AI · Labour market · DevOps
The world in 10–20 years: three scenarios instead of a forecast
Nobody knows what the world looks like after AI, robots and drones. What one can do: collect measured numbers, derive scenarios from them and ask what stays the same in all of them – from the view of someone working in IT and DevOps.
First: scenarios, not a forecast
The text separates two things: numbers that were measured and published (with a source) and my interpretation of them. The scenarios below are my reading of the numbers, not a prediction. Whoever claims to know the world of 2040 exactly is selling something.
The big picture: three technologies, three speeds
AI is software and spreads fastest: according to the 2025 DORA report, 90% of surveyed software developers use AI (nearly 5,000 respondents). Robots are hardware and grow more slowly but steadily: the IFR reported on 24 September 2026 a record stock of 5 million industrial robots worldwide, with more than 600,000 newly installed in 2025 (+11%). Drones I leave without numbers: I found no reliable statistic I would stand behind here. Qualitatively, use cases such as inspection, agriculture and logistics are emerging, where regulation mostly sets the pace.
For the labour market overall, the World Economic Forum (Future of Jobs 2025) expects 170 million new and 92 million displaced jobs by 2030, a net +78 million. Nearly 40% of on-the-job skills will change, and 41% of employers plan to reduce staff because AI takes over tasks. This is a survey of companies, not a measurement.
My focus: IT and DevOps – what the numbers say
| Source | Figure | Meaning |
|---|---|---|
| Bitkom (03.09.2026) | 79,000 missing IT specialists (2023: 149,000, 2025: 109,000) | The shortage has nearly halved in three years but has not disappeared. |
| Bitkom (03.09.2026) | 42% of companies use AI in IT; only 10% of those have cut positions because of it, 85% none | So far few cuts; 32% expect cuts in future, 65% do not. In 35% of cases AI replaces vacancies that cannot be filled. |
| Brynjolfsson et al. (Stanford, „Canaries“) | −13% relative employment for 22–25-year-olds in the most AI-exposed occupations | Experienced workers in the same occupations stayed stable. Getting started gets harder, not the whole industry. |
| Acemoglu (NBER w32487) | at most 0.66% (adjusted: under 0.53%) more total factor productivity over 10 years | The cautious counter-voice: the economy-wide effect could be small. |
| DORA 2025 | AI use 90%, linked to higher delivery throughput | According to the report AI acts as a “mirror and a multiplier”: cohesive teams gain, fragmented teams see their weaknesses exposed. |
Three scenarios for IT careers
- Slow: the effect stays small, as Acemoglu considers possible. AI is one more tool next to editor and terminal. The shortage of specialists remains, classic skills (Linux, networking, automation) keep their value.
- Medium (today’s situation, extended): AI is standard in almost every team. Teams need fewer hands for the same work, entry gets harder (the “Canaries” number), the shortage keeps shrinking. Work shifts from writing to reviewing, operating and securing.
- Fast: AI takes over most routine in development and operations, and employers do what they say they plan in the WEF survey. Small teams with a lot of responsibility become the rule, and expertise in platforms, security and regulation counts for more than typing speed.
What stays the same in all scenarios
- Someone runs the system. Delivery, monitoring, backups, recovery: when something fails, a human who understands the context is needed.
- Someone is accountable. The EU AI Act demands documentation, oversight and transparency; that cannot be delegated to a model.
- Good foundations pay off. If AI is an amplifier (DORA), it amplifies clean pipelines, tests and processes just as it amplifies chaos.
What I do with it myself
I use AI daily as a tool and check every result myself, as with PflegeOasis. I deliberately build my practice on the parts that remain in all scenarios: my own server with zero-trust access, CI/CD pipelines, Docker and Linux. Whether the scenario turns out slow, medium or fast I do not know – I can only make sure my skills are needed in each of them.
Germany and the EU
Regulation sets the pace here. The “Digital Omnibus” amendment of the EU AI Act was adopted by the Council on 29 June 2026 and has been in force since July 2026. Obligations for high-risk systems under Annex III now apply from 2 December 2027, under Annex I from 2 August 2028. Transparency duties for generative AI (Article 50) remain; for systems already on the market the marking deadline is 2 December 2026. For operators of AI applications this means documentation and oversight become part of everyday work.
On robots, Germany ranks fifth worldwide according to the IFR, with under 25,000 newly installed units, and accounts for about 41% of installations in the EU. The Bitkom figure shows at the same time: German companies already use AI widely in IT (42%) but have so far hardly cut positions.
How much routine work AI really saves today can be checked in the AI routine-work calculator, using figures from studies. How I keep part of my own working environment independent is in the home-server article, and the MeisterMeld case shows how I use AI in a real product.
Limits of this text
- Scenarios are not forecasts. They show what one can prepare for, not what will happen.
- Surveys are not measurements. Bitkom, WEF and DORA rest on questionnaires; the “Canaries” figure comes from US data and does not prove causation.
- As of October 2026. Laws and figures change; the AI Act details come from a secondary source, so please check the Official Journal before decisions. WEF and DORA are from 2025, the other figures from 2026. This is not legal or career advice.
Keep reading
AI · Productivity · Calculator
How much routine work does AI save? A calculator based on studies
The studies show very different numbers – from just under 3% to over 40%. Instead of a single number, the calculator gives three scenarios, each with its source, in hours and euros for your team.
AI · LLM · Content pipeline
AI content with human-in-the-loop: 200+ guide articles from LLM drafts
With care law, AI mistakes are not acceptable. This is how I use a language model as a writing aid without handing it the responsibility.
Regulation · AI · Transparency
AI Act Article 50: what needs doing by 2 December 2026
Most of Article 50’s transparency duties have applied since 2 August 2026. For the machine-readable marking of generative AI already on the market, a transition period runs until 2 December 2026. A checklist for developers and operators.
Career · DevOps · Field report
Career change into DevOps in Germany: my plan, my evidence, my gaps
I am looking for a permanent IT job, with DevOps and platform engineering as the best fit. This is an open account of what the market offers according to the numbers, where I come from and how I make my skills verifiable.
More from practice
Find more notes, case studies and projects on the homepage.
All notes Case study MeisterMeld Case study PflegeOasis Case study Kernbohr-Expert Back to the homepage
Comments
Join the discussion: comments run on GitHub Discussions (sign in with a GitHub account). Content from GitHub/giscus is only loaded after you click. Privacy