AI hiring is sorting talent, not lifting it

Oracle's global headcount fell by about 21,000, or 13%, in the fiscal year ended May 31, 2026, to roughly 141,000. In the same period, Q1 FY27 capex hit $28.5 billion against $8.5 billion a year earlier, full-year capex guidance sits at $90-95 billion, cloud infrastructure revenue rose 121% to $7.4 billion, and remaining performance obligations reached $664 billion. A second round of India cuts has begun, with product engineering hit hardest.

Teksands take: This is the cleanest template yet for the AI-era workforce trade: headcount down, capex up, mix shifting from generalist product and support roles toward AI infrastructure, cloud and platform engineering. CFO Hilary Maxson told staff Oracle should not operate on 'doing more with less'. Read that carefully. The constraint is not effort, it is where the effort sits. If your engineering pyramid still mirrors last decade's demand, the reallocation is coming for you too.

An AIM Research and FDE Academy report on India's Forward Deployed Engineer landscape finds employer demand for field-deployed AI builders at 95% against available talent at 45% — a 50-point gap, the largest among all FDE role families studied. Customer and consulting skills appear in 98% of FDE job postings, ahead of core engineering at 94% and applied AI at 79%. Median pay for field-deployed AI builders runs from ₹32 lakh at 0-3 years to ₹155 lakh at 15+ years, with AI deployment experience commanding a 30-50% premium.

Our take: The bottleneck has moved from AI literacy to AI accountability. The people who matter are the ones who can take a model into a bank's legacy stack, debug it in a client environment and own the outcome. That changes the assessment bar: production deployment history, enterprise integration and incident ownership beat certifications and model-building pedigree. Screen by title and you will keep losing candidates you cannot even identify.

Adidas called India strategic and cut half its Gurugram tech hub in the same breath

Adidas told employees at its Gurugram India Tech Hub on September 15 that a large slice of the technology organisation was being cut, with software engineers, data scientists and data engineers among those affected. Reports put the number above 350 of roughly 700 roles, though the company confirmed only that roles were reduced and declined to give a headcount. Some staff received exit letters immediately; others were asked to stay until December 15 to hand over work to colleagues at other technology hubs.

This is the GCC story of the quarter in reverse: a captive centre cutting deep into engineering and data roles while publicly calling India strategic. The signal is not exit, it is narrowing — fewer generalist seats, sharper AI and data mandates. For NCR recruiters, the December 15 handover cohort is a live pipeline with a hard expiry. Move early, before Q4 counteroffers do.

Four signals worth watching

The lesson: The largest organised hiring machine in one of India's biggest tech catchments is not hiring for tech. That is not a failure of the drive — it is an accurate read of where mass-entry demand actually sits: manufacturing, capital goods, BFSI operations, healthcare and apprenticeships. If your fresher funnel assumes Pune's graduate supply flows to software by default, you are competing with employers who offer faster decisions and lower friction. Treat government camps as a rival channel, not background noise.

India's tech market is shedding senior talent while its mass-entry hiring goes non-tech

Hire for what the role actually needs

The gap this week is not volume, it is fit. A 50-point shortfall in AI deployment talent, a GCC estate growing while hiring falls 28%, and a market releasing senior cloud engineers in waves all point the same way: the winning teams will be the ones that define the capability precisely before they open the requisition. Teksands works with GCCs, IT services and product companies to build hiring pipelines around the skills that actually ship — AI deployment, cloud platform, data engineering and operations — rather than the titles that look good on a job description.

Rewrite one AI requisition's screening questions this week

Take your highest-priority AI or data requisition and replace the framework checklist with three questions that test production ownership: describe a model you took into a client or enterprise environment and what broke; describe an integration you owned end to end across a legacy stack; describe an incident you debugged in production and what you changed afterwards. The FDE data says customer and consulting skills appear in 98% of postings, ahead of core engineering at 94% and applied AI at 79% — so the bar is deployment and accountability, not model-building pedigree. Run the same three questions across your next five AI interviews and compare pass rates against your current screen.