DevOps engineers build and run the systems that get software from a developer's laptop into production: build and release pipelines, cloud infrastructure written as code, monitoring and incident response. A lot of that work is code and configuration, which AI coding tools now write well, so people in DevOps, cloud and platform engineering ask whether AI will take it over. This page answers with the official 2025 to 2035 projections from the US Bureau of Labor Statistics (BLS), O*NET job data and the International Labour Organization (ILO) index of exposure to generative AI.
The short answer: not on current projections, and DevOps sits on the receiving side of the shift. The BLS has no separate DevOps occupation. O*NET lists "DevOps Engineer" as a reported job title for software developers, projected to grow 10.2 percent from 2025 to 2035, almost three times the 3.5 percent for all jobs. And when the BLS explains why network and systems administrator jobs are projected to fall 4.1 percent, it names two reasons: their duties are increasingly done by DevOps engineers and similar roles, and routine tasks are increasingly automated with tools or AI.
What the BLS projects for the jobs DevOps work falls under
| Job (US SOC) | US jobs 2025 | Projected 2035 | Change | Openings a year | Median pay 2025 |
|---|---|---|---|---|---|
| Software developers (15-1252), where O*NET lists DevOps Engineer | 1,717,800 | 1,892,600 | +10.2% | 95,300 | $135,980 |
| Computer occupations, all other (15-1299), incl. computer systems engineers/architects | 471,200 | 495,300 | +5.1% | 27,000 | $116,580 |
| Information security analysts (15-1212) | 192,900 | 233,400 | +21.0% | 14,100 | $129,180 |
| Computer network architects (15-1241) | 181,800 | 195,800 | +7.7% | 9,600 | $134,050 |
| Computer systems analysts (15-1211) | 544,400 | 587,200 | +7.9% | 32,900 | $105,850 |
| Network and computer systems administrators (15-1244) | 323,600 | 310,400 | -4.1% | 13,400 | $99,130 |
| Computer occupations (15-1200) | 5,036,200 | 5,323,800 | +5.7% | 280,000 | $109,470 |
| All occupations | 170,280,800 | 176,198,500 | +3.5% | 17,461,800 | $50,980 |
Source: BLS Employment Projections, occupational projections and worker characteristics, 2025 to 2035. Median pay is the 2025 annual median. Job titles map to occupations through O*NET: O*NET lists DevOps Engineer (Development Operations Engineer) among the reported titles for software developers, alongside Infrastructure Engineer and Systems Engineer, while Infrastructure Engineer, Solutions Architect and Systems Engineer also appear under computer systems engineers/architects, part of computer occupations, all other. The BLS gives no pay figure for DevOps alone; the occupations it falls under have medians of $116,580 to $135,980, against $50,980 for all workers.
The BLS names DevOps as where sysadmin work is going
In its table of factors behind each projection, the BLS says the share of network and computer systems administrators "decreases as job duties are increasingly performed by DevOps engineers and similar roles, including software developers, and as workers increasingly use third-party automation tools or artificial intelligence (AI) to automate routine tasks" (BLS Table 1.12, factors affecting occupational utilization). It labels these occupational substitution and productivity change. The BLS Occupational Outlook Handbook says the same in plain words: some sysadmin tasks "are increasingly being done by software developers focused on DevOps (development operations)", some are outsourced to companies that provide networks as a service, and administrators are increasingly automating routine tasks themselves.
So AI appears in the BLS data on this side of IT, but as one of two forces that move work away from the classic sysadmin job: toward DevOps and developer roles, and toward automation. For the sysadmin side of the story, see will AI replace system administrators?
For software developers, where DevOps engineers are counted, the BLS sets no AI factor that lowers demand. The BLS OOH expects strong demand from software development for AI, the Internet of Things, robotics and other automation, and from more spending on software that protects networks and infrastructure. The BLS does name AI for the neighbouring computer programmers: their work "continues to be automated, including with artificial intelligence (AI)", and that job is projected to fall 7.3 percent.
Where the jobs grow: cloud and hosting first
The industry split shows the same move. Software developers grow in almost every large industry; sysadmins shrink almost everywhere except the firms that sell computing as a service:
| Industry (NAICS) | Software developers 2025 | Change to 2035 | Sysadmins 2025 | Change to 2035 |
|---|---|---|---|---|
| Computing infrastructure providers, data processing, web hosting (518000) | 76,400 | +27.0% | 7,300 | +14.3% |
| Computer systems design and related services (541500) | 505,800 | +14.4% | 50,700 | +2.9% |
| Software publishers (513200) | 171,600 | +15.2% | 4,700 | +3.7% |
| Management of companies and enterprises (551000) | 96,000 | +4.6% | 21,800 | -5.8% |
| Manufacturing | 136,000 | +4.1% | 21,100 | -7.9% |
| Insurance carriers (524100) | 57,300 | +4.6% | 9,500 | -5.5% |
| Educational services, state, local and private | 24,300 | +3.6% | 43,100 | -8.5% |
| Government, excluding education and hospitals | 22,500 | +2.9% | 27,400 | -7.1% |
| Employment services (561300) | 36,300 | -1.8% | 7,100 | -11.7% |
| Telecommunications (517000) | 18,400 | -8.2% | 8,400 | -17.4% |
| All industries | 1,717,800 | +10.2% | 323,600 | -4.1% |
Source: BLS National Employment Matrix, 2025 to 2035, for software developers and network and computer systems administrators. The fastest growth for both is in computing infrastructure, data processing and web hosting, the industry that includes cloud providers: +27.0 percent for developers and +14.3 percent for sysadmins. In-house IT at schools, government, insurers, manufacturers and corporate head offices is where sysadmin jobs fall, which fits work moving to cloud platforms, outsourced services and automation.
Will AI replace DevOps and cloud engineers?
AI tools change the work more than they remove it. DevOps already automates by design: infrastructure as code, pipelines and configuration management exist so people stop doing repeated work by hand. AI speeds up writing that automation. What the data show by area:
| DevOps work | What the official data show | How AI tools touch it |
|---|---|---|
| Infrastructure as code and configuration | O*NET lists Ansible, Chef, Puppet, Terraform and AWS CloudFormation among the tools software developers use | Writing and editing config files and templates is text and code work AI assistants already draft; a person still owns what reaches production |
| Build and release pipelines (CI/CD) | O*NET lists Jenkins CI, GitLab, Atlassian Bitbucket and Red Hat OpenShift | Pipeline scripts are routine to generate; deciding release gates, rollbacks and risk stays with people |
| Monitoring and incident response | O*NET tasks for systems engineers/architects include monitoring system operation to detect potential problems | Alert triage and log summaries are where tools help most; diagnosing a new failure under pressure and owning the outage do not hand off easily |
| Routine admin: patches, accounts, backups | The BLS says workers increasingly use third-party automation tools or AI to automate routine sysadmin tasks | The most automatable part, and the reason sysadmin jobs are projected to shrink |
| Architecture, security and cost | O*NET tasks include verifying the stability, security and scalability of system architecture | Trade-offs between cost, reliability and security are judgment work tied to a specific business |
Source: tools and tasks from O*NET OnLine, Software Developers (15-1252.00) and Computer Systems Engineers/Architects (15-1299.08); sysadmin quote from BLS Table 1.12. The right column is our reading, not a measured score. The pattern: the parts of DevOps that look like routine administration are the most exposed, as the BLS says for sysadmins; designing systems, owning reliability and security, and responding when something breaks stay with people. Security is a growth area next door, with information security analysts projected at +21.0 percent; see will AI replace cybersecurity jobs?
Will AI replace Kubernetes engineers and SREs?
Neither is a separate job in the official data, so the answer comes from the jobs they are counted under. In the O*NET 31.0 database, Site Reliability Engineer is a job title under software developers (15-1252), computer and information research scientists and validation engineers, while Platform Engineer and Cloud Platform Engineer sit under computer systems engineers/architects (15-1299.08). Kubernetes engineer is not an O*NET title at all: Kubernetes appears as a tool these jobs use. The projections in the first table therefore apply: software developers +10.2 percent to 2035 and computer occupations, all other +5.1 percent, against +3.5 percent for all jobs. The BLS sets no AI factor that lowers demand for software developers.
Kubernetes skills are in demand, per O*NET
The Kubernetes documentation (CC BY 4.0) describes it as "a portable, extensible, open source platform for managing containerized workloads and services that facilitate both declarative configuration and automation". O*NET 31.0 lists Kubernetes for 16 occupations and marks it a hot technology for all of them, meaning employers often name it across job postings. For 8 of them it is also marked in demand, meaning employers often ask for it in postings for that specific job: software developers, web developers, database administrators, information security engineers, computer systems engineers/architects, computer and information research scientists, blockchain engineers and validation engineers. Docker carries the same in demand mark for software developers and systems engineers/architects.
What AI tools change in Kubernetes and SRE work
| Work | What Kubernetes already does | How AI tools touch it |
|---|---|---|
| Manifests and configuration | You describe the desired state of your containers and Kubernetes changes the actual state to match at a controlled rate | Configuration files are text that AI assistants draft quickly; checking what a change does to a live cluster stays with a person |
| Rollouts, restarts and scaling | Automated rollouts and rollbacks, self-healing (restarting and replacing failed containers), load balancing | Much of the routine work was automated before AI; tools add more, the same pressure the BLS names for sysadmin tasks |
| Incident response and on-call | Health checks keep containers that do not respond away from traffic | Alert triage and log summaries are where tools help most; diagnosing a new failure and deciding to roll back stay with the on-call engineer |
| Reliability targets, capacity and cost | Fits containers onto nodes based on the CPU and memory you set for each | Choosing how reliable a service must be and what that may cost is a business trade-off, judgment work |
| Secrets and cluster security | Stores passwords, tokens and keys, and updates them without rebuilding images | Security review of clusters is growing work; O*NET marks Kubernetes in demand for information security engineers |
Source: features from the Kubernetes overview; titles and technology marks from the O*NET 31.0 database. The right column is our reading, not a measured score. Kubernetes is itself automation, so the job was never about doing tasks by hand: it is about running automated systems safely. What AI takes over first is the routine layer (writing standard configuration, first-pass triage); owning reliability, upgrades without downtime, security and the call at 3 a.m. stays with people. For security work, see will AI replace cybersecurity jobs?
How exposed DevOps work is to generative AI
The ILO Working Paper 140 (2025) scores each occupation's tasks from 0 to 1 for how much generative AI could do them, and groups occupations from not exposed up to gradient 4, the most exposed. The index has no DevOps code; the groups closest to the job score as follows (how to read the ILO scores):
| Occupation (ISCO-08) | ILO band | Mean task score (0 to 1) |
|---|---|---|
| Applications Programmers (2514) | Gradient 3 | 0.57 |
| Systems Administrators (2522) | Gradient 3 | 0.57 |
| Software Developers (2512) | Gradient 3 | 0.53 |
| Computer Network Professionals (2523) | Gradient 3 | 0.52 |
| Systems Analysts (2511) | Gradient 2 | 0.49 |
| Database and Network Professionals n.e.c. (2529) | Gradient 2 | 0.49 |
| ICT Service Managers (1330) | Gradient 2 | 0.44 |
| Computer Network and Systems Technicians (3513) | Gradient 2 | 0.43 |
Source: ILO 2025 index, ISCO-08 codes. Software and systems groups sit in gradients 2 and 3, the middle to upper end of exposure for professional jobs; none reaches gradient 4, where web and multimedia developers (0.60) and clerical jobs such as data entry (0.70) sit. Exposure measures how much of the task list AI could help with, not how many jobs disappear: software developers are exposed and still among the US occupations projected to add the most jobs.
What this means if you work in DevOps
- DevOps gains work in the official data. The BLS names DevOps engineers as a destination for sysadmin duties, and software developers, where O*NET lists DevOps Engineer, add about 174,700 jobs by 2035.
- Routine administration is the exposed part. Patching, accounts, backups and repeat fixes are what the BLS says is being automated, including with AI.
- Skills that move with the work: writing and reviewing infrastructure code, platform and reliability engineering, security, cost control, and running AI systems in production.
- Where growth is: cloud, hosting and data processing grow fastest; in-house IT in education, government and telecommunications shrinks.
For the wider field, see will AI replace IT jobs? and will AI replace STEM jobs? Projections are estimates, not guarantees, and they cover all causes of change, not AI alone.
Questions and answers
Will AI replace DevOps engineers?
Not on current projections. The BLS has no separate DevOps occupation, but O*NET lists DevOps Engineer as a job title for software developers, projected to grow 10.2 percent from 2025 to 2035, against 3.5 percent for all jobs. The BLS also says sysadmin duties are increasingly done by DevOps engineers.
Is DevOps safe from AI?
No job is untouched, but DevOps is on the gaining side of the official data. AI tools write configuration and pipeline code faster; the BLS names AI and automation as reasons routine sysadmin work shrinks, while system design, reliability, security and incident response stay with people.
Will AI replace cloud engineers?
The BLS projects the fastest growth for both software developers (+27.0 percent) and sysadmins (+14.3 percent) in computing infrastructure, data processing and web hosting, the industry that includes cloud providers. Cloud engineering roles map to software developers or computer systems engineers/architects, both projected to grow.
Will AI replace Kubernetes engineers?
Not on current data. Kubernetes engineer is not an O*NET job title; Kubernetes appears as a tool, marked a hot technology for 16 occupations and in demand in job postings for 8, including software developers (projected +10.2 percent to 2035) and computer systems engineers/architects. AI tools draft configuration; running clusters safely stays with people.
Will AI replace site reliability engineers (SRE)?
Not on current projections. O*NET 31.0 lists Site Reliability Engineer as a job title under software developers, projected by the BLS to grow 10.2 percent from 2025 to 2035, against 3.5 percent for all jobs. AI helps with alert triage and log summaries; diagnosing new failures and owning reliability stay with people.
Is DevOps replacing system administrators?
In part, according to the BLS. It projects network and computer systems administrator jobs to fall 4.1 percent to 2035 because duties are increasingly performed by DevOps engineers and similar roles, including software developers, and because routine tasks are automated with tools or AI.
How much do DevOps engineers earn?
The BLS gives no pay figure for DevOps alone. Software developers, where O*NET lists DevOps Engineer, had a median annual wage of $135,980 in 2025; network and computer systems administrators $99,130; the median for all workers was $50,980.