Yazılar

OpenAI, Samsung SDS and SK Telecom to Begin Korea Data Center Build

/in /tarafından

OpenAI, Samsung SDS, and SK Telecom are preparing to begin construction of artificial intelligence data centers in South Korea as early as March, according to Science Minister Bae Kyung-hoon.

The South Korean government previously said the U.S. startup would form joint ventures with the two Korean firms to develop two facilities with an initial combined capacity of 20 megawatts. The planned infrastructure is intended to support AI model training and deployment, reinforcing South Korea’s role in the global AI supply chain.

SK Telecom confirmed it is in discussions with OpenAI regarding a data center project in the country’s southwest, though the exact timeline for construction remains under review. The collaboration reflects increasing partnerships between U.S. AI developers and Asian technology firms to expand computing capacity amid surging demand.

The initiative comes as major AI companies race to secure power, land, and advanced hardware to support rapidly scaling models. South Korea’s strong semiconductor ecosystem and digital infrastructure make it an attractive base for AI expansion, particularly as governments compete to host next-generation data facilities.

Inside SpaceX’s xAI Deal: Tax, Debt and Legal Advantages

/in /tarafından

The sale of xAI to SpaceX delivers significant tax, financial, and legal benefits for investors, according to people familiar with the transaction. The deal uses a triangular merger structure that allows SpaceX to acquire xAI as a wholly owned subsidiary—rather than fully merging operations—thereby avoiding immediate repayment of billions in debt and limiting legal exposure.

The structure keeps xAI’s liabilities, contracts, and debt ring-fenced from SpaceX, insulating the parent from potential litigation tied to xAI’s social media platform X and its Grok product. M&A attorneys say this approach is commonly used to preserve corporate insulation while enabling operational independence.

Financially, the transaction qualifies as a tax-free reorganization. xAI shareholders can defer taxes on the SpaceX shares they received until they sell. The deal also avoided triggering change-of-control provisions in xAI’s debt—critical as the company carries billions from prior financings—by routing the acquisition through intermediary entities. As a result, bondholders were not entitled to repayment, and xAI bonds rose following news of the deal.

The all-stock transaction values xAI at $250 billion and SpaceX at $1 trillion, making it the largest M&A deal on record, according to LSEG. Importantly, securities lawyers say the structure may help SpaceX avoid added disclosure hurdles ahead of a potential IPO later this year if xAI does not meet the SEC’s “significant subsidiary” threshold.

While some investors worry the added complexity could complicate valuation—combining rockets, satellites, defense contracts, AI, and social media—others say confidence in Elon Musk’s execution outweighs those concerns as SpaceX moves toward a historic public offering.

AI Method Helps Identify Which Dinosaurs Made Fossil Footprints

/in /tarafından

Scientists have developed a new artificial intelligence method that helps determine which dinosaurs made specific fossilized footprints, addressing a long-standing challenge in paleontology. Footprints are among the most common dinosaur fossils, yet matching them to the correct species has often relied on subjective interpretation. The new approach uses AI to analyze eight measurable traits in each footprint, offering a more objective classification system.

The research, published in Proceedings of the National Academy of Sciences, analyzed nearly 2,000 footprint silhouettes spanning 150 million years. The algorithm identified key features such as toe spread, heel position, load distribution, and left-right asymmetry, which together explain differences in footprint shape. Experts then mapped these traits to known dinosaur groups to guide identification of future discoveries.

Researchers say the method does not eliminate uncertainty, as footprint shape can vary depending on behavior, ground conditions, burial processes, and erosion. Still, it provides a consistent framework to compare tracks across time and locations. One notable result supported earlier findings that certain 210-million-year-old footprints from South Africa resemble bird tracks, despite predating the earliest known bird fossils by tens of millions of years.

The findings suggest AI can become a powerful tool in paleontology, helping scientists reconstruct ancient ecosystems and better understand dinosaur diversity, even when bones are absent and only footprints remain.