Yesterday, US Bitcoin spot ETFs recorded a net inflow of $314.3 million, while US Ethereum spot ETFs saw a net inflow of $179.8 million.

According to monitoring by OnchainLens, a crypto whale has withdrawn 27,290 HYPE tokens from OKX, valued at approximately $2.23 million. About two months ago, this same wallet withdrew 47,340 HYPE tokens worth around $3.08 million, bringing its current total holdings to 74,810 HYPE tokens valued at roughly $5.33 million.

According to GMGN market data, the market capitalization of PONS, a token in the Robinhood Chain ecosystem, has climbed above $120 million, now standing at approximately $121 million. The token has recorded a 24-hour gain of 36.25%, with its current price at $0.120.

Against the backdrop of Bitcoin’s resurgence and a rebound in crypto-related stocks, investment bank Mizuho believes the quality of the current cryptocurrency market rally may be better than previous cycles. The bank’s analyst Dan Dolev noted that the current upswing is not primarily driven by high leverage. Crypto-denominated open interest has fallen to a one-month low after the initial rally, indicating the funding structure is closer to being driven by spot markets and ETFs — a key factor for market sentiment. In many past crypto rallies, rapid leverage buildup often amplified volatility, and once prices pulled back, it tended to trigger a chain of liquidations. Mizuho points out that spot Bitcoin ETF inflows have been more prominent in this cycle, with a net inflow of around $1.9 billion over the past week, marking the strongest weekly inflow since October 2025, showing traditional capital channels are still supporting crypto assets. The bank expects platform-based firms like Robinhood, eToro, and BitGo to benefit the most if the rally continues, as a rebound in trading volume will directly boost revenue from brokerage, custody, and institutional services. Especially amid a recovery in retail trading, sustained ETF demand, and expanding institutional custody needs, crypto infrastructure companies are more likely to achieve stable revenue elasticity than single tokens. However, the market will still be influenced by U.S. Treasury yields, the U.S. dollar’s performance, and risk appetite. If the Jackson Hole symposium delivers a hawkish signal, or if the U.S. stock market’s AI sector continues to correct, crypto assets may face short-term pressure. Mizuho’s assessment leans toward the medium-term structure: this rally has fewer leverage bubbles, and if spot demand continues to flow in, crypto stocks will have a clearer path for earnings transmission.

According to TradingBeats monitoring, ZEC is currently trading at $784.5, down around 7.8% in 24 hours. While this pullback has cut the daily losses of the "BTC OG" whale's ZEC short position by approximately $2.16 million, its current unrealized loss still stands at $11.154 million. The whale is currently shorting 32,800 ZEC with 2x full leverage, with a position value of roughly $25.699 million, an average entry price of just $444.0, a return of about -153.4%, and a liquidation price of $2,550.7. Meanwhile, the whale is longing 1,868.3 BTC with 3x full leverage, with a position value of around $148 million, an average entry price of $77,089.9. BTC is now trading at $79,060, with this long position generating an unrealized profit of approximately $3.681 million, a return of roughly 7.7%, and a liquidation price of about $44,563. From last night to this morning, the whale added 600 BTC at an average price of $79,146.0, with a transaction volume of around $47.488 million. Currently, BTC and ZEC are the only two contract positions held by this address. However, the unrealized loss on its ZEC position has expanded to roughly 3.03 times the unrealized profit on its BTC position, leading to a combined unrealized loss of approximately $7.473 million across both positions. On-chain perpetual and address analysis tool TradingBeats is now live, supporting real-time viewing of Hyperliquid data, tracing whale operations from addresses, and delivering in-depth, comprehensive analysis.

Morgan Stanley has initiated coverage on Nvidia’s credit profile and assigned a Neutral rating. The bank notes Nvidia still boasts a robust balance sheet and cash flow, but its credit exposure will grow more complex as the company participates in the over-$500 billion AI infrastructure financing platform. Morgan Stanley’s key focus is Nvidia’s evolution from a chip supplier to a broader AI ecosystem financing enabler. Beyond direct GPU sales, the company may help cloud providers and data center operators scale computing power investments via residual value guarantees, revenue sharing, credit support, and co-financing arrangements. The bank projects Nvidia’s broad credit exposure could reach nearly $200 billion by the end of 2028. This does not alter Nvidia’s core position in AI hardware, but will prompt the market to re-evaluate its risk pricing. Recent heightened volatility in AI stocks has led investors to focus on cloud providers’ capital expenditure returns, AI revenue realization, and data center financing pressures. If future depreciation of AI computing power assets outpaces expectations, or if some clients’ cash flow falls short of market assumptions, Nvidia’s ecosystem financing arrangements will emerge as a new valuation variable.

Beating AI Express: Short positions in two Chinese publicly listed large language model (LLM) companies, MiniMax and Zhipu AI, have hit all-time highs simultaneously. Data from S&P Global shows that shorted shares of MiniMax equal 20% of its freely tradable shares, while Zhipu’s short interest stands at around 6%. MiniMax will release its interim report after Hong Kong stock market closes on August 26, while Zhipu will report its results on August 31; short sellers are increasing their positions ahead of the earnings releases. Both firms were heavily hyped when they went public earlier this year. As of now, Zhipu’s stock remains over 800% above its IPO price, while MiniMax’s is roughly 80% above its IPO price, but both have fallen more than 50% from their peak levels. After Kimi K3 launched in July, Zhipu’s stock dropped around 24% at one point, while MiniMax’s fell approximately 18%. Zhipu later rolled out its GLM-5.3 model; Jefferies noted its performance is close to Kimi K3, with single-task costs around 19% lower, yet its stock price has not seen a meaningful rebound. Stock supply has also risen. The IPO lock-up periods expired in July, with 25.68 million shares of Zhipu and 150 million shares of MiniMax becoming tradable, totaling around $115 billion at the time’s prices. Meanwhile, southbound capital continues to buy the dips, with its holdings rising to roughly 12% of Zhipu and 8.1% of MiniMax, though this has failed to lift their share prices. Analysts hold different concerns about the two firms. Hedgeye says Zhipu is constrained by price wars, limiting its ability to raise prices and expand profit margins; MiniMax, meanwhile, is “neither the smartest nor the cheapest.” The two upcoming interim reports will answer one key question: Can pure-play LLM companies generate profits amid increasingly low-cost competition?