What happened
Qualcomm, the US chipmaker, is cutting jobs worldwide. In an internal memo on 19 October, reviewed by Reuters, CEO Cristiano Amon called it a “performance-focused reduction” that follows the company's annual reviews. He said business priorities, organisational needs and the skills required would decide which positions go. Notifications were expected to start in November.
- Qualcomm employs about 20,000 people in India, mainly in engineering and design centres in Bengaluru, Hyderabad, Chennai and the NCR, according to ET.
- Industry estimates put the India cuts at around 1,000 jobs. ET had reported in October that 600-1,000 jobs could be affected.
- Reports say the cuts began in Bengaluru and included senior engineers. Analyst Pareekh Jain told The Hindu the cuts appear to span levels and aren't necessarily performance-based.
- Anonymous posts by employees claim some people with good ratings were let go. The Hindu says these couldn't be verified.
- Fiscal 2026 revenue was $43.12 billion, down 3%. Fourth-quarter revenue fell 8%.
- Qualcomm hasn't confirmed the India numbers and didn't respond to The Hindu or ET.
The gap between a performance label and what employees say they're seeing is the interesting part for HR. When a company cuts for two reasons at once, weak performance and changing business needs, how it picks people and explains each exit matters a lot.
The case
You're the HR head of a 20,000-person engineering centre in India. Headquarters wants a 5% reduction after this year's reviews, using both performance ratings and business priorities. Some projects are also being stopped. How do you decide who leaves, and what do you tell people?
1Clarify
- Is the target a headcount number or a cost number? If it's cost, other levers count too, like freezing open roles or cutting contractors.
- How many people got a low rating this year? This tells us how much of the 5% performance can honestly explain.
- How consistent are ratings across managers? If one team rates hard and another rates easy, ratings alone will punish the wrong people.
- Which projects are stopping, and are there open roles elsewhere? People on a stopped project lose their job because of the role, not their work. Some might move to open roles.
2Structure
- Size: the target, and how much other levers can cover.
- Selection: who leaves, and on what basis.
- Process checks: rating calibration, a review panel and a way to appeal.
- Communication: what we tell leavers, and what we tell everyone who stays.
- Retention: keeping the people we most need once the cuts are done.
3Analyse
Can ratings alone get you to 1,000?
The 5% comes from the reported numbers: 1,000 ÷ 20,000. The rest are assumptions for the case. Say 3% of staff, or 600 people, were rated below expectations this year. That is all performance can honestly explain. The other 400 must come from somewhere else.
| Approach | From low ratings | From the rest | What leavers are told |
|---|---|---|---|
| Ratings only | 600 | 400 people rated 'meets expectations', ranked by managers | All told it was performance |
| Two buckets | 600 | 400 roles on stopped projects | Each told the real reason |
With ratings only, 400 people with acceptable ratings are told they underperformed. They know that isn't true, and so do their colleagues. That is exactly the kind of story that spreads in anonymous posts.
The hidden cost of a cut that feels unfair
Suppose a cut seen as unfair pushes voluntary attrition up by 1 percentage point among the 19,000 who stay. That is 190 extra resignations. If replacing each one costs about ₹20 lakh in hiring and time to get up to speed, the bill is 190 × ₹20 lakh = ₹38 crore. Both approaches remove the same 1,000 jobs, so this cost is avoidable.
Attrition rise and replacement cost are assumptions. Even a fair process will cause some resignations; the point is the difference.
4Recommend
I recommend splitting the cut into two buckets and telling each person which one they're in: about 600 exits for performance after rating calibration, and about 400 roles on stopped projects.
First, it's honest, which protects the trust of the 19,000 who stay. On our assumptions, that avoids up to ₹38 crore in extra resignations. Second, people in the role bucket can be offered open roles first, so we keep skills we'd otherwise hire for. Third, a review panel checking every name, with a simple appeal route, catches cases where one manager's harsh ratings drive the list.
The risk is time. Calibration and redeployment add a few weeks, and rumours fill any silence. I'd give all staff a clear timeline on day one and finish every conversation within that window.