You’re Locking In 2026 Cloud Waste Right Now

Your finance team is finalizing 2026 cloud commitments. Engineering has Q2 modernization plans. Neither group has compared notes.

By December 31, you’ll have signed Reserved Instance contracts based on today’s VM footprint, right before half those VMs get containerized.

You’ll have budgeted for services you’re planning to retire and under-budgeted for the managed services replacing them. You’ll have an approved “temporary” migration architecture that won’t get optimized until Q3, if ever.

These Q4 decisions are made under pressure, using October data for a February reality that won’t exist by April.

In this edition, I will discuss four such decisions that become 2026 waste because no one goes back to validate them.

You’re committing to 2026 capacity before you know 2026’s architecture

Finance aims to secure 30-40% cloud discounts before year-end. Engineering wants flexibility to refactor in Q2. Both sound reasonable until you’re six months in, when you’re paying for Reserved Instances on VMs you’ve migrated off, while simultaneously paying for the containers you’ve migrated to.

The FinOps Foundation reports that managing commitment-based discounts, such as Reserved Instances and Savings Plans, is now a top priority for cloud teams, precisely because of their potential for both high savings and high waste when architectures evolve.

IDC estimates that 20-30% of cloud spend is wasted on resources that go unused, often prepaid capacity that no longer fits once workloads shift.

What can you do?

If you’re budgeting commitments in Q4 2025, commit only to workloads with stable, 90-day usage patterns.

For everything else, anything touching a 2026 modernization initiative, any system you might containerize, any workload you’re “planning to optimize”, leave it flexible or commit for one year maximum.

The 40% discount looks worse when you’re paying for both the old commitment and the new architecture.

You’re standardizing on premium services because they’re “Already approved”

Three years ago, someone chose Azure SQL Business Critical for the flagship app because it required high performance. Now it’s in the deployment template.

Every new database inherits the premium tier because it has been approved, is in the runbook, and switching requires explaining why you’re deviating from the standard.

That default path is exactly where cost leaks begin. In the HashiCorp survey, 40% of organisations cite “over-provisioning of resources” as a major contributor to cloud waste, and a further 35% point to “idle or under-used resources.

The implication is that the “premium service by default” mindset locks you into paying for yesterday’s architecture, while modern workloads run outside those patterns and at full price.

Instead of a premium-by-default approach, shift your governance to fit-for-purpose tiering and enforce that non-production environments are sized and scheduled based on actual usage, not inherited standards.

What can you do?

Before finalizing your 2026 budget, audit every “default” choice. Require justification for premium tiers. Make cost efficiency the standard, and premium the exception that requires a business case.

You’re auto-renewing premium support contracts you don’t use

Many mid-market companies purchased AWS Business Support or its equivalent from Azure years ago as insurance. The contract auto-renews every Q4. The cost? For a company spending $50,000 monthly on cloud services (~$ 600,000 annually), Business support costs approximately $3,500 per month. At $10 million in annual cloud spend, Enterprise support costs approach $670,000.

The problem is that most mid-market teams file 2-5 support tickets annually. That’s over $ 60,000 per ticket for Enterprise support, far more than hiring external consultants on demand.

Meanwhile, teams rely on internal expertise, community forums, and documentation for 90% of their issues.

Premium support made sense three years ago when the team was smaller and cloud architecture was new.

But your in-house expertise has grown. Your architects know Azure. Your engineers solve most issues without vendor escalation. The $150,000 annual support contract is now funding the capability you’ve built internally.

What can you do?

Before December 31, audit your 2025 support usage. How many tickets did you file? What severity? If you opened fewer than 10 critical cases, calculate your cost-per-ticket.

Then ask: Would that $150,000 be better spent on an additional cloud engineer or on a third-party consulting retainer?

Some providers will offer a 40% discount if you mention you’re downgrading, but only if you ask before the auto-renewal locks you in for another year.

You’re approving “Temporary” architecture that becomes permanent

Q4 migrations come with deadlines. Data centers need to close by December 31. Workloads need to move before the lease expires. So, teams estimate resource sizes conservatively; it’s better to overprovision than underprovision during a high-stakes cutover.

The plan is always to optimize later, once things stabilize.

Later rarely comes. By Q1, engineering is focused on introducing new features, rather than revisiting the sizing decisions made in December. Those “temporary” estimates calcify into permanent line items.

Industry data indicates that roughly 32% of cloud spend is wasted, primarily due to resources that were intended to be temporary, over-provisioned instances selected during migrations, or prototypes that were never shut down.

About 60% of companies use lift-and-shift approaches without re-engineering, often paying 30-40% more than they would for a cloud-native architecture.

What can you do?

Before December 31, tag every resource deployed in Q4 2025 with an expiration or review date.

Set calendar reminders for Q1 2026 to audit November and December deployments.

If something was sized “for migration safety,” schedule the rightsizing review now. Make temporary decisions visible, measurable, and time-bound; otherwise, they’ll become your most expensive, permanent architecture.

Most companies are now making a fifth decision without realizing it: AI budgets. Recent data shows that 80% of companies missed AI cost forecasts by more than 25%. Token-based pricing scales with adoption, so that $5K pilot chatbot can hit $50K monthly by February when usage takes off.

Want to audit your Q4 decisions before they lock in 2026 waste? We’ll walk you through it.