Nearly $13 million has been moved into the consolidation wallet of bankrupt crypto trading firm Alameda Research in just 24 hours, revealed data from blockchain security firm PeckShield on Feb. 2. 

The address received $6 million in Tether (USDT) and $2.5 million in Ether (ETH) from crypto exchange Bitfinex’s hot wallet, along with $4.5 million worth of USD Coin (USDC) from an anonymous source and 30,000 Lido tokens worth roughly $65,500.

The asset’s transfer is considered to be part of recovery efforts tied to bankruptcy proceedings. A spokesperson for Bitfinex told Cointelegraph that Alameda had an account on Bitfinex and the exchange is collaborating with the liquidators to refund the remaining funds.

Alameda filed for bankruptcy protection on Nov. 11, along with nearly 130 other companies controlled by FTX Group. Since then, its consolidation wallet has seen inflows from several addresses, accumulating over $26 million in ETH and $183 million in other altcoins, including $54 million in BitDAO tokens.

Related: Alameda Research had a $65B secret line of credit with FTX

The amount recovered, however, could be much bigger, as liquidators have reportedly suffered at least $11.5 million in losses — some of which were preventable — since taking control of Alameda’s trading accounts, according to a report from crypto analytics firm Arkham Intelligence.

Liquidators faced another technical obstacle when attempting to recover funds on Jan. 12. Cointelegraph reported the loss of $72,000 worth of digital assets on the decentralized finance (DeFi) lending platform Aave while liquidators attempted to consolidate funds into a single multisignature wallet.

To close a borrow position on Aave, the liquidators removed extra collateral, putting the assets at risk of liquidation. In addition, liquidators are also reported to have spent far more on gas fees than they earned from moving the funds.

Alameda Research nearly collapsed in 2018 before FTX came into play, according to new reports citing former employees. The firm’s algorithm was designed to make a large number of automated and fast trades, but was inefficient in predicting price movements.