Recently I read a column by David Fickling of Bloomberg Opinion titled “The LNG Boom Is Losing Its Biggest Bet in Emerging Asia.” His point is that emerging Asia, long touted as the future of liquefied natural gas (LNG), keeps failing to show up. Shell has named Thailand, Bangladesh, the Philippines and Vietnam as leading growth markets since its inaugural LNG outlook in 2017, and its latest one, published after the Strait of Hormuz crisis, still sees Thailand’s LNG imports in 2040 at five times today’s level.
Thailand, in Fickling’s words, is “the most prominent example” of the gas industry’s misreading. Gas still generates almost two-thirds of our electricity. Yet the draft Power Development Plan 2026 (PDP2026), released on Sept 8, aims to raise renewables from 20% to 50% of the grid within a decade, and, in one of its four cases, to as high as 89% around 2050. Energy Minister Akanat Promphan has said a core reason is to reduce reliance on “expensive and volatile spot LNG imports.”
Unfortunately, draft PDP2026 does not match the minister’s words, at least for the next 11 years.
Last month my colleagues at Climate Finance Network Thailand (CFNT) and I published “An Irresponsible Path to Net Zero,” a critique of draft PDP2026. What struck us most is that all four “cases” offered to the public are identical from 2026 through 2037 (the plan calls this “Phase 1”). Every case builds the same 50,900 MW of new capacity, including 9,100 MW of new combined-cycle gas plants (up from 6,300 MW in the abandoned PDP2024 draft) plus 540 MW of refurbished lignite at Mae Moh. The cases diverge only after 2037, i.e. after the decisions that matter most have been locked in.
Every case builds the same 50,900 MW of new capacity, including 9,100 MW of new combined-cycle gas plants plus 540 MW of refurbished lignite at Mae Moh. The cases diverge only after 2037, i.e. after the decisions that matter most have been locked in.
Our LNG imports nearly doubled from 829 million standard cubic feet per day (MMscfd) in 2021 to 1,628 MMscfd in January to July 2026, as Myanmar pipeline gas fell by about 41%. The plan’s own chart marks 2,000 MMscfd of gas demand as the threshold for “LNG import risk.” In every case, demand crosses that line in the early 2030s and stays above it through 2037.
Fickling’s sharpest point is that LNG now fails the energy trilemma on every count. LNG was never clean, and these days it isn’t cheap either: just buying the fuel for an existing gas plant can now cost more than building a brand-new solar-plus-storage system from scratch. And the security argument fell apart the moment Iranian drones closed off roughly a fifth of the global market, plunging Asian importers into blackouts and power rationing. Thai households felt this earlier this year, when the spot LNG assumption behind the Ft (the fuel adjustment charge on our power bills) jumped from USD 11.6 to 18.8 per MMBtu.
It seems extremely unlikely that adding 9,100 MW of gas capacity in Phase 1 (2026 through 2037) can help Thailand meet its NDC 3.0 commitment or lower the electricity price due to an even greater reliance on LNG imports.
Since it seems extremely unlikely that adding 9,100 MW of gas capacity in Phase 1 (2026 through 2037) can help Thailand meet its NDC 3.0 commitment or lower the electricity price due to an even greater reliance on LNG imports, new gas additions can only be seen as yet more evidence of the gas lobby’s influence over Thailand’s energy policymaking. I have no doubt that the combination of rooftop solar, batteries, and demand-side measures can replace all 9,100 MW of planned new gas by 2037. Agora and SDG Move recently ran cost-optimization modeling based on draft PDP’s numbers and found that gas capacity could be 14,000 MW lower in 2040 and 5,000 MW lower in 2050, with no new gas needed. The shift also necessitates more flexible contract terms for existing gas-fired power plants, e.g. lower minimum generation and take-or-pay conditions, so that gas can decline instead of curtailing renewables.
Now to Phase 2 (2038 to 2050). The plan in all four cases bets on the same fleet of small modular reactors (SMRs): 300 MW by 2037, expanding thirtyfold to 9,000 MW by 2050, with no fallback. That means up to 30 separate units, in a country whose nuclear regulations do not yet cover power generation.
This is a bet with a poor track record. Of 127 SMR designs worldwide, only two are operating: China’s HTR-PM, which took 11 years from start of construction to commercial operation, and Russia’s floating KLT-40S, which took 13. In the US, the estimated cost of NuScale’s flagship project more than doubled between 2015 and 2023 before it was cancelled.
Draft PDP2026 nonetheless assumes an SMR capital cost of THB 202,164 per kW, lower than what many real projects have seen, and operating cost at only 1% of capex (THB 1,785 per kW per year). Idaho National Laboratory estimates put it at THB 4,012 to 7,344. BloombergNEF estimates SMR power at around USD 372 per MWh, against USD 33 to 75 for utility-scale solar in Thailand today. (Then there is the 2,500 to 3,000 MW the plan reserves for a technology it doesn’t even name, priced at the system average so that it can never look expensive, but I will save that for another time.)
If SMRs arrive late or over budget, as they have almost everywhere, the likeliest stopgap is more gas, and therefore more LNG imports.
Vietnam, Fickling notes, is balking at LNG developers’ demands to guarantee purchase of up to 95% of output for 25 years. Thailand should ask the same question Vietnam is asking: why pay over the odds for plants that cannot even promise cheap or secure power?
I would propose at least two changes to the final plan. First, reopen Phase 1 and present genuine alternative mixes for 2026 to 2037, the period that decides whether Thailand meets its 2035 milestone under NDC 3.0. Second, price the risks the plan already acknowledges, namely LNG price volatility and carbon (its tariff comparison includes no carbon cost at all, even though the Cabinet approved in principle a draft Climate Change Act with a carbon tax last December), and benchmark SMR costs against delivered projects with a cheaper contingency, which on current evidence means more solar and storage.
I would propose at least two changes. First, reopen Phase 1 and present genuine alternative mixes for 2026 to 2037. Second, price the risks the plan already acknowledges, namely LNG price volatility and carbon, and benchmark SMR costs against delivered projects with a cheaper contingency, which on current evidence means more solar and storage.
In short: if the energy minister really wants to cut our reliance on spot LNG as he claims, the final PDP2026 is the place to show it, starting with removing the unnecessary 9,100 MW of new gas in Phase 1.