Hive Digital Technologies shares surged 25% on Monday to around $5.20, marking their highest level in more than seven months, after the company disclosed results from artificial intelligence research conducted by a Columbia University team using its GPU cluster in Paraguay.
The research performance has been submitted to the NeurIPS conference, one of the largest global AI events, scheduled for December.
Remote AI training shows strong performance
Researchers based in New York trained neural networks on Hive’s infrastructure in Asunción, focusing on improving efficiency in large language models. The company said its Nvidia A40 GPUs delivered performance comparable to more advanced hardware in certain computing tasks.
The testing evaluated latency, throughput, and distributed processing, offering insights into the viability of high-performance computing powered by hydroelectric energy.
Paraguay infrastructure underpins strategy
Hive launched the Paraguay-based platform in March to support long-distance AI training experiments. The company currently operates करीब 300 megawatts of energy capacity in the country, with another 100 megawatts under development.
The findings are now shaping a larger Tier-III data center project in Yguazú. A 100-megawatt substation is expected to go live in September 2026, with the full facility targeted for completion in the second half of 2027.
Expansion into AI infrastructure accelerates
Earlier this year, Hive announced plans for a 320-megawatt AI facility near Toronto designed to support more than 100,000 GPUs. The company also signed a three-year agreement worth $220 million with Bell Canada and Cohere to deliver AI cloud services.
That deal includes deploying over 2,300 Nvidia Grace Blackwell GPUs and is expected to generate roughly $70 million in recurring annual revenue starting in late 2026.
Market revalues business model
The share price jump reflects a broader shift in how traders are valuing the company, as its focus expands beyond digital asset exposure toward AI infrastructure.
Validation from an academic institution adds credibility to Hive’s technical capabilities, while the research suggests that software optimization can enable older Nvidia A40 GPUs to approach the performance of newer H100 systems in certain use cases.
The successful remote training between New York and Paraguay also strengthens the company’s strategy of building data centers in regions with low-cost renewable energy, potentially improving efficiency and returns.
With new long-term contracts and recurring revenue streams emerging, Hive is repositioning itself as a provider of core infrastructure for AI workloads alongside its legacy operations.
Explore how AI is transforming markets—discover web3, AI and crypto’s convergence driving the next wave of digital asset innovation.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
