Banks to sell $3 billion in loans for Meta data center
A consortium of banks, including Natixis SA, Mitsubishi UFJ Financial Group Inc., and Societe Generale SA, is selling $3 billion in loans to support the construction of a Meta Platforms Inc. data center in New Albany, Ohio, aimed at enhancing artificial intelligence infrastructure. The financing consists of two construction loans and is expected to be marketed by April 16, following Meta's previous $30 billion bond sale and a $29 billion private capital deal for another data center in Louisiana.
Meta Launches Muse Spark AI Model Amid Restructuring
Meta Platforms Inc. launched its first artificial intelligence model, Muse Spark, developed by its newly established Superintelligence Group, as part of a significant restructuring of its AI division led by CEO Mark Zuckerberg. Muse Spark, which is proprietary and available through Meta's AI portal, reportedly outperformed previous models in writing and reasoning tests but lagged in coding capabilities compared to competitors. This launch follows Zuckerberg's substantial investment in AI talent and infrastructure, with the company forecasting expenditures of up to $135 billion this year, nearly double last year's spending.
Antares Nuclear Secures Approval for Microreactor Demonstration
Antares Nuclear Inc., a California-based startup, has received approval from federal regulators to proceed with its demonstration system, becoming the first company authorized under a U.S. program designed to accelerate new fission technologies. The Department of Energy approved Antares's documented safety analysis for its Mark-0 microreactor, which is set to undergo fuel loading in the coming months, with a target for criticality by July 4. Antares plans to utilize its early reactors for defense and space applications starting in 2028.
EU Begins Implementation of Artificial Intelligence Act
The European Union has commenced the implementation phase of its Artificial Intelligence Act, focusing on compliance obligations for providers of general-purpose AI models. This phase follows the Act's adoption in 2024 and emphasizes a risk-based governance framework, requiring providers to maintain up-to-date technical documentation and ensure transparency for downstream users. The Act distinguishes between general-purpose AI and models posing systemic risks, which face additional regulatory requirements.
