LoopWHI ’26

Nutrition

survey
9,097 respondents — 7,614 answered the eat-out question, 7,632 answered protein sources, 7,669 declared a diet

Key takeaways

  • 23.5% of urban working Indians eat out or order in three or more times a week — and the share more than doubles from the 50s to the 20s, falling from 10.9% to 27.9% as age drops.
  • Bengaluru and Chennai eat out most — 30.0% and 30.4% of professionals there hit the three-plus-times threshold, against 18.9% in Pune. The frequent-takeaway habit is a metro-by-metro story, not a national one.
  • Dal is still the floor of the Indian protein plate: 80.7% rely on it. Eggs (63.5%) and chicken (56.5%) follow. The supplement layer is thin — only 12.1% use whey, 7.7% plant-protein powder.
  • The "men supplement more" gap is almost entirely a story about men in their 20s. Male whey use runs 18.3% in the 20s against 9.5% for women; by the 30s the gap collapses to two points.
  • Whey use tracks income almost linearly — from 10.0% under ₹3 lakh to 19.2% above ₹40 lakh. Engineered protein is, for now, a premium good.
  • Vegetarianism splits the metros violently: 48.7% in Delhi NCR against 16.1% in Hyderabad and Chennai — a 32-point spread, and the dietary backdrop to Bengaluru's worst-in-metro B12 deficiency.

Of 7,614 urban working Indians who answered the eat-out question in WHI 2026, 23.5% eat out or order in three or more times a week — and 8.5% do so five times a week or more. The reference point matters here. India's food-delivery platforms have rebuilt a meaningful share of the country's weekday meals around an app: the sector generated roughly ₹1.2 lakh crore in gross output in FY24, led by Swiggy and Zomato (NCAER–Prosus). The survey reads that shift from the demand side — close to a quarter of the working population now treats restaurant or delivery food as a thrice-weekly default rather than an occasion.

Eat-out / order-in frequency Share n
I don't order in or eat out 21.7% 1,650
1–2 times a week 54.8% 4,171
3–4 times a week 15.0% 1,145
5–6 times a week 4.8% 366
7+ times a week 3.7% 282

A quarter of the workforce eats out three-plus times a week

Eat-out / order-in frequency — the bottom three bands are the 23.5% 'frequent' group

I don't order in or eat out
n = 1,650
21.7%
1–2 times a week
n = 4,171
54.8%
3–4 times a week
n = 1,145
15.0%
5–6 times a week
n = 366
4.8%
7+ times a week
n = 282
3.7%
Values in %

The frequent cohort (3–4, 5–6, 7+) sums to 23.5%; 8.5% eat out five or more times a week. The dominant pattern is still restraint — three in four abstain or keep it to once or twice.

The dominant pattern is restraint, not excess — more than three in four either abstain or keep restaurant food to once or twice a week. The frequent group is a minority. But it is a metabolically loaded minority, and as the cuts below show, it concentrates in the youngest, highest-spending, most urban slice of the workforce — the same slice already carrying the steepest glucose curve in this report.

Frequent eating-out collapses with age — from 27.9% in the 20s to 10.9% in the 50s#

Age band Sample Eats out 3+ times/week
20–29 3,055 27.9%
30–39 3,257 21.9%
40–49 1,011 18.8%
50–59 229 10.9%

Frequent eating-out collapses with age

Share eating out 3+ times a week, by age band

A 2.6× fall from the 20s to the 50s — the eat-out curve runs opposite to almost every biomarker curve in this report. The 50s cell is smaller (n = 229); read it as directional.

The eat-out curve runs the opposite way to almost every biomarker curve in this report. Glucose dysfunction, blood pressure, and liver strain all climb with age; restaurant frequency falls steadily, halving and then halving again. A professional in their 20s is two and a half times more likely than one in their 50s to be eating out three times a week. The drivers are familiar — younger workers eat out more because they have fewer dependents to cook for, smaller kitchens, longer office hours, and more disposable income relative to fixed commitments. The household forms; the habit recedes.

The two curves crossing matters. The age group eating restaurant food most often is the age group whose metabolism is still mostly forgiving of it. The age group eating it least is the one that can least afford the glycaemic load. The behaviour and the biological vulnerability are misaligned by about two decades — the Blood Sugar page's central finding, that glucose dysfunction doubles from the 20s to the 30s, restated through a lunch order.

Bengaluru and Chennai eat out most; Pune least#

City Sample Eats out 3+ times/week
Chennai 415 30.4%
Bengaluru 1,611 30.0%
Kolkata 269 26.4%
Mumbai 1,050 23.7%
Delhi NCR 1,052 21.5%
Hyderabad 801 20.6%
Pune 994 18.9%

Bengaluru and Chennai eat out most; Pune least

Share eating out 3+ times a week, by city, ranked

Chennai
n = 415
30.4%
Bengaluru
n = 1,611
30.0%
Kolkata
n = 269
26.4%
Mumbai
n = 1,050
23.7%
Delhi NCR
n = 1,052
21.5%
Hyderabad
n = 801
20.6%
Pune
n = 994
18.9%

National survey average — 23.5%

Values in %

The spread across metros (18.9%–30.4%) is wider than the spread across genders. Bengaluru's reading concentrates in its young, delivery-dense cohort.

The spread across metros is wider than the spread across genders. Bengaluru and Chennai professionals are around 60% more likely than Pune professionals to eat out three or more times a week. Bengaluru's reading is consistent with the rest of its profile in this report — it is the youngest major-metro cohort in the sample, the densest delivery market in the country, and a workforce structured around long office hours and single-occupant flats. Pune at the other end fits its sleep profile too: smaller scale, more home-cooked routine, a shorter working day.

The Bengaluru reading concentrates in the young. Within the city, frequent eating-out runs 34.3% in the 20s and falls to 23.1% by the 40s — so the city-level number is largely the city's age structure expressing itself, amplified by the delivery infrastructure built around exactly that cohort.

Dal is the floor; the supplement layer is thin#

Protein sources are multi-select — respondents could name every source they use regularly, so the column does not sum to 100%.

Protein source Share who use it regularly n
Dal / chana / rajma 80.7% 6,162
Eggs 63.5% 4,848
Chicken 56.5% 4,310
Paneer / tofu / soy 55.6% 4,247
Fish 36.3% 2,767
Other 12.3% 935
Whey protein / casein 12.1% 920
Plant-protein powder 7.7% 591

Dal is the floor; the supplement layer is thin

Share who use each protein source regularly — engineered supplements (brass) sit far below the food sources

Dal / chana / rajma
n = 6,162
80.7%
Eggs
n = 4,848
63.5%
Chicken
n = 4,310
56.5%
Paneer / tofu / soy
n = 4,247
55.6%
Fish
n = 2,767
36.3%
Other
n = 935
12.3%
Whey protein / casein
n = 920
12.1%
Plant-protein powder
n = 591
7.7%
Values in %

Multi-select — respondents could name every source, so the column does not sum to 100%. The cliff between fish (36.3%) and whey (12.1%) separates food from engineered protein.

Four out of five working Indians anchor their protein on pulses. This is the single most stable finding in the nutrition data across both editions of the report, and it is genuinely a strength — dal is cheap, complete enough when paired with cereal, culturally universal, and indifferent to income. The problem is not the floor. The problem is what sits on top of it.

The bottom two rows are the more telling read. Engineered protein — whey and plant powders — reaches barely one in eight and one in thirteen of the workforce respectively. Awareness of protein as a deliberate dietary target, rather than a thing that arrives incidentally with dinner, is still confined to a small and, as the next two cuts show, distinctly privileged minority.

The "men supplement more" gap is a story about men in their 20s#

Whey use, 20s Whey use, 30s Whey use, 40s
Women 9.5% 9.9% 7.9%
Men 18.3% 11.4% 8.2%

The 'men supplement more' gap is a story about men in their 20s

Whey use by gender across age bands — the gap closes as age rises

WomenMen

Men supplement at nearly twice the rate of women in the 20s (18.3% vs 9.5%); by the 30s the gap is a point and a half, and by the 40s it is gone.

Across the whole sample, men use whey at 13.3% against 9.3% for women — a real but modest gap. Split by age, the gap turns out to live almost entirely in one cell. Among professionals in their 20s, men supplement at nearly twice the rate of women; by the 30s the difference narrows to a point and a half, and by the 40s it is gone. The headline gender gap in protein supplementation is not a fact about men and women. It is a fact about gym culture in the male 20s, and it ages out fast.

This reframes the equivalent WHI 2025 finding, which reported a flat "2.8:1 male" supplementation ratio without the age split. The ratio was real but the framing was incomplete — it read as a durable gender difference when it is really a young-male behaviour that fades as the cohort ages into other priorities.

Whey use rises almost linearly with income#

Income band Sample Whey use
Under ₹3 lakh 987 10.0%
₹3–6 lakh 2,176 11.0%
₹6–10 lakh 1,524 12.0%
₹10–15 lakh 1,029 12.1%
₹15–25 lakh 789 15.0%
₹25–40 lakh 398 16.3%
₹40 lakh+ 146 19.2%

Whey use rises almost linearly with income

Whey use by income band, low to high — a shallow but steady gradient

Under ₹3 lakh
n = 987
10.0%
₹3–6 lakh
n = 2,176
11.0%
₹6–10 lakh
n = 1,524
12.0%
₹10–15 lakh
n = 1,029
12.1%
₹15–25 lakh
n = 789
15.0%
₹25–40 lakh
n = 398
16.3%
₹40 lakh+
n = 146
19.2%
Values in %

Adoption nearly doubles from the bottom band (10.0%) to the top (19.2%). Absolute values are low — the honest read is a clean climb, not a dramatic one.

Engineered protein nearly doubles in adoption from the bottom income band to the top. The gradient is shallow but it is clean — every step up the income ladder adds whey users. Two readings fit, and both are probably true. Supplement protein is a discretionary spend, so cost screens out lower earners. And protein-as-a-target is a piece of health literacy that travels with income, gym membership, and the social networks where the habit is normal. Whichever weighs more, the result is the same: the workforce slice most able to optimise its protein intake is the slice that already has the most metabolic headroom, and the slice that would benefit most from a few extra grams a day is the slice least likely to be buying it.

Vegetarianism splits the metros by 32 points — the backdrop to Bengaluru's B12 problem#

City Sample Vegetarian Non-vegetarian Eggetarian
Delhi NCR 1,059 48.7% 38.4% 12.8%
Pune 1,004 38.6% 51.6% 9.8%
Mumbai 1,055 35.5% 56.6% 7.9%
Bengaluru 1,622 25.0% 65.3% 9.7%
Hyderabad 808 16.1% 77.1% 6.8%

Vegetarianism splits the metros by 32 points

Diet declaration by city — vegetarian / eggetarian / non-vegetarian, ordered by vegetarian share

VegetarianEggetarianNon-vegetarian
Delhi NCR
n = 1,059
48.7% veg
Pune
n = 1,004
38.6% veg
Mumbai
n = 1,055
35.5% veg
Bengaluru
n = 1,622
25.0% veg
Hyderabad
n = 808
16.1% veg
Values in

National split: 32.7% vegetarian, 9.0% eggetarian, 58.2% non-vegetarian. The diet-to-B12 link is an audience-level pattern, not an individual one — the survey diet responses and biomarker B12 readings come from overlapping but unlinked populations.

Nationally the split is 58.2% non-vegetarian, 32.7% vegetarian, 9.0% eggetarian. The metro-level figures barely resemble each other. Delhi NCR is close to half-vegetarian; Hyderabad and Chennai (16.1% each) and Kolkata (14.2%) are overwhelmingly meat- and fish-eating. This is the strongest dietary cut in the survey, and it carries a clinical tail.

The cut that does not behave as expected is Bengaluru. The city carries the worst B12 deficiency of any metro in the WHI 2026 biomarker data — 78.8% of tested reports below threshold (the Vitamin B12 and Bengaluru pages take up that floor in full) — yet its vegetarian share, at 25.0%, is below the national average and well under Delhi's. Vegetarianism alone does not explain Bengaluru's B12 floor. The likelier read is that the city's deficiency is driven less by who avoids meat than by how little of the population, vegetarian or not, gets adequate dairy and supplementation — a measurement-friendly deficit that no one is currently measuring at scale.

Fish is a coastal protein, not a national one#

City Sample Eats fish regularly
Kolkata 268 69.0%
Chennai 414 51.7%
Hyderabad 803 42.2%
Mumbai 1,051 39.2%
Bengaluru 1,612 36.8%
Pune 1,002 29.2%
Delhi NCR 1,054 19.7%

Fish is a coastal protein, not a national one

Share who eat fish regularly, by city — coastal/riverine cities (brass) sit two to three times above land-locked Delhi NCR

Kolkata
n = 268
69.0%
Chennai
n = 414
51.7%
Hyderabad
n = 803
42.2%
Mumbai
n = 1,051
39.2%
Bengaluru
n = 1,612
36.8%
Pune
n = 1,002
29.2%
Delhi NCR
n = 1,054
19.7%
Values in %

The one protein source whose adoption is governed by geography rather than income or age — there is no single national protein plate.

Fish consumption maps onto the coastline. Kolkata, Chennai, and the coastal-metro cluster sit two to three times above land-locked Delhi NCR. This is the one protein source in the survey whose adoption is governed by geography rather than income or age — which makes it the cleanest illustration of how regional the "Indian diet" actually is. There is no single national protein plate; there are several, and which one a worker eats from is set substantially by where the office is.

References

  1. 1Swaminathan, S., Vaz, M. & Kurpad, A.V. Protein intakes in India. British Journal of Nutrition 108, S50–S58 (2012). https://doi.org/10.1017/S0007114512002413
  2. 2Green, R., Allen, L.H., Bjørke-Monsen, A.L. et al. Vitamin B12 deficiency. Nature Reviews Disease Primers 3, 17040 (2017). https://doi.org/10.1038/nrdp.2017.40
  3. 3Anand, S.S., Hawkes, C., de Souza, R.J. et al. Food consumption and its impact on cardiovascular disease: importance for developing countries. Journal of the American College of Cardiology 66, 1590–1614 (2015). https://doi.org/10.1016/j.jacc.2015.07.050
  4. 4Misra, A., Singhal, N., Sivakumar, B. et al. Nutrition transition in India: secular trends in dietary intake and their relationship to diet-related non-communicable diseases. Journal of Diabetes 3, 278–292 (2011). https://doi.org/10.1111/j.1753-0407.2011.00139.x