Container Glass in APAC: What the West Gets Wrong

Talk to a Western OEM about the container glass Asia market and you'll get the same story every time: cheap labour, low automation, a region you sell a turnkey furnace line into and then leave. I've audited plants from Chonburi to Karawang over the past decade and that story is mostly backwards. The furnaces running in a lot of ASEAN plants are newer than what's still limping along in parts of Europe. What's missing isn't equipment. It's the changeover discipline that ties the equipment to output.

That gap matters more in Asia than anywhere else I've worked, because the volume growth doesn't wait for the discipline to catch up. Vietnam, Indonesia and the Philippines have added container glass capacity most years this decade, much of it feeding beverage and pharma packaging lines that used to import finished ware from China or the Gulf. The plants are real. The systems running them often aren't.

Asia's mixed-OEM fleets are the norm, not the exception

In 2017 I audited a two-furnace plant outside Ho Chi Minh City that ran an old Emhart eight-section machine on 1980s cam-timing controls next to a Heye Smart H1 installed four years earlier. Same batch house, same forehearth design philosophy, two completely different job change routines because two different OEM service contracts had trained two different crews. The plant manager knew it was a problem (and yes, both vendors' service reports said their own equipment wasn't the issue).

That's the pattern across most of the region I've seen, including plants sitting inside the AFGM membership footprint across Malaysia, Thailand and Indonesia. Mixed fleets aren't a temporary phase on the way to standardisation. They're the permanent condition, because plants add capacity in stages, from whichever OEM has the best commercial terms that year. Cord and stones get lumped into one "glass quality" line item on plenty of OEM service reports I've read. Cord points at the melter's temperature profile. Stones point at refractory wear. Two different repair budgets, and nobody's separating them at the point the report gets written. A vendor-neutral container glass consultant walks into that same two-line plant without a forming platform to protect, which is the only way to compare a defect on the old line against one on the new line without the report reading like a sales pitch for a rebuild.

ASEAN scale is outrunning changeover discipline

Across the plants we've audited under our Asia container glass work, cross-shift variance on identical SKUs is the number that gets ignored until someone measures it. In plants that haven't systemised the changeover, that variance typically runs 30-60%, meaning the same mould set and the same recipe can produce a materially different first-ware quality depending on which crew ran the swap. I've seen it hit both ends of that range in the same ASEAN plant in the same month.

The shift roles that should prevent this often exist on paper only. On the floor, when a job change runs late and the next SKU is already staged, sign-off gets skipped under time pressure more often than any plant manager wants to admit. The 0600 handover is where it shows up first. On the plants I've audited across the region, that handover misses the previous shift's swabbing data close to 70% of the time, which means the incoming crew is guessing at mould condition instead of reading it.

• Melter operator owns glass level and temperature stability, targeted to within ±2°C
• Feeder operator owns gob weight consistency, targeted to within ±0.3-0.5g
• Hot-end superintendent owns recipe lock and sign-off on any set-point change
• Cold-end inspection technician owns redraw and reject classification, not just the headline reject rate

A furnace doesn't lose you a shift. A changeover nobody's timing does, one crew at a time.

What Europe's carbon rules actually tell you about the real fight

Western consultancies love to frame Asia's disadvantage as regulatory, pointing at Europe's tightening carbon rules as proof that European plants are under more pressure to get efficient. Under EU ETS Phase IV, the Linear Reduction Factor now cuts free allocation by 4.3% a year through 2027, rising to 4.4% from 2028, and the EU Carbon Border Adjustment Mechanism moved into its definitive phase on 1 January 2026, though container glass still sits outside CBAM's product scope, a gap FEVE has already flagged as a competitiveness risk. That's real pressure, but it's aimed at plants already running 50-90% cullet ratios and hybrid oxy-fuel furnace rebuilds. It doesn't tell you anything about whether an ASEAN plant's forming line is losing two hours a shift to a job change nobody's timing.

And that's the part the regulatory conversation skips. O-I's "Fit to Win" programme has idled North American furnace capacity chasing roughly $180-200M in annualised savings, which is a rationalisation story, not an efficiency one. Meanwhile Gulf capacity out of Saudi Arabia and the UAE runs on subsidised gas 30-40% below European energy costs, which is exactly why some of that Gulf output gets routed toward South Asia and East Africa rather than competing head-on into Europe. None of that has much to do with whether a plant in Surabaya has systemised its mould change stage. Regulation shapes where capital goes. It doesn't fix a changeover.

The job change problem is universal: the fix isn't

This is why we built the Job Change Tool the way we did. It's not a generic SMED for container glass rebadge, and it isn't tied to any OEM's forming platform, because half the plants we work with are running two or three OEMs' equipment on the same floor. It maps the changeover to the nine-stage Job Change Lifecycle, from plan through post-mortem, with a named owner at every stage, so the mould shop knows exactly when its window opens and closes instead of guessing off a paper checklist taped to the panel.

Twenty-three minutes. That's the swing we measured section-to-section on one ASEAN line once the checklist replaced the notebook. Not a furnace problem. A handover problem. The same one showing up in the 0600 data gap above. Plants that treat job change as a mould-shop task lose that time every changeover, on every line, and multiplied across a five-line plant running two changeovers a shift, the number stops being a rounding error.

Founder Zaid Hassoneh built his early instincts on this the hard way, moving from the floor at O-I Brisbane in 2005 to Plant Manager by 2019, and the APAC experience behind this work comes from having stood at the section during a bad changeover at three in the morning, not from a case study deck.

The OEM sales story about Asia isn't wrong because the equipment's bad. It's wrong because it assumes the only lever worth pulling is the next machine. The plant outside Ho Chi Minh City I mentioned earlier didn't need a third furnace. It needed someone to walk both crews through the same nine stages and make them own it the same way, shift after shift.