Crypto Sector Redirects Capacity Toward AI
Crypto Sector Redirects Capacity Toward AI
Anonymous
Key Points
1
Bittensor’s Covenant-72B language model was trained across more than 70 independent nodes in a permissionless environment, which the article presents as evidence that complex machine-learning workflows can operate without relying on a centralized cloud provider.
2
Orbs’ Agentic protocol uses a cosigned oracle framework to check an autonomous AI agent’s proposed DeFi transaction against predefined risk parameters before the transaction is settled on-chain.
3
TeraWulf’s AI-hosting model allows customers to bring their own Nvidia H100 or H200 GPU clusters, while TeraWulf supplies power, cooling and connectivity; its facilities are being equipped with features such as liquid cooling, dual power supplies and N+1-redundant cooling systems.
4
Google owns a 14% stake in TeraWulf and provided a $3.2 billion backstop, according to the article, adding strategic and financial support beyond TeraWulf’s reported $3.7 billion FluidStack agreement.
5
Tom Lee’s claim that crypto will be a “huge winner” from AI was presented without a timeline, specific assets or financial mechanisms, and without figures or cited research measuring how much AI-related activity currently flows through crypto markets.
Crypto and artificial intelligence are merging into a new business model. Bitcoin miners are pivoting to host AI servers, while blockchain projects build systems for autonomous trading and decentralized computing. Crypto Insider reports that TeraWulf earned $21 million from high-performance computing in Q1 2026—more than double its $13 million from bitcoin mining. This shift could make crypto companies crucial suppliers of AI infrastructure, but it may also weaken Bitcoin's security by moving computing power away from mining.
The convergence carries both promise and risk. Investors like BlackRock see AI agents driving demand for stablecoins and tokenized computing power. But critics worry that concentrating energy, computing, and financial control in crypto hands could raise costs and undermine democratic accountability. Whether crypto becomes a major player in AI's economy remains speculative—no major AI platforms have yet adopted crypto payments at scale.
Bitcoin Miners Shift to Hosting AI Chip Clusters
TeraWulf is leading the charge away from pure mining. The company signed a $3.7 billion agreement called FluidStack to host customer-owned Nvidia H100 and H200 GPU clusters at its nuclear- and hydroelectric-powered facilities. Crypto Insider reports the business generated $21 million in Q1 2026—outpacing the $13 million from bitcoin mining. Google, which owns 14% of TeraWulf, backed the pivot with a $3.2 billion commitment.
TeraWulf's data centers are designed for heavy computing. The facilities feature liquid cooling, dual power supplies, and redundant backup systems to keep GPUs running 24/7. By supplying only power, cooling, and connectivity—while customers own the chips—TeraWulf reduced upfront costs and risk. The model appeals to AI startups and labs that need reliable, energy-efficient hosting without building their own infrastructure.
Decentralized AI Training and Autonomous Trading Agents
Two crypto projects promise to decentralize AI workflows without cloud providers. Bittensor trained a Covenant-72B language model across more than 70 independent nodes in a permissionless network—no single company controlled the process. Cointelegraph highlighted this as proof that complex machine-learning can work on blockchain infrastructure. The result was a functional AI model, though whether it matches commercial competitors remains unclear.
Orbs built an automation layer for autonomous AI trading. Its Agentic protocol uses a cosigned oracle to verify an AI agent's proposed DeFi transaction against risk rules before it settles on-chain. This safeguard prevents a rogue agent from draining a fund or executing harmful trades. BlackRock noted that AI agents could generate substantial demand for stablecoins and tokenized payment rails, though no major AI developers have yet integrated these systems.
Crypto Could Capture AI's Financial Activity—or It Could Not
Investor Tom Lee argues that crypto will be a "huge winner" from AI's growth, claiming the sector could capture much of the financial activity generated by intelligent systems. Crypto Economy cited Lee's view as a bullish case for crypto. However, Lee provided no timeline, no specific assets, and no data measuring how much AI activity currently flows through crypto markets or stablecoins today.
The reality is far more uncertain. Few AI companies use crypto for payments or treasury management. BlackRock and other major investors see *potential* in tokenized computing and stablecoin rails, but potential is not adoption. If AI firms continue using traditional banks and cloud providers, crypto's role will shrink, not grow. The pivot by miners toward AI hosting could succeed as infrastructure, regardless of whether crypto itself becomes integral to AI finance.
Bitcoin Security Could Weaken as Mining Power Shifts
Moving computing away from Bitcoin mining poses a network risk. As more hashpower pivots to GPU hosting for AI, fewer computers secure Bitcoin's blockchain. However, Bitcoin's difficulty adjustment—a built-in rule that recalibrates mining complexity every two weeks—softens the blow. Lower total hashpower triggers lower difficulty, making mining more profitable for remaining participants and incentivizing new entrants.
Critics worry about concentrated control. If a handful of companies like TeraWulf own most AI hosting infrastructure, they also control energy, computing capacity, and increasingly, financial rails. American Banker has reported on regulatory scrutiny of crypto market structure. Concentrated power over critical infrastructure could raise costs for users, threaten gig workers in less-developed regions, and complicate democratic oversight of AI systems integral to the economy.
Sentiment
Mostly PositiveNEGATIVENEUTRALPOSITIVE
Publishers
16
Articles
1
Reach
17
A
AnonymousAuthor
