- CIOsurge
- Posts
- đź’¸ AWS raises the price of guaranteed GPU capacity
đź’¸ AWS raises the price of guaranteed GPU capacity
AI Security’s Identity Gap, GPU Scarcity Reshapes AI Budgets, CIO Pressure Points for 2026
Aimply Team & Zack Tembi
January 12, 2026
Non-Human Identity Risks Emerge, Salesforce Tightens Data Access, FinOps Clarifies Multicloud Commitments
Powered by Single Fin
Welcome to this week’s edition of CIOsurge!
This week:
From my conversation with Michael Silva, AI security is breaking down in the middle layer, where identity connects intent to action and remains the least mature control surface.
AWS’s price hike on guaranteed GPU capacity signals that predictable AI infrastructure is becoming a premium resource, forcing CIOs to rethink prioritization and budgeting under volatility.
Looking ahead to 2026, CIOs face mounting pressure to govern agent sprawl, prove AI ROI with evidence, and operationalize compliance without slowing execution.
Let’s make this week a game-changer.
Stay sharp. Stay ahead.
🥂 You’re Invited, AI and Identity Forum
I’m hosting a CIOsurge Technology Executive Forum in NYC on AI, identity, and the new attack surface. No presentations, no sales pitches, just a good room of senior IT leaders comparing notes on AI-native security, non-human identity exposure, and what defending at machine speed actually requires.
If this is already on your 2026 roadmap, I’d love to have you there. Space is limited and registration is subject to approval.
đź“… Date: Thursday, January 22, 2026 | 5:00 PM - 8:00 PM
đź“Ś Location: Bar Jamon, New York, NY | 125 E 17th St, New York, NY 10003
đź”— Link: Request an invitation here
đź’ˇ Guest Expert Insights: Michael Silva
⏳ AI Security Has a Middle-Layer Problem
In my conversation with Michael Silva, he also shared a mental model that resonated immediately. He described AI security like an hourglass. At the top, teams focus on prompts and user interaction. At the bottom, they focus on data protection, DLP, and governance.
The gap is the middle. Identity and access is what connects intent to action, yet it is often the least mature layer. That would be like investing heavily in application security and data controls while skipping APIs and network paths entirely.
Once you see it that way, the risk becomes obvious. Without strong identity controls in the middle, prompt safety and data protections cannot fully compensate.
đź’¸ AWS raises the price of guaranteed GPU capacity
AWS increased EC2 Capacity Blocks for ML pricing by roughly 15%, reflecting surging demand for scarce Nvidia GPUs. While the hike affects a narrow slice of cloud spend, it hits a strategic layer of enterprise AI infrastructure, reshaping how CIOs plan for guaranteed GPU access amid ongoing supply constraints.
This is a reminder that AI infrastructure economics are no longer governed by standard cloud pricing logic. Guaranteed GPU access is becoming a premium control point, not a commodity. If your AI roadmap depends on predictable training capacity, that predictability now comes with a visible scarcity tax.
For CIOs, the real risk isn’t the immediate cost bump—it’s budgeting under volatility. Capacity Blocks won’t trigger mass migrations, but they should force tougher conversations about workload prioritization, alternative accelerators, and whether every AI initiative truly needs top-tier GPUs on demand.
- Zack Tembi
🧠Four CIO pressure points heading into 2026
CIO Dive flags four 2026 priorities: managing AI agent sprawl with orchestration/governance platforms, accelerating upskilling so tools drive real value, tracking a growing patchwork of state AI rules despite federal pushback, and proving ROI fast as boards demand measurable outcomes under tighter budgets.
Agent sprawl is the sleeper risk here. If we don’t build an inventory, ownership model, cost controls, and observability early, we’ll recreate shadow IT—except this time it can take actions. Treat agents like production services: lifecycle management, access boundaries, telemetry, and decommissioning are non-negotiable.
On ROI and regulation, the connective tissue is evidence. If we can’t show which use cases moved KPIs, what data trained or powered them, and how controls were applied, we’ll lose air cover with the CEO and the board. 2026 is the year we operationalize AI the way we operationalized cloud: governance that enables speed, not paperwork that blocks it.
- Zack Tembi
🗞️ At A Glance
đź’ˇ CIO Spotlights
Trafigura creates CIO role, appoints Jane Kilmartin to lead global tech strategy
Trafigura named Jane Kilmartin as its first-ever CIO, tasking her with leading group-wide technology and digital transformation from Geneva.
Kilmartin brings deep commodities and energy IT experience, most recently as Group CIO at Alpiq and previously at Cargill and RWE.
She will oversee Trading IT, data, risk, infrastructure, and security as Trafigura pushes toward a more digital, data-driven operating model.
Silicon Labs brings in Ian Dawson as CISO to scale security-by-design
Ian Dawson joins Silicon Labs as CISO, tasked with leading global cybersecurity strategy, IP protection, and product security.
Dawson brings 20+ years of security leadership from Lumen and Charles Schwab, spanning cloud, DevSecOps, and AI governance.
The hire signals a push to harden cyber resilience as Silicon Labs scales its IoT and wireless product portfolio.
Reply