Hive Blockchain, a prominent Bitcoin mining farm, has announced its plans to offer customers access to its data centers for training large language AI models. The company claims that its solution provides enhanced privacy compared to competitors like OpenAI’s ChatGPT. During an earnings call on June 30th, Hive’s CEO and President, Aydin Kilic, expressed the company’s commitment to providing privacy, data ownership, and AI compute workloads on their GPUs through their Hive Cloud service.
Following this announcement, Hive experienced a 2% increase in shares before the market closed on Friday. This move reflects the growing convergence of the crypto mining and AI sectors. As mining profitability becomes more challenging, miners are turning to AI as a potential avenue for sustained profitability. The AI sector is currently experiencing significant growth and attracting investors from around the world.
It remains to be seen if miners can effectively compete with tech giants such as Amazon Web Services and Google, who have the advantage of economies of scale and extensive experience in operating customer-oriented data centers. Large language models are instrumental in understanding and generating human language by utilizing probabilistic calculations. These models have been trained using GPUs, which were initially designed for faster image processing but were later found to be suitable for running AI workloads.
Hive has built an extensive fleet of 38,000 GPUs initially for Ethereum mining. Although some GPUs have since been repurposed for mining altcoins, many remain available for rent-as-a-service to support the company’s cloud offering. The potential for Hive to generate revenue by renting GPUs to AI-focused firms appears promising. The company anticipates a run rate of $1 million per year if they offer GPUs for AI training. Presently, 500 GPUs generate $230,000 in revenue per quarter.
In their recent fiscal year report, Hive disclosed $106.3 million in revenue for the fiscal year ending on March 31st. The gross operating margin accounted for $50.4 million, roughly 47% of the total revenue. However, FY2023 earnings were approximately half of FY2022’s earnings, amounting to $211.1 million, with a gross operating margin of 78% of revenue at that time. The report also indicated a net loss of $236.4 million, including non-cash charges such as depreciation, equipment impairment, and deposit impairment.
Overall, Hive Blockchain’s move to offer AI model training in their data centers presents an opportunity for the company to tap into the growing AI sector and diversify its revenue streams. By leveraging their extensive GPU fleet, Hive aims to cater to the needs of AI-focused companies while ensuring privacy and data ownership. This strategic expansion could potentially contribute to the company’s financial growth in the coming years.
**Editor Notes:**
Hive Blockchain’s decision to provide customers access to their data centers for large language AI model training demonstrates the company’s adaptability and foresight in the evolving technology landscape. By prioritizing privacy and data ownership, Hive sets itself apart from competitors like OpenAI. With the convergence of the crypto mining and AI sectors, it’s commendable to see companies like Hive exploring new avenues for profitability. As the market continues to evolve, it will be interesting to see how Hive’s GPU rentals for AI training unfold and contribute to their overall success. For more insightful news and updates on the latest advancements in AI, visit the [GPT News Room](https://gptnewsroom.com).
