Piper Sandler CIO survey finds AI spending ’beyond experimental’

Piper Sandler’s CIO Pulse Survey shows 86% of IT decision-makers are already deploying copilots, agentic AI, or fully autonomous systems, while IT budgets are expected to grow 4.8% in 2026, up 2 percentage points from 2025. Security remains the top spending priority at 73%, and 94% of respondents plan to increase spending on AI infrastructure, with Azure and AWS still leading. The findings are constructive for AI and cloud-related names such as Amazon, Microsoft, Arista Networks, Datadog, Dell, Palo Alto Networks, and Rubrik.

Analysis

The cleanest read-through is that AI capex is still broadening, but the spend is shifting from experimentation to plumbing: cloud, data quality, observability, security, and workflow automation. That tends to favor the “picks-and-shovels” cohort over pure model beneficiaries, with Microsoft and Amazon advantaged by budget share and stickiness, while Arista, Datadog, and Rubrik benefit from the need to make AI infrastructure reliable enough for production workloads. The second-order effect is that enterprise buyers are likely compressing vendor stacks: fewer point solutions survive if AI programs force procurement to consolidate around platforms that can handle identity, data governance, networking, and monitoring end-to-end.

The most important negative is not “AI adoption slows,” but that adoption starts to threaten labor and thus creates a political and budgetary overhang. If 60%+ of IT leaders are already flagging headcount pressure, CIOs will increasingly demand measurable ROI within 2-3 quarters, which favors vendors that can show immediate productivity or cost takeout and punishes names selling optionality alone. That dynamic is supportive for Microsoft and Palo Alto, but it is a subtle headwind for names like Datadog if usage growth does not convert into budget defense as scrutiny on software line items rises.

The underappreciated risk is that the current enthusiasm for AI infrastructure could become self-limiting if data quality becomes the bottleneck faster than compute does. In that case, spend migrates away from GPUs and broad cloud expansion toward governance, security, and data lifecycle tools, which is bullish for PANW/RBRK relative to more cyclical infra beneficiaries. Over the next 6-12 months, the key catalyst is whether the next CIO survey or earnings season shows AI moving from pilot metrics to hard productivity metrics; if not, the market may rotate from multiple expansion into a narrower quality rally.

Key Decisions for Investors

• Go long MSFT vs. a basket of lower-quality software names for 3-6 months; MSFT has the best mix of budget share, AI monetization optionality, and enterprise lock-in, while the downside is cushioned if AI spend broadens but remains ROI-constrained.
• Buy PANW on pullbacks and consider a PANW/RBRK pair only if security budgets re-accelerate faster than data-protection budgets; PANW should benefit sooner from AI governance and threat concerns, with a cleaner path to near-term spend capture.
• Stay long ANET as a 6-12 month infrastructure lever, but size modestly; upside is strongest if AI networking becomes the next constraint, while risk is a capex digestion phase if cloud providers pause spend after current buildout.
• Use DDOG as a relative underweight vs. MSFT or PANW for the next 1-2 quarters; observability should benefit, but it is more exposed if CIOs force vendor consolidation and demand measurable headcount or incident-reduction ROI.
• Initiate a small AMZN long into any AI-infrastructure dip; AWS remains a share-gainer if enterprise spend shifts from experimentation to deployment, but the trade works best with a 6-9 month horizon and tight risk control if cloud growth decelerates.