On January 15, 2026, Gartner published its latest worldwide AI spending forecast: $2.52 trillion in 2026, a 44% jump from $1.76 trillion in 2025, climbing further to $3.34 trillion in 2027. That figure rolls up compute, services, software, models, and security into one number — the most complete single baseline for anyone making AI product or procurement decisions this year.
The same forecast places AI squarely in the Trough of Disillusionment on Gartner’s Hype Cycle for all of 2026 — the phase where experiments and deployments fail to deliver on promises and expectations cool. The money keeps burning, but how it gets bought is changing shape. Set against our opening outlook on AI in 2026, this forecast supplies the quantitative floor under that story.
Where the Money Goes: Segment by Segment
- AI infrastructure: $1.37 trillion in 2026 (from $965 billion in 2025) — more than half the total on its own, heading to $1.75 trillion in 2027
- AI services: $588.6 billion (from $439.4 billion)
- AI software: $452.5 billion (from $283.1 billion)
- AI cybersecurity: $51.3 billion, nearly double the $25.9 billion spent in 2025
- AI models: $26.4 billion; data science and ML platforms: $31.1 billion
Spending on AI-optimized servers alone will grow 49% in 2026 and account for 17% of total AI spending. Citing Gartner data, ITPro also notes that AI-related semiconductors — processors, high-bandwidth memory, and networking components — already represented nearly a third of all semiconductor sales in 2025. The upstream of the stack has been repriced around AI for a while.
Buying Logic Inside the Trough of Disillusionment
Gartner distinguished VP analyst John-David Lovelock puts it bluntly: because AI sits in the Trough of Disillusionment throughout 2026, it “will most often be sold to enterprises by their incumbent software provider rather than bought as part of a new moonshot project.” His corollary: improved predictability of ROI must come before AI can truly scale inside the enterprise.
ITPro’s coverage adds a sobering datapoint from MIT research: in 2025, 95% of surveyed organizations reported zero ROI on their generative AI projects. Read together, the infrastructure-heavy mix makes sense — the foundation is treated as an investment you make regardless of which application wins, while application-level returns are still unproven. Gartner also cautions that AI adoption is shaped fundamentally by the readiness of human capital and organizational processes, not merely by the dollars invested.
The Cybersecurity Doubling
AI cybersecurity is the fastest-growing line in the forecast, doubling from $25.9 billion to $51.3 billion. That tracks the reality of models going into production at scale: prompt injection defenses, data leak prevention, and output compliance are shifting from afterthought patches to budgeted line items. Once AI is wired into existing workflows, protection costs ride along on the same purchase order — a clear tailwind for security vendors.
What It Means for Developers and Procurement Teams
First, the rising share of infrastructure and security budget increases the leverage of platform and tooling teams inside their companies; plan headcount accordingly. Second, “sold, not bought” means most AI capabilities will arrive as upgrades from your incumbent SaaS vendors — so redraw the build-vs-buy line by auditing those vendors’ AI roadmaps before pitching anything custom. Third, proven ROI is now the gate to scale: start with measurable, small-scope deployments, accumulate the numbers, then argue for expansion. For teams with limited resources, that discipline is an advantage rather than a constraint.
Sources
- Gartner Says Worldwide AI Spending Will Total $2.5 Trillion in 2026 — Gartner
- Companies continue to splash out on AI, despite disillusionment — ITPro
- Global AI spending will total $2.5 trillion in 2026, says Gartner — Process Excellence Network
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.