In the fast-paced world of technology and finance, a new narrative is emerging, suggesting a significant shift in capital and attention. Oracle, a long-standing tech behemoth, recently made headlines with a staggering $300 billion cloud deal tied to OpenAI, a name synonymous with the current AI revolution. Initially, its stock soared, but within two months, the market's enthusiasm cooled, with Oracle shedding that very $300 billion in market value. This dramatic turn has analysts pondering if AI promises are outpacing the cold, hard cash flows needed to sustain them. Yet, on the other side of the spectrum, a three-year-old coding tool startup named Cursor managed to raise $2.3 billion at an eye-watering $29.3 billion valuation, tripling its worth in months, on the back of promises for an AI pair programmer.
These contrasting stories spark a crucial question for investors: When AI can command such immense valuations for nascent companies, does the crypto sector still hold the same appeal for capital, or is it merely being pulled into the same overarching AI trade under a different guise? The landscape appears to be undergoing a profound re-evaluation.
The AI Money Torrent and its Impact
A closer look at the sheer volume of funding flowing into artificial intelligence reveals the underlying sentiment. Global AI startup funding reached approximately $100 billion in 2024, an astonishing 80% increase from the previous year and nearly a third of all venture capital deployed. S&P Global’s figures show generative AI funding alone exceeding $56 billion in 2024, almost doubling the prior year. The Stanford AI Index further highlights this trend, tracking private investment in generative AI at $33.9 billion for 2024, more than eight times the 2022 figures. Moreover, EY estimates that generative AI startups secured another $49.2 billion in just the first half of 2025.
For those familiar with crypto's explosive growth in 2021, this level of capital injection rings a bell. Back then, the hot trades revolved around token issuance, DeFi yields, and metaverse equity. Now, the center of gravity has decidedly shifted. The big checks are funding intensive training runs, massive data centers, and a select group of foundation model labs. Barron’s reports that roughly a third of global venture capital is now directed towards AI giants like xAI, Databricks, Anthropic, and OpenAI. Public companies are also accumulating vast debt to secure GPU capacity; Oracle, for instance, is reportedly lining up $38 billion in bonds for its cloud infrastructure expansion. Nvidia's data center revenue has, in turn, reshaped entire equity indices, demonstrating that if you seek exposure to “future cash flows from compute,” the highest beta now resides in AI infrastructure and foundation models.
This isn't to say that liquidity has vanished from crypto. Instead, it suggests that marginal dollars are now being priced against a formidable new benchmark. If a mid-size AI startup can command a $30 billion valuation, and OpenAI can discuss trillion-dollar capital expenditure plans without eliciting ridicule, the bar for a $10 billion token with limited real-world utility undeniably rises.
AI Tokens and the ASI Experiment
In a logical move, the crypto industry attempted to align itself with this narrative by packaging AI into tokens. The most prominent endeavor was the Artificial Superintelligence Alliance, a plan to merge SingularityNET, Fetch.ai, and Ocean Protocol into a single ASI token, branding the entire ecosystem as decentralized AI. Fetch.ai’s merger blog in 2024 presented a straightforward proposition: one treasury, one token, uniting three projects covering agents, data, and models.
Initially, this strategy gained traction. Billions of dollars worth of AGIX, FET, and OCEAN liquidity converged around this shared vision. Exchanges swiftly listed spot and perpetual pairs for ASI, while retail holders were offered migration bridges and a single token that neatly represented “AI” on their watchlists. It appeared crypto had found a way to condense a complex sector into a manageable derivative.
However, the alliance soon fractured when Ocean Protocol announced its withdrawal in October, seeking to de-peg OCEAN from ASI and re-list it as an independent asset. Ocean cited “voluntary association” as its reason for exiting, though Fetch.ai subsequently initiated legal action, alleging broken promises surrounding the merger and tracing conversions of over 660 million OCEAN to FET.
This governance drama underscores the dynamics of the AI token trade. It often mirrors the private AI boom, but with amplified volatility and notably less revenue. When ASI performed well, participation surged. Yet, as valuations cooled and community politics resurfaced, the “alliance” reverted to being three distinct entities with divergent agendas. From a liquidity perspective, AI tokens often behave less like a separate asset class and more like a conduit for existing crypto capital to shadow developments in private AI. While new funding rounds for Cursor or Anthropic may not directly impact ASI, they significantly influence the emotional tone and market perception, with crypto traders adjusting their AI token baskets accordingly.
From Bitcoin Mines to AI Model Farms
Perhaps the most tangible merger between AI and crypto is occurring within power contracts. Bitcoin miners have spent a decade developing data centers in regions with access to cheap energy, and now, AI hyperscalers are actively competing for the very same megawatt capacity. Bitfarms serves as a clear illustration of this trend. The company has declared its intention to completely phase out Bitcoin mining by 2027, instead re-purposing its infrastructure for AI and high-performance computing.
- Its 18-megawatt facility in Washington state is slated to be the inaugural conversion site.
- The facility will feature racks specifically designed for Nvidia GB300-class servers.
- It will incorporate liquid cooling capable of managing approximately 190 kilowatts per rack.
Bitfarms’ recent press release detailed a fully funded $128 million agreement with a prominent US data center partner. Management boldly claims that a single AI facility could potentially generate more profit than the company’s entire historical Bitcoin mining operations.
Bitfarms is not an isolated case. Iris Energy has rebranded as IREN and is pivoting its hydro-powered sites towards AI data centers, with Bernstein research projecting billions in revenue from Microsoft-backed GPU deployments. Hut 8 openly positions itself as a power-first platform, capable of directing its planned 1,530 megawatts of capacity to whichever workload offers the best returns, with AI and HPC topping the list. Core Scientific pursued this path so aggressively that AI cloud provider CoreWeave proposed a $9 billion all-stock deal to acquire it, aiming to secure over a gigawatt of data center power for Nvidia-heavy clusters, before shareholder resistance derailed the acquisition.
The pattern is consistent: Bitcoin mining provided these firms with access to affordable power, grid connections, and often hard-won permits. Then, AI emerged, offering a significantly higher dollar value per megawatt. For shareholders who have witnessed multiple halvings compress mining margins, channeling energy into GPU stacks represents a clear shift from a maturing carry trade to a robust growth opportunity. This is where the headline, “AI is eating crypto liquidity,” becomes particularly literal for Bitcoin. Every megawatt diverted from SHA-256 to GB300 or H200 is a unit of energy that no longer contributes to securing the Bitcoin network. While hash rate continues to grow through new entrants and hardware upgrades, a progressively larger share of cheap power will be dictated by AI’s burgeoning willingness to pay.
When AI Attacks the Rails: The Security Frontier
Another critical intersection between AI capital and crypto lies in the realm of security. In November, Anthropic released a report detailing what it described as the first large-scale espionage campaign orchestrated by an AI agent. A China-linked group reportedly jailbroke Anthropic’s Claude Code product, leveraging it to automate reconnaissance, exploit development, credential harvesting, and lateral movement across approximately 30 victim organizations. While some attacks succeeded, others failed due to the model hallucinating fake credentials or stealing publicly available documents. The most alarming revelation was that natural language prompts, rather than a room full of human operators, drove most of the attack chain.
Crypto exchanges and custodians find themselves directly within this blast radius. They already incorporate AI into their trading surveillance, customer support, and fraud monitoring systems. As more operations become automated through AI agents, the very tools designed to route orders or monitor for money laundering will become attractive targets. The dense concentration of keys and hot wallets inherent in these platforms makes them highly appealing to any group capable of directing a sophisticated, Claude-sized agent at a network. In the event of a major AI-driven breach at a significant exchange, the subsequent regulatory response would likely view AI and crypto as a single, interconnected risk surface sitting atop critical financial infrastructure, regardless of whether the affected venue trades Nvidia equity, Bitcoin, or both.
Is AI Truly Eating Crypto Liquidity? A Nuanced Perspective
The honest answer is that AI is doing something far more intricate and interesting than simply devouring crypto liquidity. It is fundamentally re-setting the price of risk for anything that touches compute. Venture capital that might once have pursued Layer 1 blockchain projects is now funding foundation models and AI infrastructure. Public equity investors are weighing substantial drawdowns in legacy companies like Oracle against the speculative potential of a $300 billion OpenAI cloud deal. Private markets are comfortable valuing a developer tool like Cursor on par with a mid-cap token network. Bitcoin miners are strategically rebranding as data center operators, securing long-term contracts with hyperscalers. Meanwhile, token projects are actively attempting to append “AI” to their tickers, recognizing where the current market excitement resides.
From within the crypto industry, this market dynamic might indeed appear like a food chain where AI relentlessly consumes everything in its path. However, the reality is always more nuanced and complex. Over the past two years, AI has solidified its position as the benchmark trade for future computing. This dominant narrative inevitably pulls Bitcoin infrastructure, AI tokens, and even the security of exchanges into the same overarching story. Therefore, liquidity isn't vanishing outright; it’s dynamically moving and re-pricing everything else against the one sector that has successfully convinced markets to fund trillion-dollar capital expenditure plans based largely on a promise and compelling demonstration.
Post a Comment