USDT grew on Ethereum through 2024, but smart contract holdings stalled, BIS data show

According to a Bank for International Settlements working paper released on September 15, 2026, additional USDT was minted throughout 2024 without a corresponding prolonged increase in balances held within smart-contract accounts on the Ethereum network. Meanwhile, similar accounts on Tron maintained roughly 1% of the total USDT supply for the majority of the study’s historical dataset. Taken together, these insights challenge the common premise that an expanding stablecoin supply automatically indicates a fresh influx of capital into decentralized finance.

The BIS research examines where the tokens reside rather than the specific motivations of individual holders. Because the chart tracking holder balances concludes prior to 2026, its percentages should not be interpreted as measurements from September 2026. This distinction in timing is important, as contemporary dashboards continue to display significant USDT balances across both blockchain networks.

Smart-contract accounts on Ethereum held over 20% of the network’s USDT supply during periods in 2021 and 2022. That proportion remained fairly steady between 15% and 20% until late 2024, after which it declined to roughly 10% to 15% while overall token issuance grew. This shift represents a change in the relative proportion of tokens located inside contracts, rather than a drop in the absolute token balance. The authors of the BIS paper note that the rise in issuance failed to generate a prolonged surge in contract balances.

A shrinking percentage can occur when newly minted tokens accumulate outside of contracts, even if the volume inside those contracts remains close to previous levels. The study’s Ethereum data illustrates this exact dynamic, revealing a much higher volume of USDT held in non-contract accounts as total issuance climbed, lacking a matching sustained growth in smart-contract holdings. Consequently, looking solely at the percentage shift is an unreliable method for determining whether capital was pulled out of DeFi.

Figure 10 within the research paper estimates that Ethereum contract-held USDT ranged between approximately $10 billion and $15 billion toward the conclusion of its plotted timeline, while Tron’s equivalent figures sat at roughly $1 billion or less. These are approximate values derived from charts. Ethereum’s dollar-denominated balance varied in the low tens of billions while its proportional share decreased, whereas Tron’s contract balances stayed a minor fraction of a substantially larger total supply. Because these two percentages rely on distinct, chain-specific denominators, they cannot be combined into a single metric for DeFi adoption.

What a token balance can reveal

To conduct their analysis, the researchers reconstructed USDT distribution using transfer event logs from both Ethereum and Tron. They identified smart-contract accounts based on contract deployments, categorized all other addresses as externally owned accounts, and verified the overall token supply against events involving minting, burning, and blacklist destructions. Tracking the token directly offers a perspective distinct from aggregating deposit totals reported by individual DeFi protocols, which frequently results in the same tokens being counted multiple times.

Protocol-level total value locked (TVL) tracks assets allocated to specific decentralized finance applications. Conversely, the BIS methodology tracks a single token across various addresses spanning two networks, accounting for holdings situated outside of those specific applications. This makes the BIS approach better suited for understanding how USDT is distributed among different account types, whereas protocol TVL merely highlights the scale of chosen platforms. Still, neither approach transforms a balance into a confirmed indicator of a holder’s underlying intent.

Although this distinction refines the measurement, account types remain imperfect indicators of real-world economic activity. A smart contract might hold USDT to support a bridge, a wrapped token variant, or custody solutions rather than backing a DeFi lending or trading protocol. Similarly, an externally owned address could be utilized for everyday payments, savings, remittances, or exchange custody. Therefore, Tron’s roughly 1% contract-held share does not prove that the remaining tokens functioned as payments, nor does Ethereum’s declining share demonstrate that DeFi engagement shrank.

The current scale of the market remains considerable. As of September 28, DefiLlama reported a total USDT market capitalization across all chains of approximately $183.7 billion, which included roughly $73.3 billion on Ethereum and about $92.5 billion on Tron. These figures represent a separate, third-party snapshot of total supply rather than an update to the BIS holder distribution breakdown. As such, they cannot determine whether current tokens reside inside DeFi contracts, exchange wallets, or other account categories.

Any modern assessment regarding DeFi deployment would necessitate a current accounting of balances held within recognized DeFi contracts on each blockchain, while appropriately separating bridges and custodial accounts where feasible. The historical percentages provided by the BIS cannot deliver this modern update. On their own, rising USDT totals do not prove increased DeFi usage or higher transaction volumes, nor do they provide any insight into market demand for native assets like ETH or TRX.

Frequently Asked Questions

• Does a larger USDT supply mean more capital has entered DeFi?
  No. The BIS data shows that USDT issuance can grow without a sustained rise in balances held by smart-contract accounts on Ethereum.
• What percentage of USDT do smart contracts hold on Ethereum?
  Smart-contract accounts on Ethereum held over 20% during parts of 2021 and 2022, hovered around 15% to 20% until late 2024, and later dropped to roughly 10% to 15%.
• What share of USDT do smart contracts hold on Tron?
  Throughout most of the study’s historical series, smart contracts on Tron held about 1% of the network’s USDT supply.
• Can smart contract shares be used to measure overall DeFi adoption?
  No, because account types do not reveal the exact economic purpose of the holder, and smart contracts can be used for bridges, wrappers, or custody rather than active DeFi strategies.