Phemex CEO Federico Variola has argued that artificial intelligence has so far been a “net negative” for the cryptocurrency industry, pointing to competition for capital and a growing cybersecurity burden. Variola said AI is simultaneously drawing liquidity away from crypto and giving attackers more powerful tools, creating pressures that could disproportionately affect smaller protocols and push parts of the industry toward greater centralization.
Speaking on the Chain Reaction podcast published September 15, Variola said AI-assisted social engineering and vulnerability discovery are increasing the resources required to secure crypto infrastructure. His concern is that smaller teams may struggle to absorb rising defensive costs while larger, better-capitalized organizations can invest more heavily in security, potentially weakening some of the economic advantages associated with decentralized development.
AI Competes With Crypto for Capital and Infrastructure
The capital argument extends beyond conventional venture funding. Several Bitcoin mining companies have been repurposing power capacity and data-center infrastructure for AI and high-performance computing, where demand and long-term contracts can offer an alternative revenue source to mining. S&P Global identified IREN, Riot Platforms, Core Scientific, HIVE, Cipher, TeraWulf and other miners among companies increasingly developing AI and HPC operations during 2026.
The shift has continued through large commercial agreements. Hut 8, for example, signed a 15-year AI data-center lease worth $9.8 billion in July, with Reuters noting that several former or existing Bitcoin miners are using power assets and data-center expertise originally developed for cryptocurrency mining to serve AI customers. That supports the broader infrastructure-competition argument, although it does not prove that every dollar entering AI would otherwise have entered crypto.
Variola’s second concern centers on security. He argued that AI can lower the cost of sophisticated attacks by improving phishing, social engineering and vulnerability discovery, forcing exchanges and protocols to spend more on prevention. That risk is not one-sided, however: the same technology can also strengthen defensive monitoring, anomaly detection and software review, making AI a security tool as well as a threat amplifier.
Phemex Is Building Around AI Anyway
Variola’s criticism comes despite Phemex pursuing one of the industry’s more explicit AI strategies. The exchange launched its “AI-Native Revolution” on February 19, committing to embed artificial intelligence into operations, product development and strategic decision-making rather than treating it as a standalone feature. In a subsequent company statement, Variola described AI as a structural shift in how exchanges will operate.
Phemex followed that strategy with its AI Bot on February 24. The tool analyzes factors including recent volatility, historical drawdowns and price structure to recommend trading parameters automatically, while leaving users responsible for deploying the strategy and accepting the resulting trading risk.
The apparent contradiction is narrower than it first appears. Variola is distinguishing between AI as a useful operational and trading tool and what he sees as its broader competitive and security effects on crypto as an industry. He has also argued that AI agents can assist portfolio construction and trading decisions without eliminating the need for human judgment.
The unresolved question is whether those defensive and productivity gains eventually outweigh the capital and security pressures Variola identified. For crypto companies, the measurable indicators will be where infrastructure investment flows, how cybersecurity spending evolves and whether AI-driven automation lowers operating costs enough to offset the new risks it introduces.
