Not a Cycle: A Structural Change
Technology sector layoffs between 2022 and 2025 affected hundreds of thousands of workers across companies of every size — not only the growth-stage startups that overhired during the pandemic but the most established technology companies in the world: Microsoft, Google, Meta, Amazon, Salesforce, and others that had never previously laid off at scale. The layoffs were initially described as a correction to overhiring during the pandemic’s digital-acceleration period, and that description was partially accurate for the first wave. The persistence and pattern of subsequent layoffs suggests something more structural than a one-time correction.
Understanding what’s actually happening in the technology labour market — beyond the ‘correction’ and ‘tough market’ language that every layoff announcement uses — helps technology workers, people considering technology careers, and technology industry observers make more accurate assessments of what the industry looks like going forward.
The AI Substitution Effect: Real and Partial
A portion of technology labour reduction reflects genuine AI substitution: tasks that previously required significant human time — certain coding tasks, content moderation at scale, customer service interactions, data processing, and some software testing — are being partially or substantially handled by AI tools. Companies that deployed AI assistance in these functions have reduced headcount in affected roles faster than attrition would have accomplished.
The partial qualifier matters: AI substitution in technology roles is currently most significant in specific, well-defined task categories rather than in the full scope of most technology worker roles. The argument that AI is replacing software engineers wholesale overstates the current state; the argument that AI is having no effect on technology employment understates it. The actual picture is more specific: certain software engineering tasks are faster and require less human time, but the scope of software being built and the complexity of the systems being maintained has also grown, partially absorbing the productivity improvement.
The Platform Maturation Dynamic
Technology companies that grew aggressively during the pandemic were partly growing on the temporal acceleration of digital adoption that pandemic conditions produced — e-commerce, streaming, cloud collaboration, and digital services all saw adoption compressed into a shorter time frame than would have otherwise occurred. As conditions normalised, growth rates returned to pre-pandemic trajectories and the workforce sized for accelerated growth was oversized for the decelerated growth that followed.
The companies that have gone through substantial workforce reduction and maintained or improved revenue and margins are demonstrating that they were over-hiring relative to sustainable productivity requirements — that they had accumulated organisational overhead in layers of management, coordination functions, and support roles that technology companies have historically outgrown. This doesn’t make the layoffs less disruptive for those affected, but it does indicate that the reduction has been partly a productivity-per-employee recalibration rather than purely a revenue problem.
What It Means for Technology Careers
The technology labour market in 2025–2026 is more competitive than in 2021, but technology skills remain among the most in-demand in the broader labour market. The roles with the most hiring are concentrated in AI/ML engineering, data engineering, security (cybersecurity and AI safety), cloud infrastructure, and product engineering for AI-powered products. The roles with the most displacement have been in middle management, programme management, certain QA roles, and the most commoditised parts of the engineering labour market.
The career implications for technology workers: the skills that were adequate during rapid industry growth — familiarity with an established technology stack, experience at a well-known company — are less sufficient differentiation in a more competitive market. Specific expertise in AI tools (including but not limited to ML engineering), cross-functional skills that go beyond pure technical depth, and the demonstrated ability to produce measurable outcomes are the differentiators that technology workers are finding most valuable in the current market.
The Industry Concentration Trend
One structural outcome of the post-2022 technology market is increased concentration: the companies with the most resources to invest in AI (Microsoft, Google, Meta, Amazon) have expanded their capabilities while smaller competitors have struggled to access the compute, talent, and data advantages that frontier AI requires. The democratisation narrative that characterised much of the 2010s technology era — any startup could challenge an incumbent with a better product — is being tested by the scale requirements of frontier AI development.
The open source AI movement (covered in the Tech News section of Vol. 2) partially counteracts this concentration by making capable models available without the frontier compute investment. But the most capable frontier models, and the infrastructure that makes them useful at scale, remain concentrated among a small number of very large organisations. The implications for technology industry diversity, startup ecosystems, and the distribution of AI’s economic benefits are among the most significant questions in current technology policy.