Analysis: Due to Trump's lack of a clear Bitcoin strategy reserve plan, traders believe that BTC has limited upside potential in the short term

On-chain analyst Ai Yi (Twitter handle @ai_9684xtpa) reported that following Bitget’s opening of its BTC withdrawal portal, the exchange’s total protection fund of 5,500 BTC has been flowing in batches into its hot wallet to facilitate withdrawals. As of now, 2,042.28 BTC (valued at roughly $169 million) has been transferred, with 3,457.72 BTC remaining on-chain. However, since these tokens were moved in advance, the analyst emphasized that this figure does not represent the actual amount of user withdrawals.

Xie Jiayin, head of Bitget’s Chinese-speaking region, stated during the "Bitget Event Review" live stream that in consideration of users’ actual needs, the platform will first open Bitcoin network withdrawals at 16:00 Beijing time today. By this Friday, Bitget will resume withdrawals for all assets across the platform. Withdrawal channels for regular users, KOLs, and VIP clients will open simultaneously, with no priority granted to any group. Per prior announcements, the specific withdrawal timelines are as follows: September 28 at 16:00: BTC (Bitcoin network); September 29 at 16:00: ETH (Ethereum, BSC, Arbitrum, Base, Optimism); September 30 at 16:00: USDT (Ethereum, BSC, Solana, Tron); October 2 at 16:00: other tokens, fiat currencies, and P2P services.

On-chain analyst Ai Yi (Twitter handle @ai_9684xtpa) has tracked that address casualpig.eth has re-established an ETH position after a one-year pause. The address’s last ETH purchase came when the token hit a peak of $4,751.74. Three hours ago, it withdrew 1,754 ETH from Bybit, valued at $4.65 million. The address’s prior ETH trading cycle ran from August to October 2025, during the last bull market peak. It chased the high at $4,751.74 and exited when prices fell to $4,433.10, incurring an estimated loss of $449,000.

Bitget has launched its phased withdrawal feature as scheduled, with Bitcoin (BTC) withdrawal services now available. Per the previously announced timeline, withdrawals for Ethereum (ETH)-related networks will open at 16:00 UTC+8 on September 29, while withdrawals for Tether (USDT)-related networks will be accessible at 16:00 UTC+8 on September 30. Withdrawals for other tokens, fiat currency, and C2C services will resume gradually starting at 16:00 UTC+8 on October 2. Meanwhile, Bitget CEO Gracy Chen and Greater China Head Xie Jiayin are hosting a public livestream to review the recent security incident, share the latest updates, and answer user questions live.

Bitget’s Head of Chinese-speaking Markets Xie Jiayin stated during the "Bitget Event Review" live stream that the exchange suspended withdrawals following the recent attack because the incident involved multiple blockchains, rather than a single asset or network. Before completing checks and verification at every stage, Bitget did not want to expose any users to risk. Reopening withdrawals required separate fixes and security verifications for each chain, so phased recovery by chain and asset would take time. Xie emphasized that the withdrawal suspension had no connection to Bitget’s capital sufficiency. The Bitget Protection Fund totals $464 million, which fully covers losses from the attack. Furthermore, Bitget’s Proof of Reserves is 127%, and this figure is fully public and transparent.

Beating AI Express, Insight: Xiaomi’s MiMo team has reviewed the repeated tool call issue following the launch of MiMo-V2.6. The model sometimes repeatedly invokes identical or highly similar tools, draining context without advancing task progress. In OpenCode, the share of responses with repeated tool calls for Flash and Pro once hit 1.02% and 0.54% respectively. The root cause lies in reinforcement learning (RL) reward design: training prioritizes rewarding "final task completion accuracy" but fails to adequately penalize inefficient in-process behaviors. The original rule only imposed penalties when single-round tool calls exceeded 32 times; repeated calls under 32 times incurred no penalty at all. As RL scale expanded, this "bad habit" was further reinforced. After reviewing training checkpoints, Xiaomi found the proportion of abnormal samples with single-round calls over 10 times in Flash rose from 11.1% at step 0 to 24.6% at step 20. The most direct fix would have been cutting the penalty threshold from 32 to 8 calls, then re-running ~20 MixRL steps, estimated at $2.31 million. Instead, Xiaomi trained a dedicated RL teacher to correct repeated calls, running 12 steps and ~7,000 samples, then integrating this capability back into Pro and Flash via MOPD. The full fix cost ~$90,000, just 4% of the full retraining plan, with other key benchmarks largely unchanged. The fixed MiMo-V2.6-Pro-MOPD and Flash-MOPD weights are now available, and the API has been switched to the new version with unchanged call names. Xiaomi will also reset the remaining quota for MiMo Desktop users’ current cycle.