90,000 km of Petrol vs 3,000 km of Electrons

I have been logging every fuel fill for my Hyundai Creta since 2017. Eight years, 180 fills, ₹7.28 lakh, and 18 tonnes of CO₂. Twenty months ago, I bought an electric bike and started logging that too. Here is what the data actually says.

I am, by most measures, a responsible car owner. My Hyundai Creta has been with me since late 2017, and I have logged every fuel fill, every service, without exception. I know its lifetime average efficiency to one decimal place (10.7 km/L, since you're asking). I have an 8-year dataset that would make a fleet manager weep with joy.

None of this happened by accident. I wrote about the habit itself in an earlier post, The Lost Art of Documentation, and the argument is simple: you cannot study what you did not record. Memory is unreliable, impressions are selective, and the story you tell yourself about how something went is almost always tidier than what actually happened. The logbook does not lie. It also does not flatter.

So when I bought an Ultraviolette F77 Mach 2, an electric motorcycle, about 20 months ago, I started logging that too. And like any data-obsessed person who suddenly has two very different vehicles, I wanted to compare them. Not emotionally. Not with manufacturer claims. With numbers.

This is that comparison. I will warn you upfront: it is a bit unfair, mildly humbling, and does not end with a clean moral. Which is exactly why it is worth doing.

First, the scale problem

The car has done 83,846 km since I picked it up in late 2017 with 6,435 km on the clock. The bike has done 3,000 km across 20 months. Comparing these two directly is like comparing a decade of home-cooked meals to one restaurant dinner and concluding you have solved nutrition. The sample sizes are absurdly different.

The usage patterns are also different in kind, not just scale. The car does everything: longer drives, trips, carrying things, the occasional highway run. The bike, as a two-wheeler in Bangalore, absorbs shorter commutes and city runs, the kind of distances where a petrol car burns most of its fuel warming up and idling. So wherever I can, I will normalise. Where I cannot, I will say so clearly.

That cost-per-km gap is the number that stops you mid-scroll. ₹8.68 vs ₹0.39. The bike costs 22 times less per kilometre to run, in energy terms alone. If I had done those 3,000 km in the car, I would have burned through roughly 280 litres of petrol and spent about ₹28,600 at current prices. The bike's electricity bill for the same distance: ₹1,165, give or take.

At this point the EV evangelist in the room is pumping their fist. Hold on. We have not looked at what the car has emitted.

The carbon account

Petrol emits roughly 2.31 kg of CO₂ per litre burned. The car has consumed 7,892 litres. The math is uncomfortable: 18.2 tonnes of CO₂, emitted over 8 years, mostly in Bangalore traffic.

The bike, over 3,000 km, has saved an estimated 196.81 kg of CO₂ compared to a petrol equivalent. That is meaningful, roughly the emissions from driving the car for 905 km. But it is also 1% of what the car has produced in its lifetime. The scale mismatch is real.

"The bike has saved 196 kg of CO₂. The car has produced 18,200 kg. One is a good start. The other is the context."

This is not an argument against EVs. It is an argument for not letting a small sample of clean kilometres absolve a much larger history of dirty ones. If the bike runs for the same 8 years at comparable usage, it will make a significant dent. But we are at the beginning of that story, not the end.

What efficiency actually means

The car averages 10.7 km/L. That is a respectable number for a petrol car in Indian city conditions, not brilliant, not embarrassing. The bike does 18.02 km/kWh. These are not directly comparable units, but if you run the energy equivalent through standard conversion factors, the bike is about 4.5 times more energy-efficient per kilometre than the car. Which is roughly what physics predicts. Internal combustion engines waste most of their fuel as heat. Electric motors do not.

There is one number from the bike's data that I find genuinely interesting: 12.63 kWh recovered through regenerative braking out of 166.45 kWh total consumed. That is about 7% of energy that would otherwise be lost as brake heat, fed back into the battery. It is not transformative, but it is the kind of quiet efficiency that petrol cannot even attempt. The car has no equivalent column in its logbook.

That chart matters for the EV argument more than almost anything else. The car's fuel cost went from ₹69–75/L in 2018 to ₹102/L today, a 37% increase. Over that same period, electricity tariffs in Bangalore have risen, but nowhere near as steeply. The gap between petrol and electricity as an energy source has widened in the EV's favour, and there is no reason to expect that to reverse.

The ₹1.36 lakh I have paid extra versus 2017 prices, purely due to petrol price inflation, would have bought a very decent set of accessories for the bike. Or roughly 200,000 km of electricity at current Bangalore tariffs. The fuel savings compound faster than most people expect.

The honest comparison table

What the car data says about my actual life

Eight years of fuel logs reveal things you do not consciously track. My fill frequency dropped sharply in 2020, lockdown visible in raw data. December is consistently my heaviest travel month across all 8 years, 24 fills in December versus 11 in June. My average tank gets me 468 km. I have spent ₹7,278 a month, every month, for 100 months, just keeping the car moving.

This is what documentation gives you: not just the facts, but the pattern underneath the facts. I could not have told you any of this from memory. Memory would have given me a rough sense of "I drove a lot in some years and less in others." The log gives me exactly which months, exactly which years, and the price I paid per litre on each of the 180 occasions I stopped to fill up. Every. Single. One.

None of this makes the car a bad decision. It was and is the right vehicle for what I use it for. But 8 years of logged data has a way of making abstract spending very concrete. ₹9 lakh on fuel and servicing, one fill at a time, looks different from seeing it as a single number. The single number is easier to ignore. The log does not let you.

The bike does not replace the car. They serve entirely different use cases, a two-wheeler for city runs versus a car for longer trips and variable loads. But every kilometre the bike absorbs is a kilometre the car does not have to. And at ₹8.68 vs ₹0.39 per km, the math on that substitution is fairly straightforward.

So what does it actually tell us?

EVs are cheaper to run. Significantly, not marginally. If your use case fits, relatively predictable range, charging access, city-heavy usage, the running cost argument is essentially settled. The data does not leave much room on that one.

What EVs are not, at least at 3,000 km and 20 months in, is a clean conscience. The manufacturing footprint of an EV, the grid's coal dependency, the battery lifecycle: none of that disappears because the tailpipe does. The 196 kg of CO₂ saved is real. The 18.2 tonnes already in the atmosphere from the car is also real. One does not erase the other.

The more interesting question, which 3,000 km of data cannot yet answer, is what happens at 30,000. At 60,000. Whether the efficiency holds, whether the battery degrades on a known curve, whether the grid gets cleaner over time and makes every historical kWh slightly less guilty in retrospect. The car gave me 8 years of data before I could write this. The bike has given me 3,000 km and a direction.

The data will tell the rest of the story. I just have to keep logging it.