I run a small DevOps and cloud management consultancy with six experienced engineers, and I have more than 14 years of experience across enterprise companies and startups. We've delivered multi-cluster and multi-region deployments, automated and gated CI/CD pipelines, compliance support, monitoring systems, and incident-response workflows.
Business was strong throughout the previous two years. I regularly found more work than the team could handle, and our leads came from personal referrals, professional networking, freelance platforms, and our website. Most clients were midsize companies or venture-backed startups.
That changed abruptly this year. After closing several deals per month in 2024 and 2025, I landed only two new clients in March and nothing since. Existing work also dried up during the summer, leaving the company bank account empty and forcing me to consider letting the team go.
When I contacted former clients, many had laid off large parts of their staff and reduced their infrastructure. In some cases, even the people responsible for maintaining systems we built were dismissed without replacements.
Is this happening broadly across DevOps and cloud consulting? I understand that AI is changing the industry, but the drop in demand feels unusually sudden. Is the market shrinking, or should I be repositioning the business toward a different kind of service?
4 Answers
You’re probably not alone, and it may help to separate two trends. Some clients are cutting consultants because of the broader downturn, while others are convinced that AI can replace routine platform work. AI can produce configuration quickly, but it doesn’t remove the need for someone to understand reliability, security, compliance, and the consequences of bad architectural decisions.
The market may be moving toward smaller advisory engagements and cleanup work. Companies that let AI generate their infrastructure will eventually need experienced people to review it, control costs, and make it maintainable. Building relationships and a reputation around those higher-risk problems could be more defensible than competing for generic deployment work.
The decline may be especially severe among venture-backed startups. Many of them spent their funding without reaching the growth they expected, so they’re cutting infrastructure and outside vendors regardless of the quality of the work. That doesn’t necessarily mean your technical offering is obsolete; it means the customers who used to buy it have less money and fewer people.
You might have better results with smaller, clearly defined engagements: architecture reviews, security assessments, migration plans, incident readiness, cloud-cost audits, or advisory sessions for teams building AI-heavy products. A fixed-scope service with a measurable result is easier to approve than an open-ended consulting arrangement.
A lot of organizations may simply be finishing the main cloud and Kubernetes migration cycle. Once the platform is in place, leadership often cuts the consultants who built it and assumes the remaining team can maintain it. AI tools and cheaper contractors reinforce that decision because management thinks they can get most of the way there for less money. The opportunity may shift toward reviewing and repairing systems created by inexperienced teams, rather than building every platform from scratch.
This looks like a combination of tighter budgets and AI raising the minimum capability expected from internal teams. Clients may believe developers can handle basic Terraform, Helm, pipelines, and cloud configuration with AI assistance, even when the resulting systems are fragile. The economy is also making infrastructure consulting an easy expense to postpone. Rather than selling general DevOps capacity, consider packaging specific outcomes such as cloud-cost reduction, observability spend reviews, security hardening, or compliance readiness.
That makes sense. We’ve done observability work before, but it was usually part of a larger project rather than a standalone service. I may need to turn those capabilities into a clearer product instead of presenting ourselves as general-purpose DevOps consultants.
I’m seeing the same thing internally. Teams are underinvesting in infrastructure while expecting developers and AI tools to fill the gap. It works for simple systems, but more complicated environments are already becoming difficult to operate.