As Bitcoin (BTC) mining companies expand into AI infrastructure, analyses suggest that the newly proposed compute derivatives may not fully mitigate their core risks. Futures based on GPU rental prices can cover some price fluctuations, but risks related to data center construction costs and financing remain separate.
On the 20th (Korean time), the U.S. Commodity Futures Trading Commission (CFTC) began soliciting opinions on compute derivative contracts. The review items include the compute spot market, market surveillance, manipulation potential, customer protection, and perpetual compute futures. CFTC Commissioner Michael S. Selig stated in a press release, "The U.S. cannot win the AI competition without a robust compute derivatives market."
Exchanges are also accelerating the commercialization process. CME Group announced plans to launch 'H100 Rental Index Futures' and 'B200 Rental Index Futures' on October 5 in collaboration with Silicon Data. These two products will track the hourly GPU rental price indices of NVIDIA ($NVDA) H100 and B200, with listing subject to regulatory review.
Previously, it was reported that CME Group is pushing for the launch of GPU rental-linked futures. Compute refers to the computational resources needed to train or operate AI models as a service. Until now, GPU rental prices have varied based on suppliers, regions, contract durations, and customer sizes, making it difficult to compare actual costs even when using the same semiconductors.
ICE (Intercontinental Exchange) and NATIVX also revealed plans to launch 'Energy Normalized Compute Futures' reflecting electricity costs on July 1. ICE explained that this product could provide price discovery and hedging environments similar to natural gas and electricity futures. This product is also contingent upon completing regulatory procedures.
The issue lies in the scope of hedging. GPU rental futures target compute price fluctuations. However, the AI transition for mining companies cannot be explained solely by GPU prices. Factors such as electricity, land, cooling, construction schedules, equipment procurement, and borrowing costs move together.
HIVE Digital Technologies announced on August 17 that it signed a five-year GPU cloud service contract through BUZZ HPC. The contract value is approximately $350 million (about 488.3 billion KRW). The annualized revenue is about $70 million (about 97.7 billion KRW), and BUZZ HPC's total annual revenue will increase to approximately $180 million (about 251.1 billion KRW).
In the same contract, HIVE stated that it will build 2,016 NVIDIA Blackwell Ultra GPUs with the GB300 NVL72 rack system. The expected capital investment is about $185 million (about 258.1 billion KRW). The customer advance payment is approximately $35 million (about 48.8 billion KRW), which is about 10% of the contract amount.
This contract has increased HIVE's revenue visibility, but it does not eliminate the burden of capital investment and additional equipment financing. It has been previously noted that HIVE's AI cloud contracts and construction cost burdens are significant. This indicates that the timing of actual equipment deployment and operation remains a more critical variable than the contract size.
Riot Platforms ($RIOT) announced on August 10 that it signed a 20-year data center lease for 191 MW at the Rockdale campus. The initial contract value is approximately $9.1 billion (about 12.7 trillion KRW), with a potential value of about $16.1 billion (about 22.5 trillion KRW) including a five-year extension option.
Riot plans to sequentially supply the first 96 MW by December 2027 and the entire 191 MW by June 2028. The initial development costs include a temporary financing limit of $573 million (about 799.3 billion KRW) from Morgan Stanley. Even if the long-term lease contract is confirmed, construction schedules and financial conditions must be managed separately.
In this structure, compute futures can provide some price benchmarks. However, HIVE's actual deployed equipment is from the Blackwell Ultra and GB300 series, while CME Group's initial futures are based on H100 and B200 rental prices. If the chip generations and contract structures differ, basis risks arise where actual revenues and futures prices move differently.
The academic study "Compute Futures: Who Is Left Exposed?" also pointed out that benchmarks tailored to one side of the market can hedge only that side's risks, and correlations between different H100 indices may not be sufficiently high. Basis risk is the risk that the price movements of actual holdings or contracts diverge from the price movements of hedging products.
Market reactions have been favorable towards the AI transition itself. In public reports and community discussions, stock price reactions and target price adjustments following HIVE and Riot's contract signings were mentioned. However, investment opinions and target prices can vary depending on the issuing entity and reference point, making it difficult to interpret that the business risks for mining companies have been eliminated.
The AI transition for mining companies is seen as a choice to reduce fluctuations in BTC mining revenues. However, the risks identified in this case are broader than GPU rental prices. Compute futures may serve as a tool for measuring AI infrastructure prices, but the impact on project financing and execution risks will be confirmed through liquidity and actual use cases after listing.
This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.





























