How many white petrol cars are there in Bengaluru?

There are white cars everywhere in Bengaluru. But how many of them are actually petrol cars? It sounds like a question that should have a readily available answer. A registration database might know the colour of a car and the fuel it runs on, but a public citywide cross-tab is not easy to find. So, rather than pretend we have that number, let us see how close we can get with a few defensible anchors and some explicit assumptions.

The interesting part is not the multiplication. It is deciding what to multiply, what the numbers mean, and how much uncertainty we are willing to carry. We will begin with registrations, make the missing assumptions visible, address the obvious cab objection, and then approach the question from the household side as a check.

For consistency, white means white alone, not white and silver combined. Petrol means petrol-only cars; petrol-CNG dual-fuel vehicles and hybrids are not deliberately included. The intended universe is Bengaluru’s resident passenger-car population, with commercial cabs considered separately. Registered vehicles are a proxy for vehicles in use, not a perfect count of active cars. The geographical boundary is also imperfect because different sources use Bengaluru Urban, the city or broader metropolitan definitions.

1. Begin with the strongest anchor

The Karnataka Transport Department figure reported by Business Today puts Bengaluru at 25.2 lakh registered private cars as of 30 June 2026. This is a better starting point than the older FY2024–25 figure and the extrapolation we initially considered. When a stronger anchor appears, there is little reason to preserve a more elaborate workaround. [1]

We also need to account for cars used in Bengaluru but registered elsewhere. Migration and enforcement of out-of-state vehicle rules establish that this population exists, but they do not tell us its precise share of the fleet. [2] I will therefore allow a 5% uplift, and test 2% to 8% rather than present the central assumption as measured data.

2. White is surprisingly hard to pin down

Two established automotive-coatings reports give materially different answers. BASF’s 2024 India report places white at 49%, while Axalta’s 2025 report places it at 33%, including solid and pearl white. Both are useful, but neither is a Bengaluru stock census. Their samples, methodologies and report years differ, so averaging them is a heuristic rather than a statistically weighted estimate. [3, 4]

Rather than choose the figure that makes the answer look prettier, we will use 41% white as the central assumption and 35% to 48% as a working range. This envelope is judgemental, not a confidence interval. Bengaluru’s existing fleet may differ from national new-vehicle colour preferences, and older vintages may have a different mix. The disagreement is useful precisely because it makes that uncertainty visible.

3. Petrol: stock is not this year’s sales

Autocar India reports that petrol’s share of new passenger-vehicle sales fell from 59.44% in 2024 to 53.27% in 2025, while CNG and EVs gained share. [5] But this is a flow of new sales. Our question concerns the stock of cars accumulated over many years, including older vintages purchased when petrol was more prevalent. Applying the latest sales percentage directly to the entire fleet would therefore be a poor shortcut.

Historical PPAC sales data provides context for older petrol-heavy vintages, but it does not establish today’s city stock composition. [6] I will therefore use 65% petrol for the existing private fleet and test 60% to 70%. This is an assumption, not a measured Bengaluru percentage. A reliable stock extract by vehicle class, fuel and registration year would be the most useful way to improve it.

4. What about all those white cabs?

A Bengaluru reader is likely to think of the sea of white and silver taxis. That observation matters, but two effects work in opposite directions. Commercial fleets may favour white, while high-mileage economics have historically favoured diesel and increasingly support CNG or electric alternatives. Local reports of CNG retrofitting support the direction of that argument, not a precise citywide fuel percentage. [7, 8]

A recent report puts registered cabs at about 2.97 lakh, including independent travel-agency vehicles. [7] The category may include vehicles that are not conventional passenger cars, and the available reporting does not provide a reliable current colour-by-fuel cross-tab. We should say that rather than invent precision.

For a side calculation, assume 55% to 75% of cabs are white and 15% to 30% are petrol-only. At the midpoints:

The cab adjustment is deliberately separate. The exact current registration-class breakdown and any overlap between the reported private-car and cab categories need confirmation. We are treating them as distinct for the working model, but should not count every registered cab as an ordinary passenger car or pretend the categories are perfectly reconciled.

The private-car endpoints are 5.40 lakh and 9.14 lakh. Adding the illustrative cab endpoints gives approximately 5.64 to 9.81 lakh. I would round that to a broad 5.5–10 lakh working range. These are scenario bounds, not statistical confidence limits, and the variables may be correlated.

5. A second route: households

A useful cross-check should be allowed to disagree. Bengaluru’s Comprehensive Mobility Plan household survey reported an average household size of 3.7 and roughly 20% of households with one car. The survey is older and cannot be mechanically projected to 2026, but it provides a starting order of magnitude. [9]

Using a working population of 14.5 million gives approximately 3.9 million households. A 35% car-ownership rate and 1.25 cars per car-owning household imply only about 1.71 million cars, below the registration total. At 45% ownership and 1.4 cars, the result is approximately 2.46 million, close to the registration anchor. The population figure is an approximation informed by contemporary reporting, not an official 2026 census count. [10]

Carry the check all the way through

The question is not simply how many cars exist, but how many are both white and petrol. Applying the same fuel and colour filters to the rounded household scenario gives us a second route to the actual target.

That sits reasonably close to the registration route’s 7.05 lakh private-car estimate. If we add the same illustrative cab adjustment, the household route reaches about 7.0 lakh. The comparison is reassuring, but it is not a second independent confirmation: both routes reuse the petrol and colour assumptions, and the 45% ownership and 1.4 cars scenario was partly selected to reconcile with the registration total.

A genuinely independent check would require a current, geographically matched estimate of car-owning households and cars per owning household. The 2011 Census ownership data and the older mobility survey are useful historical anchors, but neither establishes the 2026 rate. [9, 11] Differences may also arise from business-owned cars, multiple-car households, cumulative registrations, inactive vehicles and mismatched geographic boundaries.

6. What changes the answer?

The value of the model is that we can see which assumptions matter. The table varies the main inputs together. It is a scenario analysis, not a probability model; the low and high cases should not be interpreted as equally likely outcomes.

The white share and petrol share are the most consequential uncertain filters. Better local stock data would improve the estimate more than adding decimal places to the arithmetic. The next useful research step would be a Bengaluru RTO extract by vehicle class, fuel and registration year, followed by a small observational colour sample. The household route could then serve as a genuinely separate stock check.