DashTape

Daily

Desk brief

2026-09-22 · 8 high-impact headlines

BTC$85,604.14Bitcoin+5.67%ETH$2,745.15Ethereum+2.87%USDT$0.9988Tether+0.12%BNB$788.48Binance Coin+1.07%XRP$1.51XRP+7.04%USDC$1.00USD Coin+0.01%

Market cap

+5.14%

78

Extreme Greed

Headlines

Flash

  1. Tencent Holdings surged over 6%, AI assistant Muse validates consumer-grade Agent, institutions bullish on Tencent's "Xiaowei" ecosystem moat.BlockBeats news, September 22: According to Bitget market data, Hong Kong tech stocks rose in early trading, with Tencent Music up over 6%, Tencent Holdings up over 6%, and Baidu, Kuaishou, and Alibaba up over 2%. On the news front, Meta's AI assistant Muse climbed to the top of the free charts on the U.S. Apple and Google app stores. Its key capability lies in its ability to read users' social content on Facebook Marketplace and Instagram, and connect to third-party services such as Gmail, Apple and Google Calendar, and OpenTable, and even access other websites through a browser. Market analysis suggests that for Tencent, its WeChat native Agent "Xiaowei" has a similar product logic to Muse, naturally possessing social relationships, official account and Channels content, mini-programs, WeChat Pay, and a merchant ecosystem. Since June this year, "Xiaowei" has gradually rolled out small-scale gray testing.
  2. A certain whale has accumulated 14,783 ETH since yesterday, with a total value of $39.9 million.BlockBeats news, September 22, according to on-chain analyst Aunt Ai (@ai_9684xtpa) monitoring, the smart money that previously bought low and sold high ETH in August for a profit of $3.666 million has added another $19.97 million position. The address withdrew 7,216 ETH from Binance 1 hour ago, and since yesterday has accumulated 14,783 ETH, with a total value of $39.9 million, at an average withdrawal price of $2,699.44, and currently has an unrealized profit of $720,000.
  3. Meta's Muse ignites chip stock rally, with Hong Kong-listed Montage Technology and Iluvatar CoreX surging.BlockBeats news, September 22, according to Bitget market data, Meta's AI agent Muse ignited a rally in chip stocks, with Hong Kong chip stocks surging across the board in early trading. Montage Technology soared over 14%, Iluvatar CoreX rose more than 8%, and Biren Technology, SMIC, GigaDevice, Hua Hong Semiconductor and others followed the upward trend. According to BIT (bit.com) market data, U.S. stock Meta closed this morning with a surge of 11.43%, as its personal AI assistant Muse replaced ChatGPT to become the most popular free app on the U.S. iOS platform.
  4. DeepSeek plans to train an 8-trillion-parameter model, nearly triple the size of the largest existing open-source models.DeepSeek is once again scaling up its models. CEO Liang Wenfeng revealed to investors that the company is training a 2 trillion parameter model, and later plans to develop an 8 trillion parameter model. The current flagship V4-Pro has only 1.6 trillion parameters. What does 8 trillion mean? Among the currently publicly available ultra-large models with open weights, Moonshot AI's Kimi K3 has reached 2.8 trillion parameters, already the world's largest open-source model. DeepSeek's planned 8 trillion parameter model would be nearly 3 times larger in scale. However, more total parameters do not mean all parameters must be invoked in every run. Kimi K3 uses a MoE architecture, where only 104 billion of its 2.8 trillion total parameters are activated per token.
  5. DeepSeek bets on Huawei training chips, Liang Wenfeng says 'must succeed'The Information , citing people familiar with the matter, reported that DeepSeek CEO Liang Wenfeng, at a recent closed-door investor meeting, listed training large models on domestic chips such as Huawei's as one of the company's biggest bets, and said this matter "must succeed." DeepSeek expects to receive a new batch of Huawei chips usable for training as early as the fourth quarter of this year, with the specific models and quantities not yet disclosed. Liang Wenfeng believes that Huawei chips are expected to catch up with Nvidia in just a few years, and hopes DeepSeek will adapt in advance to hardware beyond Nvidia. DeepSeek is still using Nvidia chips to train its models. In May this year, Liang Wenfeng estimated that training a top-tier large model requires about 50,000 Nvidia GB300 chips, while switching to Huawei Ascend 950 would require about 200,000. At the time, the biggest problem was Huawei's insufficient supply.
  6. Crypto market surge saves 'Machi' holdings, net asset value rebounds above $11 million.BlockBeats news, September 22, according to on-chain analyst Aunt Ai (@ai_9684xtpa) monitoring, the crypto market surge saved "Machi Big Brother" Huang Licheng's positions. His contract account originally had only $800,000 left, and the total net asset value has now rebounded above $11 million. According to TradingBeats monitoring, "Machi's" current positions are as follows: ETH long position approximately $82.69 million, unrealized profit approximately $4.28 million, entry price $2,611.45, liquidation price $2,452.66. BTC long position approximately $29.09 million, unrealized profit approximately $880,000, entry price $83,225.30, liquidation price $59,335.61. HYPE long position approximately $15.18 million, unrealized profit approximately $140,000, entry price $93.08, liquidation price $36.18.
  7. Arthur Hayes: AI 'Safety First' Is Actually a Destruction of Computing Power Demand, All US Government Options Lead to Money Printing, Ultimately Bullish for BitcoinBlockBeats news, September 22 - Arthur Hayes published a new long-form article titled "Safety First," with the core argument that Anthropic, OpenAI, and SpaceX's claims of prioritizing "safety" to slow AGI development are not driven by concern for human welfare, but by economic realities. The market doesn't lack demand for AI; it wants AI at "China prices," meaning intelligence that is 100 times cheaper than current levels. Hayes points out that "Safety First" essentially means the destruction of compute demand. If spending on training new models declines and labs shift to efficiency optimization, customers will spend less on compute. The three major AI labs generate no profits, yet their compute demand underpins over $1 trillion in investment-grade debt and hundreds of billions in low-quality debt, which rely on off-balance-sheet backing from profitable tech companies like Nvidia, Broadcom, Google, and Microsoft. The real bag holders are U.S. insurance policyholders. Hayes cites Nick Nameth's analysis to reveal a "captive insurance scam": PE giants (Apollo, KKR, Brookfield, etc.) acquire insurance companies, stuff AI data center debt and AI-impacted SaaS private credit into insurance assets, and then provide fake backing through affiliated captive reinsurers with minimal capital. Nameth estimates such fake reinsurance assets total $1.54 trillion. Once AI data center debt is downgraded by rating agencies due to insufficient compute demand, insurers will be forced to add capital, but the affiliated reinsurers simply cannot pay, leading to insurer insolvency. Most U.S. states have insurance guarantee limits of only $250,000 to $300,000, and surviving insurers pay into guarantee funds only after the fact, which encourages maximum risk-taking by all parties involved. When AIG was bailed out in 2008, TARP funds ultimately flowed to Goldman Sachs and spawned record bonuses, while ordinary people only received foreclosure notices. Hayes believes this scene will repeat. For crypto investors, the conclusion is a win-win. If the U.S. government chooses to become the "buyer of last resort for compute," it will print money under the guise of national security to fund unproductive economic goods, driving up financial speculation and Bitcoin prices; if the government chooses to bail out insolvent insurers, it will similarly need to print money to cover bad AI debt, increasing money supply and driving up Bitcoin. Hayes specifically notes that the Federal Reserve unanimously voted to raise rates by 25 basis points last week, and RMP bond purchases stopped on August 14, but commercial banks have taken over by creating over $100 billion in money, and the rate hike gives banks an additional $7.5 billion annually in excess reserve interest, which will be used to expand lending and market speculation. Overall, the net effect remains stimulative. The crypto market's consolidation after a small rise in late August is about to end, dollar supply will continue to grow, and Bitcoin and select altcoins will rise. Hayes also calls this situation "incredibly beautiful," as the government will not allow the free market to stop building AI data centers, compute spot will see oversupply, AI agent usage will increase, and the surge in money printing will drive investors to chase crypto assets.
  8. The U.S. Department of Justice is investigating whether Binance violated sanctions against Iran.BlockBeats news, September 22, according to Bloomberg, people familiar with the matter revealed that U.S. federal prosecutors are investigating whether Binance Holdings Ltd., the operator of the world's largest cryptocurrency exchange, violated U.S. sanctions on Iran by failing to prevent certain transactions on its platform. The investigation is being handled by the U.S. Attorney's Office for the Southern District of New York, with the Criminal Division of the U.S. Department of Justice in Washington also participating. The relevant authorities are examining whether Binance knowingly allowed the transactions in question.