
Tech • IA • Crypto
IBM’s sharp stock drop, new AI regulation proposals, and New York’s data center moratorium highlight intensifying tensions in the race to build and govern AI infrastructure.
IBM shares fell roughly 25%, marking one of the steepest declines in its history after the company reset expectations for its server business. The drop reflects a broader shift in enterprise spending toward GPUs, memory, networking, and hyperscale cloud infrastructure, areas where IBM is less dominant. Despite strong gains in recent years—nearly doubling during the AI boom—the company is now losing share of customer technology budgets as capital flows into AI-heavy hardware ecosystems.
IBM’s business remains split across software (44%, ~80% gross margins), consulting (31%, <30% margins), and infrastructure (23%, ~60% margins). While assets like Red Hat OpenShift position it within AI orchestration, its services-heavy model faces structural limitations including slower growth and pricing pressure. The company’s historic strength in integrated systems and high switching costs has proven less decisive in the fast-moving AI era.
IBM’s dominance dates back to the System/360 mainframe era, built on reliability and deep enterprise lock-in. However, the 1981 IBM PC helped catalyze the rise of Microsoft and Intel, which ultimately captured far greater value. Later strategic pivots toward consulting and integration stabilized the company but limited its exposure to high-growth platform economics now driving AI leaders.
Demis Hassabis, CEO of Google DeepMind and a Nobel laureate, has proposed creating a U.S.-led standards body to evaluate frontier AI models for risks including cybersecurity, biological threats, and autonomous misuse. The plan includes mandatory pre-release testing, ongoing audits, and coordinated slowdowns if severe risks are detected. The proposal reflects growing concern that AI capabilities are advancing faster than governance frameworks.
While the proposal outlines mechanisms such as 30-day pre-release reviews, independent testing, and federal oversight funded by AI firms, questions remain about enforcement—especially for open-source and foreign models. Critics warn that heavy regulation could unintentionally favor large incumbents while slowing innovation or pushing development خارج U.S. oversight.
New York Governor Kathy Hochul has issued a one-year moratorium on new AI data centers requiring 50 megawatts or more, pending environmental and regulatory review. The pause targets concerns over energy consumption, water usage, and grid impact, making New York the first state to impose such a broad restriction.
Industry groups argue the moratorium could divert billions in investment to states like Texas, Virginia, and Georgia, which are actively courting data center projects. Construction leaders warn that delays in this fast-moving sector may permanently shift projects—and associated jobs and tax revenue—out of New York.
Data center development is increasingly constrained by access to power infrastructure, not land. Large-scale projects can require hundreds of megawatts and years-long timelines for substations and grid interconnection. Developers often secure land cheaply but invest heavily in electrical infrastructure before securing tenants such as AWS or other hyperscalers.
Public opposition has expanded beyond resource usage to aesthetics, prompting firms like Gensler to design data centers resembling campuses or museums instead of industrial boxes. Meanwhile, advances such as closed-loop cooling systems and use of non-potable water aim to reduce environmental strain, though skepticism persists in local communities.
The AI boom is reshaping both corporate fortunes and public policy, exposing gaps between technological acceleration, infrastructure capacity, and regulatory readiness.