Why Parabolic AI Stock Moves Are Backed by Real Demand, Not Bubble Dynamics

The Parabolic Rise in AI Stocks Has Fundamental Support

Investor Jordi Visser, a veteran macro strategist with over three decades of experience, argues that the dramatic parabolic moves in AI-related stocks are justified by unprecedented real-world demand rather than speculative froth. In a recent conversation with Anthony Pompliano, Visser explained why comparisons to historical bubbles miss the mark on what's driving today's market dynamics.

The shift from AI inference to AI action has created an explosion in token usage and computational demand. Visser traces the current surge back to approximately one year ago when companies began reporting massive increases in inference demand as AI models moved beyond simple chat responses to reasoning-based outputs. The release of Claude's Opus 4.5 marked a critical inflection point, introducing agentic AI capabilities that enable multi-step actions rather than passive responses.

"AI agents are with us. They need food and that food is not physical food. It is tokens," Visser noted, highlighting how digital workers operating behind the scenes have driven exponential growth in computational requirements.

Demand Outpacing Supply Across the AI Infrastructure Stack

Anthropic CEO Dario Amodei recently revealed that his company's business didn't just exceed projections by 10x—it grew 80 times faster than even their most optimistic scenarios. This massive demand shortfall has created supply constraints across the entire AI infrastructure ecosystem, from memory and chips to optical fiber and power generation.

Visser points to specific examples: DRAM prices began climbing in September due to inference memory needs, and companies like Corning have seen their stock prices surge as the market recognized the need for optical fiber infrastructure in next-generation data centers.

The computational shortage has transformed data center capacity into a valuable commodity. Larry Fink recently suggested that compute futures may soon trade as a standardized product, reflecting how power and processing capacity have become the fundamental building blocks of the AI economy.

Bitcoin Miners Pivot to Meet AI Compute Demand

The data center shortage has created unexpected opportunities for Bitcoin mining companies. Hut8 recently signed a nearly $10 billion deal leveraging one of its development sites, exemplifying how crypto miners are converting existing infrastructure to serve AI workloads.

"The linkage between artificial intelligence and crypto" runs deeper than most realize, Visser explained, noting that Bitcoin miners already possess the two critical components of compute: chips and energy.

Global Supply Chains and the DRAM Shortage

Roundup's new DRAM ETF attracted $5 billion in assets within just weeks of launching, demonstrating intense investor appetite for exposure to memory chip manufacturers. The ETF provides U.S. investors access to Asian semiconductor companies like Samsung and SK Hynix that were previously difficult to trade.

Recent trade data shows massive technology imports flowing into the United States to support AI infrastructure buildout, while energy exports help balance the trade deficit. This global supply chain reflects how AI development requires components from manufacturers across Korea, Japan, and other international markets.

Why This Isn't the Dot-Com Bubble

Visser pushes back against bubble comparisons, arguing that demand currently exceeds supply—the opposite of historical manias. Jensen Huang has projected that the AI infrastructure buildout will require $90 trillion over the next decade, representing a fundamental technology replacement cycle comparable to upgrading every legacy system to AI-capable alternatives.

"This is a structural thing," Visser emphasized, noting that today's labor market allows people to maintain employment and continue participating in markets even during volatility, unlike previous crashes that resulted in millions of job losses.

The market volatility that does occur tends to be sharp but short-lived—what Visser calls "speed crashes" rather than multi-year downturns. Stocks have experienced 30-50% drawdowns that recover in weeks rather than years, creating opportunities for investors who embrace rather than flee from volatility.

Tokenization and Crypto's Role in AI Infrastructure

Looking ahead, Visser expects tokenization to become a major theme starting in summer 2025, with significant developments anticipated in July. He's positioned accordingly, buying Bitcoin and Ethereum in anticipation of this convergence between AI agents and cryptocurrency infrastructure.

"AI agents need tokens," he noted, suggesting that the guardrails and speed required for AI systems align closely with blockchain technology developed for crypto applications.

Coinasity's Take

Visser's analysis offers a compelling counter-narrative to bubble warnings by focusing on supply-demand fundamentals rather than chart patterns. The 80x demand surge at Anthropic and rapid asset accumulation in specialized ETFs suggest real structural shortages rather than speculative excess. For crypto investors, the convergence thesis between AI compute needs and blockchain infrastructure—particularly around tokenization and stablecoins—deserves close attention. The Bitcoin miner pivot to AI data centers may represent an early indicator of how digital asset infrastructure becomes critical to the broader AI economy. While parabolic price action naturally triggers caution, the underlying thesis that we're seeing infrastructure buildout for a multi-decade technology transition appears well-supported by corporate capex commitments and demonstrable usage growth.