Drooid Logo
Back to story perspectives

Full Breakdown

Agentic AI Spurs Infrastructure, Finance, and Security Overhaul

6/11/2026, 3:05:44 AM

Core Event: Agentic AI Gains Momentum Across Sectors

Recent webinars and a BU construction symposium, together with Ripple’s XRPL AI Starter Kit, show AI agents moving from research to production in health care, construction, and payments.

Background: From Assistants to Autonomous Agents

Large-language models now enable reasoning, transaction execution, and self-replication, as a University of Toronto worm demonstrates by discovering and exploiting vulnerabilities without human input.

Key Investments & Capacity

  • SK Telecom plans a gigawatt-scale AI cloud in Korea, first factory due 2027.
  • Crusoe has contracted 4.9 GW of AI-optimized data-center capacity (part of a 40 GW pipeline), Paymove raised €2.12 million for agentic payments, and Ripple added payment capability.

Official Statements

NVIDIA CEO Jensen Huang called AI clouds “the backbone of new AI clouds.” SK Telecom Chairman Chey Tae-won said the NVIDIA partnership will “tackle GPU, memory and energy challenges.” Ripple highlighted the XRP Ledger’s deterministic finality and low, predictable costs for autonomous commerce. Paymove CEO Piotr Mazur described the platform as a “cohesive payment ecosystem for partners seeking straightforward, cost-effective solutions.”

Criticism & Opposition

The Toronto worm shows open-weight models can generate novel exploits, raising containment concerns. OpenZeppelin co-founder Manuel Aráoz warned AI coding agents make smart-contract security “too asymmetric.” Oklahoma Ethics Commission director Lee Anne Bruce Boone noted no disclosure rules for AI-generated political ads. Marketing analysts warned that “AI slop” erodes consumer trust when brands rely on fully automated content.

Conflicting Reports & Gaps

Mercer surveys find 53 % of executives deem AI ROI “too early,” while Tom’s Hardware reports mixed evidence on AI-driven layoffs.

Quotes

  • “University of Toronto researchers have built and tested a proof-of-concept AI-driven computer worm that uses a locally hosted open-weight large language model to reason its way through a network, generate tailored attack strategies for each target it encounters, and replicate itself, all without human intervention and without touching a commercial AI service.” — Nicolas Papernot, University of Toronto
  • “The worm identified an average of 31.3 vulnerabilities and gained elevated access on 23.1 hosts, roughly three-quarters of the hosts it actively targeted.” — University of Toronto research team
  • “AI agents are beginning to transact, pay for services, and settle value autonomously, creating demand for financial infrastructure built for machine-to-machine commerce.” — Ripple, XRPL AI Starter Kit
  • “Coding agents are superhuman at finding vulnerabilities, and smart contract security is too asymmetric: defenders need to fix every bug while attackers need just one exploit to steal funds.” — Manuel Aráoz, OpenZeppelin

What’s Next

SK Telecom’s AI factory launches in 2027; Nebius admits its first cohort in September 2026; the Oklahoma Ethics Commission drafts AI-ad disclosure rules; regulators monitor autonomous malware. The convergence of agentic AI, dedicated compute, and emerging policy will shape the next digital transformation phase.