AI Layoffs Tracker: Which Jobs & GCCs Fall First

Every week a new headline announces thousands of jobs "lost to AI," and if you run a GCC or a delivery org, each one lands like a threat to your headcount and your roadmap.

The problem is that the coverage is almost entirely doom and almost never net — it counts the cuts but ignores the demand curves that are simultaneously going vertical.

This tracker fixes that: one continuously-maintained, evidence-first map of who is cutting, who is hiring, and which roles and Indian GCCs are actually exposed versus quietly surging.

How to Read This Tracker (And What Counts as an "AI Layoff")

Most "AI layoffs" lists are really restructuring lists with an AI label stapled on for clicks. Before you act on any number in this tracker, it helps to know how we separate signal from spin.

This matters more here than on a typical news page, because employment is a livelihood topic. We date-stamp every figure, link the primary source where one exists, and frame everything as decision support — not a prediction about any individual's job.

AI layoffs vs. restructuring: the line that matters

We classify a reduction as AI-linked only when the company explicitly ties roles to automation, agent deployment, or a compute-funding trade-off — not when "AI" merely appears in the same press release as a cost cut.

That distinction is the difference between a structural shift you should plan around and an ordinary downturn wearing an AI costume. The tracker flags which is which in the "net signal" column.

How we score GCC exposure

For Indian GCC leaders, the relevant question isn't "is this company cutting?" — it's "is this company cutting the kind of work my centre does?" So each row carries a GCC-exposure tag: High, Mixed, or Low.

High means the cuts hit roles commonly offshored to India. Low means the cuts are concentrated in functions GCCs rarely own. That single lens turns a scary headline into a planning input.

The 2026 AI Layoffs Tracker: Who Cut, Who Hired

The table below is the heart of this hub. It pairs the headline reductions with their GCC exposure and a plain-language net signal, so you can scan the landscape in under a minute.

Banking & financial services

Banking is the most-watched sector because it is both heavily offshored and heavily process-driven — exactly the profile AI compresses fastest. HSBC's reduction is the flagship case, and it reads very differently for India than the headline suggests.

The nuance: the automated work skews toward standardised middle- and back-office processing, while the bank still needs people who own risk, judgment, and client outcomes.

Read the full breakdown in our analysis of why HSBC's reduction may spare India's capability centres.

Enterprise software & infrastructure

Oracle's ~30,000-role reduction is a different animal: it's less "AI replaced these people" and more "we are reallocating payroll into a roughly $300B compute bet."

The roles most exposed are legacy database and admin functions, not the cloud and AI-infrastructure roles the company is leaning into.

Atlassian's move is subtler still — the announcement named a number but stayed quiet on the functions, which is itself a tell about where agent tooling is replacing repeatable work first. Both cases are covered in depth in the Oracle and Atlassian spokes of this hub.

Indian IT & the GCC ripple

For India, the second-order effect matters more than any single company's headline. When global firms automate the standardised work that was historically offshored, the pressure lands first on cost-arbitrage seats — the model some commentators bluntly call "bums on seats."

But that same pressure is accelerating the shift toward capability centres that own products and outcomes.

Which side of that line a centre sits on is the whole game, and we map it in our guide to how AI is reshaping GCC jobs in India.

The Counter-Narrative Nobody Puts in the Headline

Here is the insight most coverage gets backwards: 2026 is not simply a year of AI destroying jobs. It is a year of violent reallocation, and the demand side of that reallocation is larger and faster-moving than the cut side — it just doesn't make a frightening headline.

While roles were cut, two demand curves went vertical

Two numbers reframe the entire debate. Agentic-AI job postings are up roughly 280% year-on-year, and demand for forward-deployed engineers — the people who embed AI into real enterprise workflows — has risen by around 800%.

At the same time, the titles themselves are mutating: "AI Engineer" is absorbing the old "ML Engineer" label, while genuinely new roles like AI Evals Engineer and Context Engineer have appeared from nothing.

The forward-deployed engineer surge is the single clearest upgrade path out of an exposed role, and we treat it as the hub's "surge" anchor.

Dig into the demand data in our breakdown of why forward-deployed engineer demand is up 800%.

The net number — and why doom-only coverage hides it

So what is the net effect? Honestly, nobody has a clean, audited figure — and that uncertainty is exactly why doom-only coverage wins the click. A vertical "+280%" demand curve is harder to photograph than an empty office floor.

The defensible position for a leader is this: treat the cuts as concentrated and real, treat the demand surge as broad and real, and plan for reallocation rather than disappearance.

For the India-specific data, see our analysis of the AI job displacement statistics that rarely make the headline.

Which Jobs Are Actually Exposed (Hint: It's Tasks, Not Titles)

The most common misconception in this entire debate is that AI replaces jobs. It doesn't — at least not first.

It replaces tasks, and a job disappears only when enough of its tasks are automated that the remaining work no longer justifies the role. That changes how you assess exposure. A senior title with a high share of repeatable, well-documented tasks can be more exposed than a junior one full of ambiguity and judgment.

The high-exposure cluster

Roles compress fastest when their work is repeatable, text- or data-heavy, well-documented, and loosely coupled to physical or relational context.

Standardised reporting, first-line support, routine code generation, and high-volume back-office processing all sit here. This is precisely the work that was easiest to offshore a decade ago — which is why the GCC overlap is so high.

The surprisingly durable cluster

The durable roles aren't the obvious "human" jobs people reach for. They're the ones that combine accountability for an outcome with the ability to direct and verify AI's output.

We map the full exposure picture, including the safe-looking roles already shrinking, in the dedicated spoke on which jobs AI is replacing first. The companion list of durable, higher-paying paths lives in our guide to AI-proof careers.

The GCC Fault Line: Cost-Arbitrage Centres vs. Capability Centres

If you take one structural idea from this hub, make it this: in 2026, India's GCC landscape is splitting along a single fault line, and which side a centre is on predicts almost everything about its AI exposure.

Why cost-arbitrage GCCs are shrinking

Cost-arbitrage centres were built to do known work more cheaply. Their entire value proposition is labour cost per unit of standardised output — and that is exactly the value AI collapses, because an agent does standardised output cheaper still.

When the moat is "we are cheaper at repeatable tasks," AI doesn't just narrow the moat; it drains it.

Why capability centres are still hiring

Capability centres own products, outcomes, and decisions. Their value is judgment, context, and accountability — the things AI augments rather than replaces.

The strategic mandate for any GCC leader is therefore brutally clear: move up the value chain from cost centre to capability centre before the arbitrage moat fully drains.

The frameworks for measuring that shift sit in our AI Workforce Transformation Strategy pillar.

A Leadership Playbook for the Disruption

Tracking the disruption is only useful if it changes what you do on Monday. Here is the action layer, split for the two audiences this hub serves.

For PMO directors and workforce planners

Start by mapping AI exposure at the task level across your portfolio, not the org-chart level.

Then re-baseline capacity assuming agent-augmented throughput, and redirect the freed capacity toward outcome-ownership work rather than treating it purely as a cost saving.

The teams that win this cycle are the ones that convert automation gains into expanded scope.

For individuals: the reskilling on-ramp

If your role sits in the high-exposure cluster, the durable move is not to out-compete the agent at repeatable tasks — it's to become the person who directs, verifies, and owns the outcome of agent work.

Our structured path is laid out in the reskilling-for-agentic-AI spoke, a 90-day plan that doesn't assume a CS background.