Smoke-Free Transition · A read of the public data

How fast is PMI moving from cigarettes — and where is the trend pointing?

Four short reads on shipment volumes, revenue mix, market segmentation and launch readiness — built end-to-end on publicly disclosed PMI (2016–2025), WHO and World Bank data.

By Augusto Parrino · Built for Sr. Analyst, Business Intelligence (req PMIPMIGB25517)
Where the numbers come from. Every figure in this report is from a public source: PMI's own annual reports for revenue and shipment volumes, the World Health Organization for smoking and tax data, and the World Bank for GDP and urbanisation. Nothing is estimated or simulated. The full source list is at the bottom of the page.
§

The five numbers to remember

If you only read this section, take these home.

Smoke-free revenue (2025)
41.5%
Smoke-free revenue (2016)
2.7%
Cigarettes sold vs 2017
-20.3%
IQOS sticks sold vs 2017
+317%
Year smoke-free passes 50%
≈ 2032
I

How fast is PMI moving away from cigarettes?

Smoke-free products went from under 3% of PMI's revenue in 2016 to 41.5% in 2025. Cigarettes shipped are about 20% lower than 2017, while IQOS sticks have grown more than four-fold over the same window.

What stands out: revenue is shifting faster than volume because each IQOS stick sells at a higher price than a cigarette. In 2025 IQOS sticks were about 20% of total stick shipments (151 bn vs 607 bn cigarettes), but smoke-free as a whole pulled 41.5% of revenue. Some of that gap is product mix — smoke-free also includes oral nicotine (ZYN) and e-vapor — but the directional read is clean: each smoke-free unit earns materially more than a cigarette, and that price premium is where the margin tail lives.
II

If today's trend keeps going, when does smoke-free overtake cigarettes?

Ten years of disclosed revenue (2016–2025) fit to a degree-2 curve, projected to 2032. The shaded area is the realistic range — what's plausible, not a guarantee.

The headline: the central trajectory now reaches 50% of revenue by 2030, with a plausible band of 42% – 58%. The 50% milestone lands around 2032.

What's interesting: an earlier version of this report, fit on only 2016–2023, projected the 50% crossing for ~2028. Adding the 2024 (39.0%) and 2025 (41.5%) actuals — both about +2.5 pp/year — has flattened the curve. The early IQOS-launch years (2016 → 2017 jumped ten points) had imprinted an upward curvature that the more recent, steadier growth doesn't sustain. Worth running every reporting cycle — the trend is real but more linear than the early years suggested.
III

Which markets behave alike?

Most launch plans group countries by region — Europe, LatAm, Asia. I grouped them instead by smoking habits and economy and let the algorithm decide. Bubble size shows how urban each country is.

Group Smoking rate (%) GDP per person (US$) Urban (%) Cigarette tax (%)
Mature heavy-smoking 27.9 28,835.9 78.5 63.1
Emerging, lower-income 24.6 9,864.8 62.1 46.6
Urbanised mid-income 17.8 17,178.7 87.7 49.8
Wealthy, low-smoking 16.6 63,000.2 85.4 53.2
What this surfaces — and what it doesn't: the four groups don't map cleanly to regions. The "Emerging, lower-income" group spans four continents (Indonesia, the Philippines, Poland, Russia, South Africa). The "Mature, heavy smoking + high tax" group bundles France, Greece, Italy, Spain, Czech Republic, South Korea — and Turkey, despite Turkey's much lower income, because the smoking and tax profile pulls it in.

Caveat I'd flag in a meeting: with 24 markets and 4 features, this is a useful first cut, not a definitive map. Japan ends up in "Urbanised mid-income" alongside Argentina and Brazil — defensible mathematically (high urbanisation, moderate smoking) but a stretch commercially. Use clustering to challenge regional assumptions, not to replace them.
IV

Which markets look most ready for IQOS next?

I compared every market to the 19 countries where IQOS is already on sale (snapshot end-2023, on purpose — see the validation note below), scoring how similar each one looks on smoking rate, income, urbanisation and cigarette tax. Markets shown in red are not yet launched.

Markets PMI hasn't entered yet — ranked by similarity

Market Similarity (%) Smoking rate (%) GDP per person (US$)
USA 97.8 15.6 76,657.2
Turkey 95.1 31.1 10,897.8
Australia 92.5 12.4 65,169.5
Brazil 72.6 12.5 9,281.3
Argentina 70.4 24.3 13,962.2

What pushes a market's score up or down

What we look atEffect on scoreRelative weight
More people living in cities↓ pushes score DOWN
Higher GDP per person↑ pushes score UP
Higher cigarette tax↑ pushes score UP
Higher smoking rate↑ pushes score UP
The catch — and the more useful read: the model says these unlaunched markets look a lot like the ones where IQOS already wins. But every one of them is blocked by tobacco-product regulation, not by commercial fit:
  • USA — ITC limited exclusion order (Sept 2021) over a Reynolds patent dispute removed IQOS from the market; PMI bought back the US rights from Altria (deal Oct 2022, transition complete 30 Apr 2024) and restarted with an Austin direct-to-consumer pilot on 27 Mar 2025.
  • Australia — TGA confirmed the prohibition of heated tobacco in August 2020 (nicotine kept as Schedule 7 under the Poisons Standard).
  • Brazil — ANVISA RDC 46/2009 banned electronic smoking devices; RDC 855/2024 (Apr 2024) reaffirmed and explicitly extended the ban to heated tobacco products.
  • Argentina — Resolución 565/2023 of the Ministry of Health (27 Mar 2023) prohibits the commercialisation of heated tobacco; Massalin Particulares suspended its USD-300M Merlo-plant expansion in response.
  • Turkey — combined import and production ban makes commercial sale of heated tobacco unviable; PMI does not officially commercialise IQOS in the market.
Read with that overlay, the model output is more useful: it identifies markets where the commercial profile is right but the regulatory door is shut. That changes the analyst's question from "should we enter?" to "what would have to change in the regulatory landscape — or our route to market — for that door to open?"
Post-hoc validation — the USA case: I trained the model on the end-2023 snapshot of IQOS availability on purpose, so the unlaunched-market predictions could be checked against subsequent reality. The model gave the USA the highest readiness score among unlaunched markets. On 27 March 2025, PMI relaunched IQOS in the USA via a direct-to-consumer pilot in Austin — its first commercial re-entry after the 2021 ITC exit. In other words, the market the model picked as "looks most like one we already won" is the one PMI did, in fact, move on first. One data point isn't proof the framework generalises, but it's the right kind of signal: prediction → action → outcome, with the model written down before the action happened.
✻

What this report is — and what it isn't

A few honest notes on scope, so the charts read the right way.