NVIDIA launches OpenShell reference architecture to provide security protection for enterprise-grade AI agents.

According to an official announcement, Upbit will list Axis Robotics (AXIS) with KRW, BTC, and USDT trading pairs.

NVIDIA: Supported by capacity expansions in the United States, Taiwan (China), and Singapore, the company projects its total semiconductor testing capacity will double in 2027 compared to 2026.

Following yesterday’s U.S. after-hours trading, SpaceX released its second post-IPO quarterly earnings report. Data shows the company’s revenue rose 161% year-over-year and turned profitable. In response to the news, Wall Street institutions have lifted SpaceX’s target prices: Morgan Stanley set SPCX’s target at $300 with an "Overweight" rating; Wells Fargo raised SPCX’s target from $214 to $274; Citigroup increased it from $200 to $240; UBS hiked it from $210 to $230.

According to TradingBeats’ monitoring, two addresses on Hyperliquid—0x77dd and 0x0392, suspected to belong to a single entity—added a total of ~$15.89 million in ETH positions over the past two hours, at an average entry price of $2,453.50. The other two associated addresses recorded no position adjustments today. As of press time, combining the four linked addresses, this whale holds a total of ~$382 million in BTC and ETH long positions, with an overall unrealized loss of ~$8.21 million. Breakdown: - ETH long positions: 109,161.92 ETH total, valued at ~$268 million, with a weighted average entry price of $2,543.57, unrealized loss of ~$9.69 million. - BTC long positions: 1,373.60 BTC total, valued at ~$115 million, with a weighted average entry price of $82,285.74, unrealized profit of ~$1.48 million, partially offsetting ETH’s losses. The four contract accounts have a combined equity of ~$43.63 million. Calculated as total nominal exposure divided by total equity, overall leverage stands at ~8.77x. Notably, the ETH long positions of today’s added addresses 0x77dd and 0x0392 have estimated liquidation prices of $2,254.71 and $2,224.81 respectively, ~8.2% and 9.4% lower than the current mark price of $2,454.80. Associated addresses: 0x0392a716dbee1661ea781881826b928daeb5d7d9 0x77ddcc4b6440b3a32b7802f053105c2fad3ae0e9 0x60468ba27448f88b19f52ed02f1a8c52f7a0b21d 0x799fB74ec743a65865bD80E30bed8A2216A15f4C

Beating AI Express, a news outlet from Dongcha, reports that DeepReinforce, an AI team founded by renowned NLP scholar Li Jiwei, has added a new 122B MoE model to the Ornith-1.5 series. The Ornith-1.5 series was first released back in August this year with three variants: 397B, 35B-A3B, and 9B; this update fills the gap for a mid-sized model. The new model’s weights are now publicly available, with versions offered for FP8, NVFP4, GGUF, and MLX. Ornith-1.5-122B uses the same architecture as Qwen3.5-122B-A10B, with a total of ~122B parameters and roughly 10B activated parameters per inference. It supports a 256K native context window, image understanding, tool calling, and inference. Official self-tests show the 122B model scored 84.5 on SWE-bench Verified (non-hack), outperforming Qwen3.8-Flash-Next’s 82.8. On CyberGym it scored 80 vs. 78, NL2Repo 51.2 vs. 45.2, and SWE-bench Pro (no-hack) 76.9 vs. 43.9. However, it still lags on Terminal-Bench 2.1 with scores of 81.2 vs. 83.4. Notably, Qwen’s official team previously scored 62.5 on SWE-bench Pro using Claude Code Harness, higher than Ornith’s 43.9 measured via OpenHands, as the two test conditions are not fully aligned.

Beating AI Insight News: The DeepSeek open-source plugin project dsh-our-free-model, which has around 32,000 stars on GitHub, has suddenly received a 20,000-character "lawyer's letter". The poster claimed to be acting on behalf of the rights holders of WorkBuddy and CodeBuddy, accusing the project of reverse-engineering Tencent’s official client, bypassing interface verification, and using free credits to access AI models. The letter demanded the developer cease relevant activities within 3 days and delete the entire repository within 7 days, warning of potential litigation and regulatory reports if not complied with. However, the project had already removed Tencent-related channels at 10:29 AM that day. At 4:21 PM, another contributor submitted a pull request (PR) attempting to bypass Tencent’s client verification by modifying request content to restore model access. Approximately 39 minutes later, the "lawyer's letter" was posted. The contributor subsequently voluntarily withdrew the PR, and the code was not merged. What’s particularly odd is that the letter lacks key details: no law firm name, handling lawyer, authorized rights holder, contact information, or official seal. The account that posted the letter was only registered on October 16. The letter also claimed the developer had submitted code to restore access just 3 minutes after removing Tencent-related channels, but public GitHub records show the PR was actually submitted nearly 6 hours later. No official confirmation from Tencent regarding the letter has been released. Numerous developers in the community have questioned its authenticity, suspecting someone is using the lawyer’s letter to exert pressure.