Note: The X account is not owned by the original Enron entity. Users should be aware of the risk.

According to a report from The Information, UK-based emerging cloud services provider Nscale has informed potential investors that it currently holds approximately $103 billion in contracted revenue, following a $45 billion compute power deal with Anthropic. Relevant documents show that Nscale had disclosed around $51 billion in contracted revenue to investors before finalizing the Anthropic agreement, a figure that has since grown further via the Anthropic order and other contracts with AI clients. The documents note that Nscale’s average contract term is roughly 5.7 years, translating to an annualized scale of about $180 billion for the total contracted value, though the documents do not specify actual projected revenue. Nscale plans to conduct its initial public offering (IPO) as early as this month.

Representatives from major global economies have unanimously agreed to adopt the U.S.-proposed governance principles for artificial intelligence (AI) and emerging technologies. U.S. Secretary of Commerce Lutnick said Group of 20 (G20) member states unanimously adopted the framework, dubbed the "Carolina Principles," on Wednesday, a key win for the Trump administration and Silicon Valley. The principles call on G20 nations to implement differentiated regulatory approaches across industries, account for potential costs of delaying technology rollouts when testing new technologies like AI, and strengthen cooperation between governments and private enterprises in emerging tech testing. Michael Kratsios, White House science and technology policy advisor, stated the U.S. delivered a clear message to attendees: it aims to drive global economic growth and innovation. The principles will be submitted for formal approval by world leaders at the G20 summit hosted by Trump in Doral, Florida this December.

Beating AI News Flash: US Commerce Secretary Lutnick said artificial intelligence company Anthropic has mended ties with the Trump administration, with months of sharp disagreements over national security issues seemingly resolved. Lutnick noted the two sides had a "highly public rift" before, but Anthropic "has come around." He added that Trump and other administration officials helped bridge the gap and secured the company's support, stating, "We've straightened things out, and now everyone is on the same page." Earlier, the US Commerce Department briefly imposed export controls on two of Anthropic's top AI models in June, lifting the restrictions two weeks later after Anthropic took steps to address the government's national security concerns.

According to Bitget market data, U.S. stocks closed on Wednesday: the Dow Jones Industrial Average posted a preliminary 0.56% gain, the S&P 500 rose 0.46%, and the Nasdaq advanced 0.45%. Nvidia (NVDA.O) and Oracle climbed 3%, SK Hynix (SKHY.O) and Micron Technology (MU.O) gained over 2%, while Dell (DELL.N) surged 15.7%. Among crypto-related concept stocks, MSTR fell 1.35%, CRCL dropped 0.95%, COIN declined 1.05%, BMNR slipped 1.28%, SBET edged down 0.61%, and PURR fell 0.31%.

Beating AI News Brief: According to a Wall Street Journal (WSJ) report, the U.S. government has filed court documents supporting OpenAI’s position in its copyright dispute with publishers including The New York Times, arguing that using news content to train AI models aligns with the "fair use" doctrine under U.S. copyright law. The U.S. Department of Justice (DOJ) stated in the filings that restricting AI companies from using copyrighted content to train large language models would violate fundamental copyright principles and hinder innovation and scientific progress. The New York Times sued Microsoft and OpenAI in December 2023, alleging that ChatGPT and Microsoft Copilot used a large volume of the Times’ content to train their models and generate responses. The U.S. government claims AI training is "highly transformative" to original works and does not harm the market value of copyrighted content; tightening fair use rules would likely leave only large tech companies able to afford licensing fees, undermining the competitiveness of the U.S. AI industry. Parties involved are expected to submit new motions after completing the evidence exchange, requesting the court to rule on core disputes.

Beating AI News Flash: Meta released an update to its Muse Spark 1.3 model on Wednesday, claiming the version delivers significant performance improvements in coding and agentic tasks. Meta AI head Alexandr Wang said the new model is "competitive with frontier models" and will pave the way for future personal AI agent products, enabling AI assistants that can work on users’ behalf around the clock. Muse Spark 1.3 carries the same price as the prior version, a strategy Wang called "aggressive". Meta added that its "contributor tier" option has been well-received by developers. The program lets Meta use developers’ work to refine models, drastically cutting costs for coding products, with a "meaningful double-digit percentage" of developers opting in. Alexandr Wang noted that as model capabilities advance, security has become a critical internal priority for Meta, and the company is boosting investments in security and alignment. Muse Spark 1.3 will launch on Muse Code and Meta API on the same day, while the highest inference version will be released after additional security testing is completed.