The short answer: market, spec, process, then supplier
Most HNB buyers interview suppliers first. That is backwards. Pass the market gate, lock the spec, pick the process routes, then interview suppliers. A line is downstream of regulation and product definition: change either, and the line changes with it.
One disclosure first. We are Ccobato, an HNB stick maker that also supplies turnkey lines and equipment, per our own website. We profit when you buy a line; we lose repeat business when it fails. So every gate below is a question we expect to be asked.
In 2022 the EU withdrew the flavored heated tobacco exemption once retail sales exceeded 2.5% of Union tobacco sales [1]. Every flavored line aimed at that market became a question the day the exemption fell. Mordor Intelligence, a third-party market research firm, sizes that market at USD 51.29 billion in 2025, rising to USD 58.53 billion in 2026 [2].
Growth plus moving regulation is why the 17 gates that follow run in the order a buyer faces them. Each ends in a question you can paste into an RFQ.
For your budget, two anchors. Time: our own published figures run from samples in 1–2 months to new capacity within 6 months, against 2–3 years for conventional sheet routes; gate 14 has the full sequence.
Money: no honest single number exists for a line’s cost. Any total quoted before your spec is a placeholder, not a budget. The budget forms gate by gate, and the priciest line item is usually not machinery (gate 15).
1. Pass the market gate before you buy a single machine
Market access sits upstream of everything the machines do: buy the line first and you may own capacity for a product no regulator will admit.
In 2022 the EU withdrew the flavored heated tobacco exemption after retail sales of heated tobacco products passed 2.5% of total Union tobacco sales, with volume up at least 10% in at least five member states [1]. You cannot compute that Union-level threshold from your own sales, so track the legislation instead: the directive was published in November 2022, and member states had to transpose it by 23 July 2023 at the latest [3]. That gap, roughly eight months, was the transition window.
And a flavor ban like this one hits the product, not the machines. A line in transit for flavored sticks is not scrap; it is a line whose spec changed mid-flight. What changed: the recipe, the packaging, the labeling station. The rod maker does not.
The US gate works differently: FDA grants marketing authorization product by product, and only where the net population-level benefits to adult smokers outweigh the risks to youth [4]. It cleared three heated tobacco products in one round in January 2023 [4]. No portfolio-wide approval, no guarantee the next SKU passes.
Southeast Asia moves faster and splits finer. Indonesia raised its minimum retail price benchmarks (HJE) effective January 1, 2025, while excise rates stayed unchanged from 2024 under PMK No. 96/2024 [5]: one policy variable moved, the other did not.
The Philippines went further: under Advisory 2025-1487, sellers of vape and novel tobacco products with medicinal or therapeutic claims must apply for licenses as pharmaceutical establishments [6]. If your sticks will carry health claims, decide the claim before you buy the machines. The claim picks your regulatory track, and the track sets the evidence your factory must produce. A line validated to tobacco standards is not one audited under the safety, efficacy, and quality language the Philippine FDA cites [6].
Model three policy variables per market: the product gate (can you sell it), price tools (excise rates, HJE minimums), and machine licensing (does the equipment need a permit). Each variable has its own source: the regulator’s texts, the tax authority for price tools, the licensing authority for machines. Then mark three states.
Baseline: rules in force stay as they are, like Indonesia’s unchanged excise rates [5]. Tightening means the trigger fires: Indonesia’s HJE rise [5], or the EU 2.5% mechanism [1]. Opening means authorization actually granted, as the FDA’s three January 2023 clearances showed [4]. Size the line for the tightening state, plan production for the baseline, and treat the opening state as a call option.
For markets outside the three above, GCC, Africa, or Central Asia, use the same three kinds of sources this section was built from: the regulator’s own texts, from a Union directive [1] to a national FDA advisory [6]; the revenue authority’s tax and licensing notices, with Indonesia’s PMK as the template [5]; and the trade press as a change alert. In markets where the framework is still forming, as in much of the Gulf, check draft rules and consultation papers, not only laws in force.
The market is rising while all this tightens. Mordor Intelligence’s 2025 and 2026 figures add about USD 7 billion to the category in a single year [2]. Both curves run at once, and the line you buy sits between them.
That is also why our zero-nicotine herbal series exists for markets such as Japan that restrict nicotine: the product and its line match the gate rather than fight it.
2. Read the giants’ capacity map: you are competing for line slots
Buying a line is buying time on someone else’s engineering calendar. That calendar is full.
PMI’s Papastratos Aspropyrgos plant in Greece is one of only three in Europe dedicated exclusively to heated tobacco units, runs 24/7, exports to more than 40 countries, and more than doubled productivity in the last three years, per PMI’s own case study [7]. KT&G’s new factory in Indonesia was scheduled to begin operations in 2026, per a Tobacco Asia report from January 2026 [8]; its expanded Kazakhstan plant already contributed in full last year [8]. The map does not end with those two: PMI makes sticks in Indonesia as well, and JTI’s Ploom and BAT’s glo run their own stick plants across Asia and Europe.
So what does a normal lead time look like? Custom production machinery is built to order and ships in months, not weeks. Counting transport, installation, and commissioning, the total calendar is longer than the build alone. Count backwards from that calendar: a line you want running next year is a conversation you start now.
Then lock the slot, not the price. Pay a deposit only against a slot confirmation with a dated engineering start; a deposit on an undated promise buys nothing. Ask which projects occupy the engineering team right now and when the next opening is; a supplier that cannot answer is selling a date it does not control. Trade speed for priority: giants pay in volume, but their approvals move slowly, so a buyer who arrives with the spec locked, the patent review clean, and payment ready can take a slot the giant is still arguing about.
3. What a complete HNB line actually looks like: six stages, three engineering disciplines
Follow one stick from leaf to carton and a turnkey quote stops being a price and starts being a chain you can audit.
Stage one turns leaf into sheet; stage two forms the rod. PMI’s account of its own factory: leaf ground into cast leaf, rolled into a continuous rod, cut into tobacco sticks measuring 120 mm in length [9]. Materials engineering, then speed engineering. On speed, buyers reach for the wrong benchmark: Hauni’s PROTOS-M4, 2017, runs up to 10,000 cigarettes per minute [10]. The only public HNB reference is an Imperial Tobacco patent filing saying a typical heated tobacco line processes thousands of consumables a minute [11]: thousands, not 10,000. The ceiling sits in the sheet material, not the machine; the next gate shows why. No public HNB yield benchmark exists in anything we reviewed, so the supplier’s yield promise belongs in the gate 13 acceptance contract, not your comparison spreadsheet.
Stage three is filters, stage four assembly, stage five QC, stage six packing. Imperial’s patent walks the back half: a combiner joins the tobacco portion, bore filter, paper tube, and mouthpiece filter from individual buffers, then a packer and a transportation case filler close the line [11]. Filters and assembly are precision engineering: the tolerances decide whether every stick heats the same. QC gets its own gate, number 10. Packing is plain conventional engineering.
The vendor map is thinner than buyers expect, and the thinness is information. Rod making is the only stage with named makers in the public sources we reviewed: Hauni, which holds the conventional benchmark [10], and Montrade, an HTP specialist whose rod maker concept, in its own words, “works for all bobbins or webs” [12]. Sheet making has no publicly named specialist; the only detailed public account of that stage comes from a tobacco company describing its own factory [9], because the majors build it in-house. Expect to extract the real vendor list through the RFQ, stage by stage.
If you are converting a vape line, the overlap is real but shallow. E-liquid filling has no counterpart for stages one and two; those are new builds, not conversions. What transfers: the rod concept is explicitly cross-category, “works for all bobbins or webs” [12]; packing, case filling, and labeling are general-purpose stations your line already runs; and QC transfers its data discipline, not its test targets.
Which stage costs the most? No supplier publishes a stage-by-stage price list, which is why black-box quotes survive. The engineering structure still gives an anchor: the expensive half is the tobacco half. Sheet making runs the longest process chain, rod making is the specialist core, packing is a commodity. Equipment makers see the same split: customers’ biggest investments these days are in HTPs [12]. Demand the stage-by-stage breakdown. A supplier that quotes one total cannot be checked.
We know the first two stages from the inside: cast-leaf HNB sheets, tobacco extracts, and cut stems are part of our supply scope, and our cast-leaf sheet capacity runs at 1,000 tons per year.
4. Pick the sheet route first: cast leaf is fragile, sticky, and decides everything downstream
Section 3 put the speed ceiling in the sheet material, not the machine. One equipment maker puts the material plainly: “It’s fragile and tends to be sticky, so it needs to be treated in a special way” [12]. Fragile and sticky are not two faults; they are two edges of one narrow process window.
Cast leaf starts from tobacco plant parts unsuited to cut filler, such as stems and dust [13]. Ground particles and a binder, kept to the minimal amount of adhesives necessary [14], form a viscous slurry cast onto a moving metal belt to become the tobacco web [13]; the dried film then has almost no binder network holding it together. Casting and drying generate their own scrap: a PMI patent filing puts the reject rate at about 3 to 20 percent [13].
Dry it too little and the film clings to tooling; dry it too much and it cracks. That is what special handling means [12].
The same patent shows the fork: a low-viscosity, high-water slurry can instead form reconstituted tobacco in a process that resembles paper-making [13]. Tobacco Asia counts four HNB-dedicated RTL processes, from paper making to pressure rolling [15].
The difference is where the tobacco ends up. Paper making coats the load on afterward, capped by the coating rate, with the aerosol former on the surface, where it leaches and absorbs moisture [15]. Cast leaf carries the load inside the web: more smoke volume and aroma [15], paid for in fragility and stickiness [12]. A flavor-heavy, high-load stick belongs to the cast-leaf family; coating control and low binder belong to paper making, China’s HNB research hotspot [15]. Pick the route from the product, not the catalog.
5. Lock the stick spec before the quote: 0.5 mm changes every machine (and why retrofits fail)
A quote without a locked stick spec prices a product that does not exist yet.
A third-party patent application describes the IQOS Heat Stick as about 7.2 mm in diameter and about 45 mm in length [16]. That is a patent filing’s wording, not an official brand spec sheet; read it as an anchor. Move the diameter by 0.5 mm. The forming molds change. The cutting tooling changes. The inspection windows change. Every station that touches geometry changes spec.
Where does your own spec start? From the device, not the stick. Pick the device family your sticks must fit and the cavity is your reference; for IQOS-class devices the patent record above supplies the anchor. If the device is your own, you write the spec. Fill in the rest by reverse engineering the reference product: measure diameter, length, and each segment’s length on production sticks, and mine patent filings as a public geometry record.
Lock in this order, and hand the supplier the same list:
- diameter and length
- tolerance band per dimension
- segment count, with each segment’s length and material
- sheet type (section 4)
- heating route (section 6)
A supplier that quotes a full line before asking for these is selling inventory, not engineering.
The retrofit fantasy dies against the same numbers. PMI’s own account: the rod is cut into 120 mm sticks [9], then cut again to the final length in the combiner. Hauni’s PROTOS-M4, a 2017 conventional cigarette maker, runs 10,000 cigarettes per minute [10]; fragile cast leaf cannot take that rate (section 4). Four of the six stages in section 3, sheet, rod, filter, assembly, lock to your geometry and sheet type; what transfers is packing plus QC’s data discipline, two stages at the cheap end: section 3’s cost split puts the money in the tobacco half and packing is the commodity. A retrofit saves the cheap end and rebuilds the expensive core. Ask the seller one question: which stations, exactly, with what samples? Then demand the retrofit quote and the new-line quote side by side; the reuse ratio falls out of the two quotes. Budget for the purpose-built line until proven otherwise.
6. Heating route before line layout: blade, induction, or peripheral
Can you change heating routes after the line is built? Almost never. The route decides the core equipment, the testing regime, and where the money sits. Only the cheap end from the previous gate survives a swap: packing and QC data handling.
With blade heating, a resistive blade sits in the device and doubles as the temperature monitor [9], so the hard engineering, heater alignment and temperature control, lives on the device side. It is the past route, the one that built the category’s defining product generation, and its device-side engineering is the most proven of the three.
Induction puts the heating element inside the stick. The inductor’s field induces a current in a susceptor inside the tobacco [17], and the precision burden lands on the stick line. It is the newest commercial wave: PMI’s ILUMA generation runs on what PMI calls the SMARTCORE INDUCTION SYSTEM, and the equipment makers’ patenting followed the route, with a published Hauni application, publication only with no confirmation of grant and no second-source cross-verification, centered on the exact positioning of flat and strip-shaped susceptor elements [18]. That station gets its own gate next.
Peripheral heating wraps the heat around the outside of the consumable: no blade, no metal in the stick. It is the oldest concept in the category. Philip Morris’s Accord and its Heatbar follow-up both heated from outside, and BAT’s first glo generation took the same approach. Today the route survives in simpler chamber-style devices, a minority position.
The three do not differ by a universal price tag; they differ by where the money sits. Blade concentrates spend in the device’s heater and control. Induction adds a metal part to every stick, so part of the cost rides in the consumable, per unit, on top of a positioning station blade lines do not have. Peripheral keeps the stick cheap and pushes the thermal design into the device. Any head-to-head route pricing you see before the stick spec is locked (the previous gate) is indicative at best.
Before any quote, ask each candidate for its route-to-machine mapping table. A qualified one follows this pattern:
| Route | Key process | Bottleneck station | Test focus |
|---|---|---|---|
| Blade | Heater blade forming and insertion | Heater-to-stick alignment and temperature control | Blade position tolerance; heating profile repeatability |
| Induction | Susceptor forming and placement inside the stick | Susceptor insertion and centering (the next gate) | Susceptor position; field coupling; per-stick heat-up consistency |
| Peripheral | Chamber heater assembly in the device | Heat transfer from the chamber wall into the plug | Wall temperature uniformity; wall-to-plug contact |
Judge any supplier’s table against those four columns. A version that lists machines but names no bottleneck station and no test method per route is a catalog, not a mapping. Send it back before you compare quotes.
We sit on the induction side of this decision. Our susceptor design is a single alloy with three calibrated working temperature bands, 300–320 °C, 321–340 °C, and 341–360 °C. That is our claim, not an independent finding: demand the patent filings, the recognition-rate test data (we claim a stable 90%–100%), and reference sticks from paying customers. A claim of uniqueness that cannot produce those three is a slogan.
7. The hardest station in an induction line: susceptor insertion (plus the cooling section’s patent trap)
The susceptor forms the aerosol at roughly 340 degrees Celsius [17]; off center, the heat lands unevenly. That tolerance is the station.
Ask a supplier which insertion route it uses and why. Straight push-in bends, kinks, or breaks the element, the failure Hauni’s filing sets out to fix [18]. Hauni’s published route, no grant confirmed, seats the segment in a receiving pocket on an insertion body, drives it in with a linear forward movement, then withdraws, leaving the segment behind [18]. PMI feeds segments continuously into a continuous material, an extremely efficient mass-production route in its patent’s words [19]. Pre-assembly fixes the susceptor into a positioning base before the matrix is filled; our own integration patent takes that direction. A supplier that cannot name its route has not built it.
The price gap between induction lines lives at this station: a metal part bought on every stick, and rejects that scrap a finished stick whole. No supplier publishes this station line by line, so make the supplier split it open: equipment, susceptor cost per stick, reject commitment.
The cooling section asks equipment and patent questions at once. PMI’s published application pins the reference structure in numbers: a crimped and gathered polymeric sheet, 50% to 90% longitudinal porosity, materials from PE, PP, PVC, and PET to PLA, cellulose acetate, and aluminium foil [20]. Have counsel start from US20150027474A1: family and status as of August 2026, then claim-chart your element against it. Your room depends on which claimed features your structure avoids; we will not guess the outcome, and gate 16 covers the whole-line clearance. Ask both: forming capability, and freedom to operate.
8. The device side is a temperature-control business, not an assembly business
Fix where the device sits in your turnkey scope before you evaluate a quote for it. The stick line in sections 3 through 7 is one manufacturing system; the device is another, built on consumer-electronics assembly. If device assembly lands in your package, this gate applies to that segment.
The device has one job: reproduce the aerosol window on every unit. PMI’s internal-susceptor filing puts that window at about 340 degrees Celsius [17]. In blade routes the heater is also the sensor [9], so batch-to-batch drift in heater material reads as a temperature error on finished devices.
Start with the material. Ask for the heater’s incoming inspection spec and its supplier-change control. Then ask how the supplier verifies finished-device heating profiles: which sensor and sampling plan, and what profile data it hands over at acceptance. Set those limits from your own product temperature target, not the supplier’s catalog.
Our own device engineering sits in control logic, not assembly. Our patents cover dual-mode dynamic temperature compensation, which feeds center-heating power into the peripheral control as a feedforward term so the radial temperature gap converges, and current-sampling control that counts puffs, stops heating automatically, and guards against dry heating on an empty chamber. A supplier with no temperature story fails this gate before its automation gets discussed.
9. Probe the supplier with glycerin uniformity: NIR shows who engineers the process
Glycerin is the aerosol’s raw material: it protects tobacco from overheating and creates the visible aerosol [9]. Feed it unevenly and sticks run hot and dry, or over-produce.
CN101261221A, granted as CN101261221B, comes from the Zhengzhou Tobacco Research Institute of CNTC, China National Tobacco Corporation’s research arm. Its finding: the amount of glycerin in tested tobacco leaves is directly proportional to the content of the feed liquid, and variance across the batch reflects feeding uniformity [21]. The sample sits under a near-infrared diffuse reflection probe [21]: offline sampling, and it quantifies glycerin, not moisture.
You can put inline control in the RFQ, but never take it on faith. Ask which instrument, how often it samples, on which installed lines. No adoption figure exists in anything we reviewed, so treat the offline device as the baseline and inline claims as unproven until the supplier names hardware and sites.
Acceptance needs a number, and no universal band exists to copy. The patent gives the metric: variance of glycerin readings across a sampled batch. Set the band against baseline data from your approved product, as gate 8 does for temperature, then write it into acceptance. A supplier that cannot name its metric, sampling plan, and band, backed by trial-run data, has never run the measurement.
Put one line in the RFQ: “State the feeding-uniformity inspection method included in the line and the acceptance criteria attached to it.” No answer, no shortlist. A supplier that has never measured glycerin distribution has never engineered for stick-to-stick consistency.
10. Treat QC as its own process stage: inspect, reject, verify
Most buyers read the QC items at the end of a quote as options. Wrong category. QC is a process stage with a closed loop: inspect, reject, store, verify. Gate 9 covered the feeding side; here is the stick end.
The inspection spec is already public. Engineers at China Tobacco Hubei Industrial Co., Ltd. published a machine-vision design for heated cigarette end faces in the Journal of Light Industry (2024): an industrial camera and flash controller capture the end face, a Canny edge algorithm traces the contour, and binarization catches hollowing and looseness. Accuracy on those defects was 99% or higher, the defects were eliminated online, and the design was sized for high-speed line operation [22]. Nobody sells it off the shelf, but the published detail lets any integrator copy it: put the spec in the RFQ and let integrators bid it.
Rejection is a separate leg. A published application (publication only, no confirmed grant, no second-source cross-verification) sequences raw-material, appearance, and end-face vision stations and removes defective cigarettes at a removing opening on the line [23]. Inspection without a rejection station is a suggestion box, not a quality system.
Neither the paper nor the patent names the stage’s cost, and no public source gives the yield gain from adding one. Get the ROI inputs from the quote instead: the stage as its own line item, detection and rejection rates committed in writing, set against your scrap cost.
Verify lives off the line. ISO 20778:2018 is the standardized laboratory puffing and collection regime for heated tobacco products; nearly all published HTP emission studies follow it or the Health Canada Intense regime [24]. It measures your product, not your line. Line records prove every stick is identical; an ISO 20778 lab run proves that identical stick can be measured the way brand audits and regulators do. Write batch-release testing under a standardized regime into acceptance; the same review notes the need for standardized protocols [24].
Run every supplier’s QC list against the loop. Inspect, reject, store, verify. Miss one leg and the yield ceiling is fixed.
11. ATEX first: dust safety is the costliest afterthought
Buyers assume the hard part of an HNB line is process precision. The people who build equipment for tobacco plants see a different wall. Siempelkamp, which builds fans for tobacco plants, states it directly: “The greatest challenges in the tobacco industry lie in protection against dust explosions (ATEX)” [25].
The mechanism most buyers never model is the second blast. A primary explosion raises dust that has accumulated on various surfaces into the air, and the suspended cloud can fuel a significantly more powerful and destructive secondary explosion [26]. That guide is general dust-safety engineering, not tobacco-specific; the physics transfers cleanly. Fine sheet and cutting dust settles on every beam; the first small event is what arms the big one.
Start with your own powder, not a textbook number. Two numbers define a dust: Pmax, the maximum explosion pressure, and Kst, the maximum rate of pressure rise. Kst is what separates them, sorting dusts into St 1 (up to 200 bar m/s), St 2 (200 to 300), and St 3 (above 300), and that class sizes your protection design. Your dust is your blend, your moisture, your fines, so send samples of your actual sheet dust and cut-stem fines to a 20-litre sphere test lab per EN 14034. A lab fee and a report later, you know your class.
Have an explosion-protection consultant or the supplier’s engineers classify the plant per IEC 60079-10-2: Zone 20 where a dust cloud is present continuously or often (inside mills and dust collectors), Zone 21 where it appears occasionally in normal operation (around rod-forming feed points), Zone 22 where it is unlikely and brief (most of the hall). Each zone maps to an equipment category (1D for Zone 20, 2D for 21, 3D for 22), each backed by a certificate under Directive 2014/34/EU. Certification runs through an EU Notified Body (the NANDO database lists them) or an IECEx certification body. Certificates are cheap next to the line; what costs is a machine that fails its inspection after installation.
So the zoning map and the per-zone protection plan go into the RFQ as a mandatory line: “Provide the ATEX zoning drawing for every station handling sheet, dust, or powder, the equipment category per zone, and the protection approach (venting, suppression, or isolation) per zone.” A supplier that treats dust safety as a footnote is quoting you a building that may not pass its own permitting.
12. Build for the Audit, and Check the Machines’ Legal Paperwork
The machine gets you into the building. The documentation gets you the orders. Plan the line as if an auditor will walk it, because one will.
Regulators are moving toward exactly that. In April 2023, FDA proposed new requirements for tobacco product manufacturers covering the manufacture, design, packing, and storage of their products [27]. The proposed rule has not been finalized; it remains in the proposal stage, so treat it as direction of travel, not current law. The Federal Register text says the rule, if finalized, would set requirements across manufacture, preproduction design validation, packing, and storage [28]. Brand owners have already been here for years: BAT’s supplier code lists machinery and point of sale materials suppliers among its Higher-Risk Indirect Suppliers, and it audits that group [29]. Your line is the machinery half of that sentence.
So make the documentation a deliverable, not a favor. DQ/IQ/OQ/PQ validation packages and batch traceability are industry-standard handover items: ask for them as a separate line item in the RFQ, with a price, so the quote has no hidden markup and no hidden gap. A supplier that will not price its documentation package is pricing your audit risk instead.
Then check the machines themselves, because equipment is a regulated object in its own right. The UK runs a licensing scheme for tobacco products manufacturing machines under excise law, with seizure and forfeiture of the machine as a consequence of operating without the license, per HMRC [30]. Before you sign, verify the supplier’s export license record and the provenance of the machines. Illegal equipment cannot be saved by a good process line.
Our own turnkey scope, per our website, includes compliance assistance for TPD and JFRL as part of one-stop delivery, and our certification list reads MSDS, TPD, JFRL, ISO 9001, ISO 14001, and GMP.
13. Design acceptance around the death valley: centerlines, OEE, and data as evidence
Turnkey projects die in hand-off and ramp-up, not at signing. Write your acceptance terms against the known failure modes.
Two case studies from general manufacturing, not HNB, show where the valley is. An Exponent failure analysis found automated machinery that was excessively flexible and susceptible to vibration that caused imprecise drilling and fastening [31]. Structure first: if the frame flexes, no control system recovers your precision. And a Zarpac engineering guide names the number one packaging-line hand-off mistake: lack of documented machine centerlines and documented line settings for each product produced on a line [32]. Undocumented centerlines mean the operators inherit the line by folklore, and every changeover becomes an experiment.
The same two lessons transfer to HNB acceptance cleanly. Put centerline documentation in the contract annex: every station’s setpoints, materials, and changeover settings, written down before handover. Put ramp-up liability in the payment milestones: a yield curve the supplier commits to across the ramp-up period, with money tied to it.
Then fix the metric. Theoretical units per hour is the seller’s number. OEE is the truth, because yield, changeovers, and bottleneck behavior eat the theoretical rate, and unintended consequences almost always negatively impact line OEE [32]. Accept no quote that leans on UPH alone.
The third acceptance surface is evidence. Your detection systems are your third party: the machine-vision logs and rejection records described in section 10 double as acceptance data. If the line inspected, rejected, and stored every defect during the acceptance run, then speed and yield claims either match the data or they do not. Write that into the contract: acceptance decisions run on detection-system data, not on the supplier’s summary.
Three clauses, one goal: document the centerlines, pay against the ramp, and accept on data.
14. Bridge with a pilot line: ask for two quotes, not one
A pilot line is not a luxury. The Chemical Engineer, the IChemE publication, puts the engineering consensus plainly: pilot plants are an essential tool for de-risking process scaleup [33]. Effects that are invisible at lab scale can derail a full-scale line, and the standard defense is to find them on small equipment first.
Buying a full line while the product is still moving is a classic capital mistake: you would be scaling a target that is still changing. So bind the pilot decision to the supplier decision. Ask every candidate for two quotes: a pilot line and the full line, plus an equipment reuse list showing which pilot assets carry into the full line. A supplier that cannot produce both has never run a phased project.
This is also a filter with a high hit rate. The supplier that resists the pilot quote is the supplier that expects you to buy the whole risk in one payment.
The bridge is real time, not lost time. Our own delivery pattern, per our website, runs samples to small batches in 1–2 months, mass production in 4–6 months, and new capacity within 6 months, against the 2–3 years our site attributes to conventional sheet-based routes. Our figures, stated as such.
15. The most valuable line item is the recipe package, not the machinery
Machines are the visible half of a turnkey deal. The invisible half is worth more: the process package.
Equipment can be bought. Recipes cannot. Reconstituted sheet formulation, coating recipes, and the process parameters that match a substrate to a heating curve are accumulated assets, not catalog items. They never appear on a machine list because they were not built in a machine shop. Consider the sheet: one recon supplier’s product contains 90 percent tobacco and the minimal amount of adhesives necessary [14], made on a band cast process also known as slurry-type recon [14]. Those numbers were earned on pilot equipment, the same pilot plants the scaleup literature treats as essential tools [33]. A competitor can buy the same band cast machinery and still be years behind the recipe.
Two hard questions separate the machine seller from the capacity seller. Does the supplier run its own pilot line, and how detailed is the process support clause in the contract? No pilot line means no place where recipes were developed; a thin clause means no commitment to transfer what it knows.
Our answer to the recipe question, per our website: a 30+ year flavoring team, a tobacco extract base in Yunnan, and a CTO with 20+ years in flavors and fragrances. Our company’s own statements. When you compare quotes, price the process support clause first and the machine list second. That is the line item that decides when the line actually makes salable product.
16. Clear the patent field before you wire the deposit
Patent risk is asymmetric, and the asymmetry decides when you run the clearance.
Look at how it ended for the giants. Philip Morris and BAT fought what JUVE Patent called one of the biggest global patent battles, starting in 2018 when Philip Morris sued BAT in Japan over its heated tobacco products [34]. It ended in 2024 in a global settlement, to the mutual satisfaction of both parties [34]. A giant can absorb six years of multi-jurisdiction litigation and buy the ending. A company entering the category cannot. That is our reading of the pattern, not a legal finding, but the arithmetic is simple: the settlement was paid out of litigation budgets you do not have.
The blockade is an industry-level problem, not a niche one. Yunnan has set up the province’s first industrial technology research institute for heated tobacco core materials and auxiliary materials in Yuxi, with a stated mission of breaking through foreign patent barriers [35]. When a state tobacco system builds an institute for patent clearance, the field is dense enough to justify an institution.
The density shows up at the component level. Even one structural part, the aerosol-cooling element, sits behind its own published patent application: a crimped and gathered polymeric sheet, optionally polylactic acid, filed by PMI [20]. Multiply that by every segment, every heating route, and every device feature, and the clearance question stops being legal trivia. It becomes a purchasing decision.
So run FTO before the deposit, not after. For the exact process route your line will use, obtain a freedom-to-operate opinion from counsel. Make it the first item in your supplier due diligence, ahead of price. A supplier that cannot discuss FTO for its own route is asking you to buy its risk.
Our own position, per our website: 30+ invention and PCT patents with FTO analysis completed. Our claim, stated as such.
17. The two-way interview: your veto checklist and their questions
Every gate in this article collapses into one conversation. Run it in both directions.
A good supplier opens with questions about your target market and your stick spec, not with a machine price list. The questions it asks are its engineering biography: whether it has worked in your target markets, whether it knows your sheet route, whether it cares about your spec before quoting. Watch what it asks first.
Then run your veto checklist. These are pass-fail questions, each one backed by a gate earlier in this article. Do you have a real volume production case we can visit? Walking the floor beats every brochure. Can you show export licenses and equipment provenance? Remember the UK machine licensing scheme, where an unlicensed tobacco manufacturing machine faces seizure and forfeiture [30]. Do you run your own pilot line, and will you quote both a pilot and a full line with a reuse list? Is feeding-uniformity inspection inside the scope and the acceptance criteria? Will you price the validation documentation package as its own line item? Will you sign a ramp-up yield curve tied to payment milestones?
Ask one process question in the supplier’s own language: have you run cast leaf in production? It is fragile and tends to be sticky, so it needs to be treated in a special way [12]. The supplier that has handled it answers with specifics; the one that has not answers with adjectives.
Red flags are a shorter list. Only device experience, no stick experience. Using customer lines as its validation site. Suspected refurbished equipment. A litigation history on process patents. No export record or equipment provenance at all. Any one of these ends the interview.
Close with the question that matters more than the line count: who developed the supplier’s process? Bought-in process knowledge walks out the door with its owner. In-house process stays in your factory.
We run this interview from the seller’s side, as Ccobato: 400 million sticks sold annually across 42 countries, a business model that spans OEM/ODM, private label, co-packing, core materials, technology licensing, joint factory establishment, and HNB production line and equipment supply, offered as complete or partial turnkey packages.
The Master Checklist at a Glance
Seventeen decision gates, one row each, distilled from the sections above into a four-column table you can paste into an RFQ. Nothing here is new; everything here was argued earlier.
| Gate | Check | Ask in the RFQ | Walk away if |
|---|---|---|---|
| Market access | EU 2022/2100 trigger (retail sales > 2.5% of Union sales), FDA authorization, local policy moves | Which markets and policy scenario is this line designed for? | No answer on regulation |
| Capacity and lead time | Engineering team’s current project load; dated slot confirmation | Which projects occupy your engineering team right now, and when is the next dated slot? | Cannot name its load or a dated slot |
| Stage-by-stage pricing | Six stages priced separately; black-box total detection | Break the quote into the six stages and price each separately | One total, no breakdown |
| Stick spec | Diameter, length, tolerance, segments locked; 0.5 mm changes molds, cutters, inspection windows | Give the spec-to-machine mapping for our exact dimensions | Supplier never asks for your spec |
| Sheet route | Cast leaf fragile and sticky; casting reject rate about 3–20% | Which sheet process does the line assume? Show samples from that route | No answer on fragile, sticky cast leaf |
| Heating route | Blade, induction, or peripheral | Provide the route-to-machine mapping: route, process, bottleneck, test | No mapping table |
| Susceptor insertion and cooling | Insertion concept; cooling section forming and centering | How do you position and verify the susceptor? What is the cooling section plan? | No insertion concept |
| Device temperature control | Heater incoming control, control loop; about 340 °C aerosol window | How is heater incoming quality checked and the loop held? | No temperature story |
| Glycerin uniformity | Feeding uniformity (offline NIR sampling) | State the feeding-uniformity inspection method and its acceptance criteria | Not in scope |
| QC chain | Inspect, reject, store, verify; end-face vision accuracy 99% or higher | Walk through the four legs on your line, priced as line items | QC quoted as an add-on |
| ATEX | Dust zoning per station and per-zone protection | Provide the ATEX zoning drawing and protection approach per station | Treated as a footnote |
| Audit paperwork and machine legality | DQ/IQ/OQ/PQ, traceability; UK machine licensing (seizure and forfeiture); export record, provenance | Price the documentation package as a separate line item; show export licenses and provenance | Will not price its docs; no record, no provenance |
| Acceptance terms | Centerlines, OEE (not UPH), ramp-up curve | Centerlines in the annex, ramp-up in payment milestones, acceptance on detection data | Refuses data-based acceptance |
| Pilot bridge | Pilot plus full line, reuse list | Quote both lines with an equipment reuse list | Cannot quote both |
| Recipe package | Recipe and process support clause | Detail the process support clause | Machines only, no process package |
| FTO | Freedom to operate for the chosen route; 2018–2024 PMI and BAT settlement | Show or fund an FTO opinion for the chosen route | Cannot discuss FTO |
| The interview | What the supplier asks you | Ask them what they ask you | Opens with a price list |
FAQ: turnkey HNB production lines
What does a turnkey HNB production line include?
Six stages: sheet and tobacco section, rod making, filter section, assembly, QC, and packing (section 3). The heavier half is the recipe and process package that decides when the line makes salable product (section 15).
How much does a turnkey HNB production line cost?
No honest single number exists. Any total quoted before your stick spec is locked prices a product that does not exist yet: move the diameter by 0.5 mm and the molds, the cutters, and the inspection windows all change (section 5). No supplier publishes a stage-by-stage price list, which is why black-box totals survive (section 3). Demand the stage-by-stage breakdown. A supplier that quotes one total cannot be checked.
How long from order to production?
No independent industry benchmark exists in the public sources we reviewed, so every number you hear, ours included, is a supplier claim. What moves the total lead time is predictable. Stick spec: until it is locked, the molds and cutters have no dimensions to make, and no supplier can start (section 5). Heating route: it decides the core equipment, and changing routes means changing the core, not adjusting the schedule (section 6). Target-market regulation: the market gate sits upstream of the line, so authorization timing caps how soon a finished line can ship product (section 1). Our own published figures: samples and small batches in 1–2 months, mass production in 4–6 months, new capacity within 6 months, versus 2–3 years for conventional sheet routes (section 14). Calibrate instead of trusting: demand dated slot confirmations from two or three suppliers and compare them, and write the committed number into your acceptance terms.
Do I need a license for the machines?
In the UK, yes. HMRC licenses tobacco products manufacturing machines under excise law, and an unlicensed machine can face seizure and forfeiture [30]. For any other jurisdiction, run three checks. First, pull the machine licensing and excise rules from your target country’s tobacco authority, customs, or tax authority website; the HMRC notice shows what such rules look like. Second, demand a target-market compliance matrix from the supplier, as listed in the documents question below. Third, have local counsel confirm both the machine rules and the product rules before you sign.
What documents do I need to evaluate an HNB turnkey supplier before signing?
Ask for a stage-by-stage price breakdown, a route-to-machine mapping table, a validation documentation package priced as its own line item, export licenses and equipment provenance, an FTO opinion for the chosen route, a ramp-up yield curve tied to payment milestones, and a target-market compliance matrix that maps your market’s rules to the machines and documents that satisfy them (sections 3, 6, 12, 13, 16).
What questions should I ask to tell a real turnkey supplier from a machine reseller?
Two hard questions: does it run its own pilot line, and how detailed is its process support clause? One process question: has it run cast leaf in production? Resellers answer in adjectives, suppliers in specifics (sections 15, 17).
How do I know a turnkey HNB supplier’s certifications will hold up in my target market?
Ask for the standard set: MSDS, TPD, JFRL, ISO 9001, ISO 14001, and GMP (section 12). Then run the comparison yourself. List the standards your target market’s rules actually require, line them up against the supplier’s certification list, and write any missing standard into the RFQ as a condition. A certificate your market does not require buys nothing; a missing one blocks everything.
Our turnkey scope, per our website: flavor development and formulation behind the recipe package (section 15), manufacturing through the pilot line bridge (section 14), and TPD and JFRL compliance assistance from the certification list (section 12). Bring us your target market and your stick spec; the interview starts with our questions, not a price list.
References
[1] Commission Delegated Directive (EU) 2022/2100: withdrawal of certain exemptions in respect of heated tobacco products. https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32022L2100
[2] Heated Tobacco Products Market Size, Share & 2031 Growth Trends Report (Mordor Intelligence). https://www.mordorintelligence.com/industry-reports/global-heated-tobacco-products-market
[3] Commission Delegated Directive (EU) 2022/2100: withdrawal of certain exemptions in respect of heated tobacco products. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32022L2100
[4] Three Heating Products Authorized in the U.S. (Tobacco Reporter, Jan 26 2023). https://tobaccoreporter.com/2023/01/26/fda-authorizes-three-new-heated-tobacco-products/
[5] Indonesia’s Vape Excise Revenue Rises 7.38% in 2025 to $170.4M Amid Broader Tobacco Excise Decline (2Firsts, 2026-01-27). https://www.2firsts.com/news/indonesia-e-cigarette-tax-revenue-reaches-284-trillion-rupiah-in-2025-up-738-from-2024
[6] Philippines Requires Licenses for Vape Products With Health Claims (Tobacco Reporter). https://tobaccoreporter.com/2025/12/22/philippines-requires-licenses-for-vape-products-with-health-claims/
[7] Papastratos 180 shift: From legacy to leadership (PMI case study). https://www.pmi.com/sustainability/case-studies-and-market-stories/papastratos-shift-legacy-leadership-greece
[8] Record Sales and New Growth for KT&G (Tobacco Asia, 2026-01-13). https://www.tobaccoasia.com/news/record-sales-and-new-growth-for-kt-g/
[9] Manufacturing the Tobacco Sticks (PMI Science, Scientific Update). https://www.pmiscience.com/en/news-events/scientific-update-magazine/manufacturing-the-tobacco-sticks-/
[10] Hauni’s PROTOS-M4 for the 10000 CPM Segment (Tobacco Asia, 2017). https://www.tobaccoasia.com/manufacturing-news/hauni-protos-m4-for-the-10000-cpm-segment/
[11] WO2024223579A1 – Heated tobacco system, device and consumable (Imperial Tobacco Ltd). https://patents.google.com/patent/WO2024223579A1/en
[12] Poised for Growth (Tobacco Reporter, April 2023), Montrade HTP equipment. https://tobaccoreporter.com/2023/04/01/poised-for-growth-5/
[13] WO2017202538A1 – Method for the preparation of a cast sheet of homogenized tobacco material (Philip Morris Products SA). https://patents.google.com/patent/WO2017202538A1/en
[14] In Demand (Tobacco Reporter, October 2023), reconstituted tobacco suppliers. https://tobaccoreporter.com/2023/10/01/in-demand/
[15] Reconstituted Tobacco’s Growing Popularity in China (Tobacco Asia / TobaccoChina Online, 2020-06-30). https://www.tobaccoasia.com/features/reconstituted-tobacco%E2%80%99s-growing-popularity-in-china/
[16] EP4275522A2 – SMOKING TOOL (Intellectual Discovery Co Ltd). https://patents.google.com/patent/EP4275522A2/en
[17] US20170086508A1 – Aerosol-generating article with internal susceptor. https://patents.google.com/patent/US20170086508A1/en
[18] Manufacture of stick-shaped smoking article products (DE102019126621A1, Hauni / Körber). https://patents.google.com/patent/DE102019126621A1/en
[19] CN107529812B – Method for producing an inductively heatable tobacco rod (Philip Morris Products SA). https://patents.google.com/patent/CN107529812B/en
[20] US20150027474A1 – Aerosol-generating article having an aerosol-cooling element. https://patents.google.com/patent/US20150027474A1/en
[21] CN101261221A – Near-infrared tobacco feeding uniformity testing device.
https://patents.google.com/patent/CN101261221A/en
[22] Design of a quality inspection system for heated cigarette ends based on machine vision. http://qgxb.zzuli.edu.cn/zzqgxb/en/article/doi/10.12187/2024.06.012
[23] Heating non-combustible cigarette on-line quality detection equipment and method and production equipment (CN110663995A). https://patents.google.com/patent/CN110663995A/en
[24] Aerosol Emissions from Heated Tobacco Products: A Review Focusing on Carbonyls, Analytical Methods, and Experimental Quality. https://pmc.ncbi.nlm.nih.gov/articles/PMC10747376/
[25] Tobacco industry – Siempelkamp Industrial Fans GmbH. https://industrial-fans.siempelkamp.com/en/branchen/tobacco-industry/
[26] Dust Explosion Safety: Best Practices for Industrial Environments (ATEX.CENTER Europe). https://atex.center/dust-explosion-safety-best-practices-for-industrial-environments/
[27] FDA’s Proposed New Requirements for Tobacco Product Manufacturing Practices – 04/12/2023 (CTP Newsroom). https://www.fda.gov/tobacco-products/ctp-newsroom/fdas-proposed-new-requirements-tobacco-product-manufacturing-practices-04122023
[28] Requirements for Tobacco Product Manufacturing Practice (Federal Register, proposed rule 2023-04591). https://www.federalregister.gov/documents/2023/03/10/2023-04591/requirements-for-tobacco-product-manufacturing-practice
[29] BAT Supplier Communities: Supplier Code of Conduct and supplier audit programmes. https://www.bat.com/sustainability-and-esg/communities/supplier-communities
[30] Excise Notice 2004, Tobacco Duty: tobacco products manufacturing machine licensing scheme (HMRC). https://www.gov.uk/government/publications/excise-notice-2004-tobacco-duty-tobacco-products-manufacturing-machine-licensing-scheme/excise-notice-2004-tobacco-duty-tobacco-products-manufacturing-machine-licensing-scheme
[31] Why Did Automated Manufacturing Machinery Fail to Achieve Performance Requirements? (Exponent). https://www.exponent.com/case-studies/why-did-automated-manufacturing-machinery-fail-achieve-performance-requirements
[32] Avoiding The 3 Deadly Mistakes Of Packaging Line Hand-offs & Start-ups (Zarpac). https://www.zarpac.com/avoiding-the-3-deadly-mistakes-of-packaging-line-hand-offs-start-ups/
[33] What is Your Pilot Plant For? (The Chemical Engineer, IChemE). https://www.thechemicalengineer.com/features/feature-series-the-process-scaleup-journey-5-what-is-your-pilot-plant-for/[34] Philip Morris and BAT end global patent clash over heat-not-burn cigarettes. https://www.juve-patent.com/cases/philip-morris-and-bat-end-global-patent-clash-over-heat-not-burn-cigarettes/
[35] Yunnan’s first research institute for the industrial technology of heated tobacco core materials and auxiliary materials was established (China Tobacco Online).
https://www.tobaccochina.cn/html/rdgz/645110.shtml