With Crypto and Risk Asset Fears Escalating, Amazon Accelerates Its AI Push

Amazon Accelerates AI Push as Crypto Miners Shift Gears

Amazon is deepening its foray into artificial intelligence with the launch of Trainium 3, a new chip designed to challenge Nvidia’s GPU dominance. Available through Amazon Web Services (AWS), Trainium 3 promises up to four times the training speed of its predecessor while maintaining the same energy footprint.

The chips are integrated into Amazon’s “UltraServers,” with each cluster capable of running 144 units simultaneously. This expansion positions Amazon to compete directly with Google and Nvidia in large-scale AI workloads, including training advanced language models.

Google’s dominance in AI model development—reportedly giving it an 87% chance of producing the leading model by year-end—has triggered what OpenAI CEO Sam Altman described as a “code red,” highlighting the intensifying competition in the sector.

However, building large-scale AI infrastructure is no small feat. High energy and space demands are barriers that few tech giants can solve alone. Crypto miners, with existing large-scale data centers, are stepping in to fill the gap. Firms like Core Scientific, CleanSpark, and Bitfarms are transforming energy-intensive operations into AI-ready facilities, positioning themselves as utility providers for hyperscale AI computing.

IREN, a former bitcoin miner turned neocloud firm, recently surged after securing a $9.7 billion AI cloud deal with Microsoft. TeraWulf similarly signed a $9.5 billion AI infrastructure joint venture with Fluidstack, backed by Google. These companies already control gigawatts of power and data center infrastructure capable of supporting high-density AI clusters.

The market outlook, however, remains cautious. Bitcoin has declined more than 17% over the past 30 days, the CoinDesk 20 index dropped 19.3%, and the NASDAQ 100 fell roughly 1.5% after a larger drawdown. Analysts warn that the AI infrastructure boom may resemble past tech bubbles. OpenAI’s multi-trillion-dollar infrastructure commitments rely heavily on continued high demand. Bain & Co. estimates that a slowdown could create an $800 billion shortfall, requiring $2 trillion in combined annual revenue by 2030 to meet projected AI compute needs.

If demand for AI slows, hybrid operations could face liquidity pressures similar to the 2022 crypto crash, with potential ripple effects across broader risk assets.

For now, crypto miners-turned-AI providers are betting on a new digital gold rush—powered by GPUs rather than ASICs.