Tech companies are cutting management layers as AI takes over more coordination work, but startups should be careful before copying the playbook too quickly.
The manager who only tracks status, passes messages upward, and schedules meetings is suddenly looking expensive. In a market where founders are being told to do more with smaller teams, Big Tech’s latest experiment is hard to ignore: fewer layers, more hands-on leaders, and AI tools handling some of the work that once justified a full management stack.
According to Business Insider, companies including Coinbase, Block, Meta, Snap, and Atlassian have been moving toward flatter structures as executives push for speed, lower costs, and higher individual output. Coinbase CEO Brian Armstrong has been especially direct about the direction of travel, arguing that the company does not want pure managers and expects leaders to stay close to the actual work. Block has used similar language around player-coaches, while Meta’s earlier push against managers managing managers gave the wider industry a phrase it could easily repeat.
That matters because startups often treat large tech companies as a preview of their own future. If Meta or Coinbase cuts management layers and keeps moving, a first-time founder may conclude that management itself is the problem. That is too simple. The real question is whether AI is replacing managerial work, or merely exposing how much weak management was built around forwarding information that software can now summarize in seconds.
There is a practical reason this trend has momentum. Modern AI tools can write meeting summaries, turn messy project updates into readable reports, flag delays, draft performance documentation, and help teams search across internal knowledge. Productivity dashboards can show which tasks are stuck, which teams are using AI tools, and where work is slowing down. A founder who once needed three people to collect and translate that information may now see one capable operator supported by software.
But coordination is only part of management. The harder work is judgment. Someone still has to decide when a product delay is a normal tradeoff or a sign that the roadmap is broken. Someone has to notice when a strong engineer is quietly burning out, when a junior hire needs coaching instead of pressure, and when a team is avoiding a hard conversation because the metrics still look fine. AI can surface signals, but it does not own the outcome.
This is where the endangered manager story becomes more complicated. The layer being cut is often not leadership, but management theatre. If a manager cannot coach talent, improve execution, clarify priorities, or raise the quality bar, AI makes that weakness more visible. Once the reporting and scheduling work is automated, the remaining value of the role has to be real.
For experienced operators, that may be a good thing. The best managers were never just meeting routers. They were translators between strategy and execution, protectors of focus, and the people who knew when a team needed air cover. In a flatter company, those skills become more important, not less, because fewer layers means mistakes travel faster and ambiguity reaches the front line sooner.
Founders should not confuse flat with focused
Startups have a special temptation here. A lean org chart looks disciplined. It signals urgency to investors, keeps payroll lower, and reduces the number of people who can slow a decision. In the early stages, that can be exactly right. A ten-person startup does not need a heavy management structure, and a founder who hides behind managers too early usually creates distance from customers, product, and hiring quality.
The risk comes when flatness becomes an ideology. A company with 40 people, five workstreams, new managers, and aggressive revenue targets may not be more efficient simply because it has fewer titles. It may just be pushing invisible management work onto employees who were hired to build, sell, or support customers. The cost still exists. It shows up as unclear ownership, duplicated decisions, weaker onboarding, and senior individual contributors spending their best hours resolving coordination problems.
That is especially dangerous for first-time founders in 2026, because AI can make a strained team look healthier than it is. Weekly updates may read better. Roadmaps may appear cleaner. Performance notes may be more polished. But if no one is developing people, enforcing standards, or resolving conflict, the company is not lean. It is under-managed.
The better lesson from Big Tech is not to eliminate managers reflexively. It is to make every management role prove its connection to output. In a startup, that means managers should be close enough to the work to understand tradeoffs, skilled enough to coach people through harder problems, and accountable enough that their teams move faster because they are there. Some will be player-coaches. Some will manage larger teams with AI support. Some roles should disappear. The distinction matters.
Founders should also be honest about stage. At seed, the founder is often the manager, recruiter, product lead, and escalation path. At Series A, the company needs repeatable execution. By Series B, mentorship and quality control become part of the operating system. Cutting managers at the wrong moment can save cash while quietly damaging the machine that turns strategy into work customers can actually use.
The next phase of startup hiring will likely reward managers who can operate with sharper tools and fewer layers around them. The old middle layer is under pressure, and much of it deserves scrutiny. But companies that treat AI as a replacement for leadership will learn the difference quickly. Software can compress information. It cannot build accountability on its own.
Also read: AI may repeat the China shock with bigger gains for business • Founders are learning that startup obsession has a personal cost • AI biosecurity risk is becoming a startup compliance problem
