Stablecoin payment network Atum completes $13.5 million funding round led by Variant.

Nov 11 Update: #Bitcoin ETFs: 1D NetFlow: -1,649 $BTC(-$121.78M)?? 7D NetFlow: -15,624 $BTC(-$1.15B)?? #Ethereum ETFs: 1D NetFlow: -11,512 $ETH(-$23.65M)?? 7D NetFlow: -127,828 $ETH(-$262.66M)??

Global central banks purchased more than 321 tons of gold in the third quarter of 2026, marking their largest quarterly gold purchase since the fourth quarter of 2024 and the fifth-largest such quarterly procurement on record. The volume of gold bought by central banks rose 8% from the over 296 tons purchased in the second quarter of 2026, and surged 453% compared to the first quarter of 2026. China and Poland led demand in the third quarter, adding over 63 tons and 40 tons respectively. Year-to-date, global central banks have acquired more than 675 tons of gold, a 6% increase over the same period in 2025.

On Trade.xyz, Anthropic’s IPO pre-market perpetual contract (xyz:ANT) has surged over 10% in nearly six hours, currently quoted at around $106.7, with a 24-hour gain exceeding 8.6%. The contract trades at a roughly 37% premium to the upper end of the company’s newly announced IPO price range of $70–$78, corresponding to an implied listed valuation of approximately $1.4 trillion and a fully diluted valuation of around $1.92 trillion. Earlier market reports noted that Anthropic’s blockbuster IPO was oversubscribed several times. Yesterday, Anthropic officials stated the company aimed for a $1 trillion valuation in its IPO, and was seeking a fully diluted valuation of $1.4 trillion. Anthropic expects to set its IPO price next Wednesday, with trading set to begin the following Thursday. The Anthropic IPO pre-market perpetual contract on Trade.xyz references one Class A common share of Anthropic, PBC (Nasdaq ticker: ANT), and operates in the pre-listing market, reflecting the market’s assessment of its expected share price. Trade.xyz added that once Anthropic officially goes public, the contract will convert into a standard XYZ perpetual contract.

UK AI cloud service provider and data center developer Nscale, backed by Nvidia, is considering delaying its planned initial public offering (IPO) with an approximate valuation of $35 billion, as management evaluates whether to push the timeline to next year. Nscale has not yet made a decision on the matter and did not immediately respond to a request for comment. The company originally planned to list on the New York Stock Exchange under the ticker symbol NSCL, and had been preparing to go public as soon as possible. Headquartered in London, Nscale filed for an IPO in September this year, targeting to raise around $3.5 billion at the time, including $2 billion from Nvidia. Just days ago, Microsoft declined to exercise its option to expand its existing agreement for computing power rental at Nscale’s Texas data centers.

According to on-chain analyst Ai Yi (@ai_9684xtpa), F2Pool co-founder Wang Chun withdrew 3,996 ETH worth $8.23 million four hours ago, then staked the ETH on Lido and deposited the wstETH into the Spark protocol.

Tether has launched the open-weight model QVAC BotPsy, an 800-million-parameter system built on Alibaba’s open-source Qwen 3.5 architecture, which converts user daily requests into local tool calls on personal devices. In internal tests, Tether said the model set a new benchmark for tool call performance among models with fewer than 1 billion parameters, and outperformed all tested models with up to 1.2 billion parameters. Across four key tests—BFCL v4, ACEBench, τ-bench, and ToolSandbox—its results roughly doubled those of the base model Qwen 3.5-0.8B-Think, and outperformed Qwen 3.5-2B-Think, which has 2.5 times more parameters. Specifically, it scored 46.1 points in BFCL’s multi-turn tool call test, nearly double the 23.8 points of the 2B model. Tether added that BotPsy reduces the need to send personal data like emails and calendars to centralized service providers, and can detect when no suitable tool is available to avoid generating false tool calls. Tether CEO Paolo Ardoino stated that users should decide for themselves what content the assistant can access and which external services it can connect to. The model’s weights have been published on Hugging Face.