On August 5, 2026, Lisa Su tossed a seemingly industry-internal product announcement into the global asset pricing pool—AMD officially announced the mass production of the AI data center rack system Helios, while also releasing key figures countering NVIDIA's latest AI racks: Helios theoretically processes approximately 30% more inference tokens for each dollar spent. This is not a simple performance declaration but rather an expectation of downward pressure on the "unit dollar computing cost curve." Meanwhile, global electricity demand for AI data centers is on the rise, with U.S. gas power generation and natural gas delivery becoming a core beneficiary chain. Pipeline operators like Energy Transfer LP are pushing expansion projects to provide a thicker energy base for this wave of AI computing frenzy. On another invisible front, the U.S., Iran, and Oman are close to reaching a 60-day temporary agreement regarding the Strait of Hormuz—Iran and Oman will divert ships in and out of the channels, initially waiving tolls for 60 days and planning to clear naval mines, which means there is downward space for global oil and gas transportation disruption risks and oil and gas price volatility premiums in the short term. On the computing supply side, Helios's mass production and the battle for "dollar computing efficiency" begins; on the energy front, the expansion of natural gas infrastructure combined with the easing of Strait of Hormuz risks will lower medium and short-term energy risk premiums, with these two macro curves being pulled downward at the same time. For BTC and ETH, this is not just an improvement in the external environment but a recalibration of risk premium and valuation frameworks: the miner cost curves, the correlation with high-beta tech assets, and the on-chain financial games surrounding AI narratives all must find a new balance under the new expectations for computing costs and energy prices.
Helios Mass Production: From Performance Arms Race to Cost War in AI Computing
The timing of Helios's entry into mass production is subtle: Before this, AMD was often seen as NVIDIA's "backup," and even with single-GPU performance improvements, it struggled to shake up the procurement patterns at the data center level. Helios's direct entry as a data center-level AI rack system means AMD is no longer just competing for discourse at the chip level but is now bringing "whole rack computing"—a high-margin, long-cycle contract—onto the negotiation table. Compared to NVIDIA’s latest AI racks, AMD’s core selling point is not "faster speeds" but rather "up to 30% more inference tokens per dollar." This effectively tells all CFOs: what truly deserves optimization is not peak computing power but how many model calls and business scenarios each dollar of capital expenditure can yield. The competitive narrative shifts from "performance arms race" to "dollar computing efficiency," as the market begins to reassess AI infrastructure using total cost of ownership and return on investment measures, rather than merely focusing on TOPS and parameter scales.
Once the rack-level systems are compressed on the "inference tokens per dollar" dimension, the marginal cost of AI inference is rewritten. Companies will no longer discuss “whether to cut AI Capex” in budget meetings, but rather “if there’s reason to extend or even amplify the investment cycle for AI infrastructure under a lower cost curve.” What mass production of Helios releases is precisely this expectation of a downward shift in the computing cost curve: AI data centers are still high in energy consumption, but the economic improvement of output per unit of computing power, combined with falling energy risk premiums, drives the market to re-evaluate the “computing + energy” mid-term constraints. For the equity market, this means the growth trajectory of AI-related assets has been pulled a bit forward; for the crypto world, the expectation that “AI infrastructure is cheaper and more widespread” will be interpreted as a renewed appetite for technology and high-beta assets, providing a foundational storyline for ongoing financial games surrounding AI narratives.
AI Computing Deflation: Warming Tech Stocks and the Crypto AI Magnifying Glass
When Lisa Su threw the Helios mass production signal to the market on August 5 and emphasized that compared to NVIDIA's latest rack, it could boost the "inference tokens per dollar" by up to 30%, equity investors received a clear “AI computing deflation curve.” This means the marginal cost of computing power in AI data centers is overall pushed down, and AI cloud providers can use lower capital expenditures to build larger models and longer application lines without sacrificing performance, quietly adjusting the profit assumptions for tech and growth sectors and resetting valuation tolerances. Historically, when sentiment in the U.S. tech and growth sectors warms, BTC and ETH often exhibit some correlation with high-beta indices like the Nasdaq. The competitive advantage of Helios in terms of computing efficiency essentially raises the baseline for risk appetite in global risk assets, for on-chain assets, this resembles a unified adjustment of “external beta.”
However, within the crypto world, the magnifying glass falls on public chains, Rollups, and AI-themed tokens that tout AI narratives. Over the past few years, whenever there are favorable catalysts like new model releases or chip iterations in the traditional AI world, this group of “AI-friendly” infrastructure and tokens tends to outperform the market— their sensitivity to AI capital spending cycles is far greater than that of macro primary assets like BTC and ETH. As Helios makes “dollar computing efficiency” the battleground for competition and as the market begins to bet on an extended AI boom cycle, the rotation mechanism of crypto funds will also be activated— parts of the risk budget may shift from “neutral positions” like BTC and ETH to higher-beta AI sectors, amplifying the macro tech narrative into on-chain thematic trading and relative yield betting.
Data Centers Compete for Electricity: Natural Gas Pipeline Expansion and Miner Electricity Price Games
As Helios lowers AI computing costs, what is genuinely pushed up is the "willingness to pay for electricity." According to a report from Bloomberg (cited by TechFlow), electricity demand from AI data centers continues to rise, becoming an important factor driving gas power generation and natural gas delivery demand in the U.S. For natural gas pipeline operators, this means they can lock in a steadier and more substantial cash flow channel using long-term electricity purchase contracts and the high load factor of AI customers. Thomas Long, co-CEO of Energy Transfer, mentioned in an analyst conference call that the company is advancing multiple natural gas pipeline projects to meet the incremental demand brought by gas power generation and expects to announce more projects in the coming months—pipeline expansions directly rewrite market expectations regarding the U.S.'s medium- to long-term gas supply capabilities and electricity price paths.
For Bitcoin miners, this curve is their lifeline. The cost of gas power generation is a significant component of electricity prices in certain areas; when pipeline expansions improve gas supply and compress regional price differentials, it can smooth electricity price fluctuations, reduce the risk of power outages in extreme scenarios, and enhance the predictability of “computing factories.” However, on the other side, AI data centers sit at the front of the load curve, locking in electricity and gas resources with a higher willingness to pay, pushing mining operations towards the fringes of the same electricity grid and gas source. The result is that Bitcoin miner costs per kilowatt-hour and shutdown price ranges will no longer simply be dictated by “where the cheap electricity is” but will be priced based on “how much AI data centers are willing to pay for that electricity,” subsequently reshaping the geographic distribution of computing power between the U.S. and overseas, and the risk premium structure of BTC under different cost curves.
Easing in the Strait of Hormuz: Falling Energy Risk Premiums Support Risk Assets
As electricity is repriced by AI data centers, the tail risks in offshore energy transportation are also being compressed. The U.S., Iran, and Oman are nearing a 60-day temporary agreement concerning the Strait of Hormuz (according to a single source), allowing vessels entering the Persian Gulf to transit through Iranian waterways, while those leaving Hormuz towards the Arabian Sea will use Omani waters. For the initial 60 days, no transit or transportation fees will be charged, and they have agreed to clear naval mines in the area to restore safe passage. The Strait of Hormuz itself is a vital choke point for global crude oil and liquefied natural gas, and its safety directly reflects on oil and gas prices and volatility. Such temporary arrangements, even for a limited time, are sufficient to shave off part of the "supply disruption" geopolitical premium in pricing, thereby alleviating the market's worst expectations for future energy prices and inflation.
The combination of macro variables thus experiences a slight adjustment: on one side, the decrease in tail risks in the Strait of Hormuz compresses energy risk premiums, easing the pressure of high oil and gas prices on inflation expectations and nominal interest rates; on the other side, the mass production of Helios and the improvement in dollar computing efficiency strengthen the narrative of “continued expansion in AI capital spending.” For crypto assets, this resembles a “relatively mild inflation + high-tech capital spending” environment—energy prices no longer threaten macro stability in the short term, yet still support the expansion of electricity infrastructure, thereby increasing market tolerance for high-beta tech assets. Historically, BTC and ETH have shown some correlation with such assets, and when augmented by the increased sensitivity of on-chain projects surrounding AI narratives to external good news, their appeal compared to traditional safe-haven assets is repriced during this window of opportunity.
Three Key Curves: Computing, Energy, and BTC Risk Premium
When Helios mass-produced, and the efficiency of inference tokens per dollar could theoretically increase by up to 30% compared to NVIDIA's latest racks, the marginal supply curve of AI computing is pushed downward; concurrently, electricity demand from AI data centers drives gas power generation, and U.S. pipeline operators accelerate natural gas delivery project expansions, attempting to flatten electricity and gas price paths in the near future; if the 60-day temporary agreement in the Strait of Hormuz is smoothly executed, it reduces the geopolitical risk premium on energy transportation channels. The convergence of these three curves means that the risk premium of BTC and ETH is no longer simply “the tale of interest rates,” but is jointly driven by the interest rate path, energy costs, and the risk appetite for tech stocks. Investors need to recalibrate their correlations with U.S. tech sectors and energy sectors, especially as on-chain assets revolving around AI narratives are highly sensitive to external positives. Looking forward, the key variables determining the shapes of these three curves are relatively clear: first, how AMD and NVIDIA’s supply rhythms and pricing games for AI rack products continue to drive or slow down the deflation of computing costs; second, the trends in U.S. electricity and natural gas prices and their feedback on Bitcoin miners' profit margins and the pace of computing expansion; third, the execution details of the Strait of Hormuz's temporary agreement and whether potential geopolitical escalation risks will raise energy risk premiums again. Until these variables provide clearer direction, the pricing of BTC and ETH will continue to search for equilibrium within the new coordinate system of interest rates, energy, and technology.
Join our community to discuss and grow stronger together!
AiCoin Exclusive Hyperliquid Benefits: https://app.hyperliquid.xyz/join/AICOIN88
AiCoin Exclusive Aster Benefits: https://www.asterdex.com/zh-CN/referral/9C50e2
On-chain Telegram community: https://t.me/AiCoinWhaleData
On-chain community: https://www.aicoin.com/link/chat?cid=N6OVMor5g
AiCoin On-chain Twitter: https://x.com/aicoinwhaledata
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
