Jenn Tejada Warns AI Agent Drift Raises Failure Risk as 2026 Hyperscaler Capex Hits $725 Billion
Updated
Updated · Forbes · Jul 2
Jenn Tejada Warns AI Agent Drift Raises Failure Risk as 2026 Hyperscaler Capex Hits $725 Billion
1 articles · Updated · Forbes · Jul 2
Summary
Jenn Tejada said AI systems moving from experimentation into production are creating harder-to-detect failure modes, especially agent drift that can compound into multiple breakdowns before humans notice.
Early detection is becoming critical because AI failures often surface differently from traditional software outages, pushing AIOps platforms to monitor models alongside broader digital infrastructure.
Tejada said companies will increasingly want systems that watch their agents, judge performance independently and let humans interrupt automated work before small issues become major incidents.
$725 billion in projected 2026 hyperscaler capital spending — nearly double the past year, according to BNP Paribas — is fueling broader AI infrastructure buildouts and jobs beyond San Francisco.
The $7 trillion AI boom is here, but will it create shared prosperity or a new era of digital colonialism?
With AI agents managing critical systems, how do we ensure humans can still pull the plug when things go wrong?
As AI's energy demand strains national power grids, can we innovate our way out of a looming infrastructure crisis?
AI Agent Drift and the 2026 Infrastructure Surge: Enterprise Risks, Governance, and the $725B Challenge
Overview
By mid-2026, the focus in artificial intelligence has shifted from simple adoption to managing its unpredictable nature. AI agent drift—where an AI system gradually and subtly deviates from its original purpose—has become a major enterprise risk. This drift is hard to detect because AI systems are designed to learn and adapt, making changes that are often invisible at first. As a result, organizations must pay close attention to how AI evolves and operates, not just what it produces. Understanding and controlling this drift is now essential to prevent AI from quietly leading businesses in the wrong direction.