Atlassian is cutting 1,600 jobs so it can shove more money into artificial intelligence. And yes, that’s as awkward as it sounds.
Because floating around this announcement is the company line attributed to CEO Mike Cannon-Brookes: AI isn’t here to replace workers. It’s here to “augment” them. Meanwhile, 1,600 people are about to find out exactly how “augmented” their rent payment feels without a paycheck.
This is the tech industry’s favorite magic trick: swear the robots aren’t coming for your job, then quietly reduce headcount to “fund innovation” and keep Wall Street happy.
1,600 jobs cut so the AI budget can grow
Let’s translate the corporate-speak. Atlassian says the layoffs are about freeing up cash to accelerate AI investment. That tells you two things at once: AI is expensive, and the company doesn’t want to wait for revenue growth to pay for it.
Building AI into business software isn’t a one-time “add a button” project. You’re paying for models (whether you build them, license them, or rent them), cloud compute, storage, bandwidth, and the unglamorous stuff, security and data controls, so your customers’ internal project plans don’t end up in the wrong place.
And when sales cycles tighten and customers start trimming software budgets, “we’ll fund it with growth” turns into “we’ll fund it with cuts.” In software companies, payroll is the biggest, fastest lever to pull. So they pull it.
The risk, is that you cut the very people you need to pull off the AI push, product folks, engineers, support teams who understand the messy reality of customers. But layoffs look clean on a spreadsheet. AI rollouts rarely do.
The CEO says AI won’t replace workers. The math says otherwise.
Cannon-Brookes’ reported message fits the standard Silicon Valley script: AI “augments” humans; it doesn’t replace them.
Here’s the problem: jobs don’t disappear only when a machine does the exact same role, start to finish. Jobs disappear when the amount of labor needed to produce the same output drops. If a support team can handle more tickets with fewer people because AI drafts responses and summarizes threads, you don’t need a press release announcing “AI replaced support reps.” You just… don’t backfill openings. Or you cut.
In the short run, AI can look like a helpful assistant, writing, summarizing, searching, helping code, triaging requests. In the medium run, companies rebuild workflows around those tools. That’s when “assistant” starts looking a lot like “substitute,” even if nobody says the quiet part out loud.
Atlassian’s move also exposes the real question executives dodge: when productivity rises, does the company use it to grow faster, or to run leaner? A 1,600-job cut is a pretty loud answer.
AI in enterprise software isn’t cheap: models, cloud bills, data, and security
AI costs stack up in layers.
First: model access. Maybe you build. Maybe you partner. Maybe you pay a third party. Either way, it’s not free.
Second: infrastructure. AI features chew through compute. Every prompt, every summary, every “smart” search can carry a per-use cost. That’s why so many software companies are quietly reshaping pricing, gating AI behind premium tiers, limiting usage, or charging add-ons, because otherwise margins get eaten alive.
Third: governance and security. Atlassian sells tools used by engineering teams and business teams to manage projects, tickets, documentation, often sensitive internal material. If AI is reading and generating content inside those systems, the guardrails have to be serious. That means more compliance work, more security review, more legal oversight, more specialized engineering.
Here’s the irony: the “efficiency” technology often requires more people at first to make it safe and reliable. Cutting staff while ramping AI can work only if the company is aggressively reallocating roles, not just shrinking the org and hoping the remaining employees sprint faster.
Another layoff in tech’s long hangover, and a reputational gamble
Atlassian isn’t operating in a vacuum. Tech has been trimming headcount for multiple quarters after years of hiring sprees. The reasons vary, slower demand, margin pressure, strategic pivots, but AI is increasingly the banner companies wave while they move budgets around.
There’s also a symbolism here that’s hard to miss. Software companies sell “productivity” to customers. When they announce layoffs to fund AI, they’re admitting how they expect productivity to show up first: internally, as fewer humans doing the work.
For customers, that can read two ways. Maybe it signals Atlassian is serious about shipping AI features. Or maybe it signals thinner support, more bugs, slower fixes, and roadmaps that look great in keynotes but wobble in real life.
And for Atlassian’s reputation, especially with developers and product teams who tend to have a high tolerance for blunt truth and a low tolerance for corporate spin, this is a risky moment. If you’re going to cut 1,600 jobs, don’t hide behind “AI won’t replace people.” People can do the math.