.S. who face restrictions or high costs buying American shares through local brokers ^[12]. In those hours the contract has no live exchange price to follow ^[11]. The source does not say how trade[XYZ] or Variational marks positions while the underlying market is shut. With two-thirds of open positions on one venue, the Sunday price on trade[XYZ] is the price for two-thirds of the sector's open positions ^[3].

Three outcomes fit these numbers. Variational keeps the pace of recent weeks, now close to $1 billion ^[4], and the market ends up with two venues nearer in size. Or trade[XYZ] grows with the market and holds its two-thirds. Or an outage or a bad mark at the leader pushes traders toward a tail where no venue holds more than $274.61 million ^[8].

I'd expect the first, because the only trend in the data is a sole incumbent giving up a third of the market while its own book grew 17 times ^[20][16]. The case against is that share comes off a 100% starting point easily when the first rivals arrive, that the leader still holds four times its nearest competitor ^[14], and that Variational's acceleration is only a few weeks old ^[4]. If trade[XYZ]'s share stops falling while total open interest keeps rising, the concentration is settling in and this view is wrong.