金链之光
06/22 16:40
HIVE Digital (HIVE) Stock Soars 22% Following Columbia University AI Performance Validation
HIVE Digital shares climbed more than 22% Monday following Columbia University’s validation of its Paraguay AI GPU facility.
Columbia researchers discovered HIVE’s Nvidia A40 GPUs delivered performance comparable to H100 systems on specific LLM pretraining tasks.
AI training experiments were executed remotely from New York on hardware positioned over 5,000 miles away in Asunción, Paraguay.
The study findings have been submitted for presentation at NeurIPS, a prestigious machine learning conference.
HIVE’s 100MW power substation in Yguazú, Paraguay is scheduled for energization in September 2026, with a Tier III data center facility to follow.
Shares of HIVE Digital Technologies (HIVE) experienced a significant rally exceeding 22% during Monday’s trading session, reaching above $7 in morning activity, following academic validation of its Paraguay-located artificial intelligence GPU infrastructure by Columbia University researchers.
The academic investigation, conducted in partnership with Columbia’s Department of Industrial Engineering and Operations Research, revealed that HIVE’s Nvidia A40 GPU systems deployed in Paraguay achieved performance levels equivalent to Nvidia H100 configurations on certain large language model pretraining operations — specifically for models up to 1.4 billion parameters.
The research team spent two months optimizing their codebase specifically for A40 architecture before conducting comprehensive measurements of throughput, latency, and bandwidth metrics. When adjusted for baseline hardware capabilities, the A40 performance data corresponded closely with previous H100 benchmark results for their particular application.
Additionally, the researchers performed serving throughput and latency evaluations on the 1.4B-parameter model, complemented by standard benchmark testing using LLaMA model architectures.
The research paper has been formally submitted to NeurIPS — the Conference on Neural Information Processing Systems — widely recognized as one of the premier forums for machine learning research.
Remote AI Training Across Continents
A particularly notable aspect of the investigation involved the geographic separation: the research team based in New York City executed multiple iterative training cycles on GPU hardware physically located in Asunción, Paraguay — spanning a distance exceeding 5,000 miles.
Executive Chairman Frank Holmes emphasized that the findings demonstrate “high-performance computing does not need to be limited by geography.” CEO Aydin Kilic characterized the A40-H100 performance equivalence as a “powerful result” that confirms the company’s engineering-focused strategy.
The experimental data collected during this collaboration will form the technical foundation for HIVE’s forthcoming AI computing campus in Yguazú, Paraguay.
Paraguay AI Facility Development Progresses
HIVE continues advancing its large-scale artificial intelligence and high-performance computing installation in Yguazú. Civil construction work on a 100-megawatt electrical substation has been completed, with commissioning activities planned for this summer and complete energization projected for September 2026.
Construction of a Tier III specification data center facility is slated to commence in fall 2026. HIVE anticipates the infrastructure will become operational during the latter half of 2027.
The company positioned the Columbia partnership as independent third-party validation supporting its strategic transformation beyond Bitcoin mining operations toward the AI infrastructure sector.
Holmes further noted that the initiative represents a significant achievement for Paraguay as well: “Paraguay has the power, the strategic location, and now the proof point.”
The 100MW substation commissioning represents the next critical near-term objective in HIVE’s infrastructure development timeline.
The post HIVE Digital (HIVE) Stock Soars 22% Following Columbia University AI Performance Validation appeared first on Blockonomi.
#World Cup Predictions: 100,000 USDT Daily#2026 World Cup Posting Challenge on HTX Square#TradFi Trading Strategies Sharing Challenge
1分享
全部评论0最新最热