OAK Research

🟣 TLDR: What happened at Binance? According to the latest estimates, around $6.3B was stolen from @binance across roughly ten blockchains overnight. That makes it the largest hack in history, far ahead of the 2025 @Bybit_Official hack, which resulted in $1.5B in losses. ◽ In its initial report, Binance said the hackers appear to have breached its systems through a DevOps engineer's laptop. In a statement released this morning at 9:47 AM, however, the exchange disclosed new intrusion vectors and described the attack as "complex." ◽ On-chain investigators (@zachxbt, @chainalysis, @tayvano_) converge on the fact that the attack was carried out by TraderTraitor, the North Korean hacking group behind the @KelpDAO / @LayerZero_Core ($292M) and @bitget ($388M) exploits. xitter.gq/zachxbt/status/2116030… ◽ According to @blockaid_, part of the hack was reportedly made possible by exploiting EIP-7702, introduced with Ethereum's Pectra upgrade. Put simply, this EIP allows a user-owned account (EOA) to delegate certain operations to a smart contract after signing an authorization with its private key. In this case, the attacker obtained an authorization signed with the key to Binance's main Ethereum hot wallet, handing control over to a malicious smart contract. Once the delegation was live, the contract was able to move all of the assets to the attacker's addresses in a matter of minutes. ◽ Of the $6.3B stolen, @tether and @circle managed to freeze $1.56B and over $800M respectively. The Arbitrum Security Council, for its part, froze 8,426 ETH along with other assets. In total, according to Binance, around $3B has been frozen by various parties, nearly half of the total amount. ◽ At around 10 PM last night, the hackers deposited $60M in USDT on Aave V3 (Ethereum Core) and borrowed 19,762 WETH (around $44.6M), the idea being to swap an asset that can be frozen (USDT) for one that can't. The protocol immediately froze almost all of its markets to keep the situation from getting worse. For the record, the hacker's position remains overcollateralized and poses no risk to Aave depositors. Meanwhile, @LlamaRisk recommended temporarily lowering Slope2 on the stablecoin reserves and on WETH on the Aave V4 Ethereum Core Hub, so that borrow rates don't exceed 10% at maximum utilization. Withdrawals had pushed utilization of the main stablecoin reserves to between 92% and 100%, with borrow rates ranging from 10% to 36% APR. ◽ Binance says user funds are safe and has pledged to cover all losses through its SAFU fund, which holds around 15,000 BTC (≈ $1B), as well as the company’s own capital. Withdrawals, suspended last night, should resume within 48 hours, according to @heyibinance. ◽ The good news is that the market hasn't overreacted: $BTC and $ETH are down a fairly modest 3% and 3.9%, while $BNB has lost just 6.6%. Credit where it's due to all the protocols that moved quickly to contain the damage: @tether, @circle, @ripple, @Paxos, @arbitrum, @aave and @USDT0_to. Without them, this hack would have been on an entirely different scale. We're continuing to monitor the situation and will keep you posted on any developments.

🟣 OAK Research turns two today. We launched on October 29, 2024, with the goal of building the first French crypto research firm and going deeper than the daily noise. Two years later, our focus has moved from technical deep dives to investment theses. We see ourselves first as a team of investors, and everything we publish is meant to help you make better decisions. Here’s where we stand today: ▫️ 800,000+ page views on the website since launch ▫️ Nearly 8 million views across our socials over the past year ▫️ Nearly 70,000 people following our work That growth also opened doors. This year, we started working with protocols we wouldn’t have imagined signing at launch, including @ethena, @aave, @maplefinance, @Plasma, @re, @PythNetwork or @krakenfx. The team grew along the way, from three people at launch to nearly ten today. Huge thanks to @Artem_Oak, @LilianAliaga_, @DecryptTV, @oak_intern, @_prenomnom_, @MaximilienP51, @0xApart and @Prometherr for building OAK every day. And thank you to everyone who reads, shares and trusts our work. Year three starts now. Job’s not finished.

🟣 Ethena Pay launches Earn+ and offers up to 12% APY on USDe. Launched a month ago in around fifty countries, @EthenaPay already has about 1,400 active users, who have deposited more than $8 million in ethereum:0x4c9edd5852cd905f086c759e8383e09bff1e68b3 into the app and have already surpassed $750 000 in weekly spending with the card. Until now, these deposits automatically earned up to 6% per year. With Earn+, open to all app users, this yield can rise to as much as 12%. There is no lock-up, and Ethena Pay charges no fees on deposits or withdrawals. But how can Ethena Pay offer such a yield on USDe? It all relies on a well-known looping strategy in DeFi, which Earn+ makes accessible to as many people as possible. The funds are deposited into an external vault managed by Upshift and Steakhouse. This vault repeats the same operation on an Aave v4 market deployed on @avax : ◻️ It deposits USDe as collateral. ◻️ It borrows ethereum:0xdac17f958d2ee523a2206206994597c13d831ec7 against this collateral. ◻️ It uses this USDT to buy back USDe, which it in turn redeposits. ◻️ It repeats the process until reaching leverage of about 11x. Regarding the security of the operation, on @aave v4, as on v3, the USDe oracle is pegged to USDT. A temporary depeg of USDe on the market therefore does not trigger a liquidation on its own. For now, the vault is capped at $4.5 million in total. Each user can deposit up to $20,000 in Standard, $50,000 in Pro and $200,000 in VIP. The global cap will be raised as it fills up.

🚨 $HOOD drops more than 10% after releasing its earnings On paper, the quarter had plenty to reassure the markets. @RobinhoodApp reported record revenue of $1.42 billion in Q3 (+12% year over year), driven by record volumes: • $992 billion in notional equity volume (+53%), • 905 million options contracts (+48%), • 16.4 billion event contracts on prediction markets (up more than 7x). However, several signals worried investors, and the stock, which had already lost nearly 20% over the past month, fell sharply again. Here is what weighed on it: ◻️ Earnings below expectations. Diluted EPS came in at $0.51, versus roughly $0.56 expected, with net income down 14% to $476 million. Part of the gap comes from an unrealized loss of $67 million on Robinhood Ventures Fund I, the rest from a tax expense that nearly doubled ($149 million versus $78 million a year earlier). ◻️ Record volumes, but a declining take rate. The number of options contracts traded rose by 48%, while options revenue grew by only 31%. Robinhood now earns about $0.44 per contract, compared with nearly $0.50 a year ago. This is explained by its compensation model. Market makers pay Robinhood for each contract, and that amount decreases when spreads tighten or when clients place multi-leg orders. And management acknowledged that spreads were even slightly tighter at the start of Q4. ◻️ Decline in crypto revenue Crypto revenue fell 55% to $120 million, a much steeper decline than that of volumes (-38%). The take rate thus dropped from about 33 to 24 basis points, partly because Bitstamp, whose fees are much lower, now accounts for 58% of the group’s crypto volumes. ◻️ Slowing net deposits. Robinhood brought in $15.7 billion in Q3, compared with $21.7 billion the previous quarter, a decline of about 28%. This slowdown weighs heavily, especially since, according to management, the majority of new deposits come from people who are already Robinhood customers. This shows that the platform is finding it increasingly difficult to attract new customers and therefore to expand its user base. ◻️ Rising costs Opex (operating expenses) climbed 16% to $739 million, notably because provisions for credit losses rose from $26 million to $71 million with the growth of the Gold Card. As a result, net margin fell from 44% to 33%. To sum up, Robinhood did record record volumes this quarter, but its revenue did not keep pace. Added to this are the decline in crypto revenue, the slowdown in deposits and rising costs, which did not reassure investors at all. In our view, the recovery could partly come from Robinhood Chain. In its first quarter on mainnet, the chain entered the top 5 blockchains by revenue and volume. It recorded more than $80 billion in DEX volume, more than the app and Bitstamp combined, and more than $50 million in chain fees. If crypto activity picks up again, it could become a real growth driver for Robinhood.

🟣 Do the DEXs with the most listed pairs necessarily generate the most volume? Since the new perps DEX meta, the listing race has accelerated significantly. Each platform keeps listing new markets (stocks, commodities, pre-market) in order to attract new traders, to the point that the number of available pairs has now become a growth argument in its own right. However, our infographic shows that this figure is not necessarily a relevant indicator of a protocol’s actual activity: ◻️ @Aster_DEX and Variational are the two platforms offering the most markets, with 592 and 574 perps markets respectively, compared with 346 for Hyperliquid. Yet @HyperliquidX processed $7.22 billion in volume over the last 24 hours, almost twice as much as Aster and Variational combined ($3.9 billion). ◻️ On Aster and @variational_io, around three quarters of markets record less than $100,000 in daily volume, and more than 40% don’t even exceed $10,000. ◻️ On Hyperliquid, the situation is completely different. 141 markets exceed $1 million in daily volume, including 59 above $10 million, which is more than Variational, Lighter and Aster combined. ◻️ The same pattern can be seen on @Lighter_xyz, which is the DEX with the fewest listed markets (199 currently). Yet the DEX generates more volume than Aster, with $1.87 billion recorded versus $1.53 billion for its competitor. The number of listed markets is therefore currently more of a marketing argument than a genuine indicator of activity. That said, having many illiquid pairs is not necessarily a disadvantage for a perp DEX. Indeed, as long as the infrastructure and maintenance costs of these markets (oracles, risk parameters, market making) remain low, each of them can remain profitable despite limited volume. They also broaden the platform’s offering, which can attract new traders, and some of them may eventually become liquid.

🟣Plasma One hits a new record 223 days. That’s how long it took Plasma One to surpass ~$100 million in spending through its card. Launched on March 17, @Plasma card crossed this milestone on October 26, in just over 7 months. No other crypto card with payments verifiable onchain had reached this threshold so quickly. For a sense of scale, @ether_fi, which held the record until now, took 371 days to reach this threshold. Kolo and Karta needed around 600 days, and Gnosis Pay 643 days, almost three times as long as Plasma One. An important note: this ranking only includes cards whose volumes can be tracked onchain and are therefore 100% reliable. KAST, Tria and RedotPay are therefore not included, nor is Wirex One, which inherited Wirex’s existing customer base.