What shape could this business take?

Every option the memo evaluated varies three things: what form the product takes, which route it travels, and how the plant is configured. Underneath all fifteen candidates carried into the case sits one shape, unexamined because it was never named. A project company buys raw biogas at the mill gate, upgrades it, moves it across two borders, and sells it delivered to a Singapore buyer. Compressed or liquefied, pipeline or ship, clustered or per-mill, it is the same business each time: own the molecule from the fence to the meter, and earn the difference.

That shape has one property the memo establishes in detail and then treats as a fact of nature. The molecule has to cross a piece of infrastructure owned by the two companies whose gas it displaces, and the memo's own conclusion is that they have no reason to let it. Nine conditions hang off that shape. Two of them are rated long shots. One of them can end the case on its own.

So the question here is not which option, but which shape. Six are worked below. The first four share a property worth naming in advance: each one removes at least one of the memo's nine conditions entirely, rather than improving the odds on it.

The shape. Stop seeking third-party access on the Trans Thai-Malaysia line. Sell the biomethane to PTT, injected into PTT's own Southern network at or near Chana in Songkhla, and let PTT do whatever it does with a molecule of gas. The green attribute travels separately, by mass balance or book-and-claim, to a Singapore buyer, matched against an equivalent volume that PTT or PETRONAS already delivers to Singapore through the line that already runs.

Why it changes the answer. The memo's argument for refusal is exact and it is an argument about one shape: both owners sell the gas this project displaces, so neither will carry it for a rival. That holds when the project is a competing shipper. It stops holding when the project is a supplier. Under this shape PTT is not asked to carry a competitor's molecule past its own customer. It is offered a low-carbon molecule to buy, into a system the memo already establishes has room, since even all Southern mills aggregated is 1 to 3% of the gas flowing from the Gulf of Thailand. The party whose refusal ends the case becomes the party who signs the first contract.

What would have to be true. Three things, and they are separable.

First, the buyer and the scheme must accept a chain of custody that is not a traced molecule. This is the crux and it is the only real question in this what-if.

Second, PTT must buy at a domestic price. It will not pay a Singapore price for gas delivered in Thailand, and it does not have to. The memo's plant-gate cost is $10.33/MMBtu and the client's own model prices compressed biomethane domestically at $14.42, at liquefied petroleum gas parity.

Third, the attribute must carry the rest. On the memo's own numbers, the physical export earns $20 against $12.33 delivered, a margin of $7.67/MMBtu. This shape earns $14.42 against $10.33 plant-gate, a margin of $4.09, plus whatever the attribute fetches. The two are equal when the attribute is worth $3.58/MMBtu. Above that, this shape pays better than the export the memo recommends, and it pays it without the route, without the $1.15/MMBtu of unsourced estimate inside the delivered cost, and without the two conditions that currently clear by $1.11 and $1.00.

Does it look true here. The mechanism is well established elsewhere and it is the mechanism, not the outcome, that transfers. California's Low Carbon Fuel Standard credits renewable natural gas injected anywhere on the North American pipeline system against compressed gas withdrawn in California; the physical molecules need never enter the state, and the regulator certifies this as displacement (Norton Rose Fulbright; CARB Guidance 19-05). In Europe the same job is done by ERGaR's mass balance scheme, which exists precisely to monetize the green value of exported renewable gas without tracking the physical cross-border flow (ERGaR).

What was true in both cases and would have to be true here: one physically interconnected system, and a rule-maker willing to accept accounting in place of metering. The first condition holds. The memo establishes the Thailand-Malaysia-Singapore interconnection and records that PETRONAS Gas lists Thailand among its four entry points. The second is not established. We searched for whether Singapore's biomethane sandbox accepts book-and-claim or mass-balance delivery and found no statement either way, in the regulator's own release or elsewhere. The one transaction signed under the pilot is reported as the first physical bio-LNG supply deal in Asia (Bioenergy Insight), which is evidence that the pilot's first deal was physical and not evidence that a later one must be.

Verdict: worth a look, and it changes the sequence rather than adding to it. The memo's stage one already schedules a written approach to PTT and PETRONAS Gas. This is the same approach carrying a different question, and the different question has a plausible yes where the current one has a reasoned no. It costs nothing to ask both in one letter.

The shape. Do not buy biogas. Contract to treat the mill's effluent. Build, own and operate the digester and the upgrading train on the mill's site, take the palm oil mill effluent as an input, return treated water, and keep the gas. The mill pays a treatment fee, or at worst pays nothing and gives up a waste stream it currently flares or vents.

Why it changes the answer. Feedstock is $6.22 of the $12.33 delivered cost, half the stack and the largest single line in it. The feedstock condition clears by 0.66 THB/Nm3, one of the three narrow margins the memo flags. If the gas arrives at zero cost, that condition stops existing and the offtake price floor falls from $18.89 to somewhere near $12.67 on a straight subtraction, which is approximate because the floor is solved at 13% rather than added up. Even approximately, it turns the 60%-take case that returns 2.8% into something a business could survive. It also enlarges the resource: the 40% of Southern mills with no digester, which is what caps the pool at 42 mills well before the 70-mill count does, become addressable, because the digester is built by whoever wants the gas.

Does it look true here. The shape is real and it is running in this country, in this industry, at the mills this case is about. Asia Biogas, whose Krabi and Rayong positions the memo already identifies, describes developing biogas projects in the cassava starch and palm oil industries on a build own operate transfer basis, treating wastewater and monetizing the gas and the carbon (Asia Biogas). The industrial gas industry runs the identical shape at far larger scale: Linde builds, owns and operates air separation plants on customer sites so the customer avoids the capital and the operations (Linde), and Air Products has sold on-site gas this way since 1940 (Air Products). SIG is a joint venture with a US industrial gas major, so this operating model is already inside the project company.

What was true in those cases and has to be true here is that the customer is worse off doing it itself. In Thailand, for a palm oil mill, it is not. A CDM validation document for a palm oil mill at Lam Thap in Krabi Province states it plainly: "Being the least cost option for wastewater treatment, which also meets the legal discharge limit, the treatment of Palm Oil Mill Effluent (POME) in anaerobic open lagoons without biogas capture is a prevalent and standard industry practice in Thailand" (UNFCCC CDM). A mill facing no compliance pressure will not pay a treatment fee for something a pond already does legally and cheaply.

What survives is the weaker version, and it is still worth having. Not a fee, but free or near-free gas in exchange for a digester the mill neither funds nor operates. Against the memo's finding that a mill with a digester nets about 2.18 THB/Nm3 by burning gas in an engine, a mill without one nets nothing, so the offer is credible to the 40% without and worthless to the 60% with.

Verdict: worth a look, restricted to the mills that have no digester. It is not a replacement for the feedstock contract at mills that already capture. It is a second supply pool at a different price, addressed with a different instrument, and it carries capital the client's model prices at zero.

The shape. Invert the previous one. The business is a methane-avoidance carbon project developer operating under the Thailand-Singapore Article 6.2 implementation agreement. It builds digesters at mills that currently vent, sells authorized credits into Singapore's carbon tax, and treats the biomethane as a by-product to be sold wherever it fetches most, including domestically.

Why it changes the answer. Every other shape in this chapter competes for a fixed pool of gas. This one makes more of it. And its revenue is gated by none of the four conditions the memo rates hardest: not the pipeline, not the Singapore price, not take-or-pay, not cluster geometry. The memo names this instrument, notes that the funding gap for digesters and the funding mechanism for them sit in two different sections of its own text, and does not size it.

Does it look true here. The methodology is among the most established in carbon markets. Twelve methane-recovery projects were registered in Malaysia's palm sector under the Clean Development Mechanism by March 2009 alone, expected to deliver an average 612,097 tCO2e a year (MDPI), and three such projects have been reviewed in Southern Thailand specifically (ResearchGate). Asia Biogas funds its own expansion from carbon revenue. The memo establishes the bilateral agreement was signed on 19 August 2025 with an eligibility list published on 18 November 2025, and that Singapore's carbon tax is S$45/tCO2e against up to 5% of a payer's taxable emissions.

The fragility is additionality, and this case contains both halves of the argument against itself. For it: the Krabi validation quote above is exactly the counterfactual an additionality claim needs, since an uncovered lagoon is legal and standard. Against it: Thailand pays a biogas power feed-in tariff of THB 2.0724/kWh, and this very project proposes to pay 4.00 THB/Nm3 for the gas, which is evidence that a digester pays for itself without a credit. Those two arguments cannot both be made in the same submission. If the digester is bankable on gas revenue, the credit is not additional. If it is not bankable, the credit is additional and the gas costs whatever the credit does not cover.

Verdict: needs something we do not have. An eligibility and additionality determination against the methodologies actually published on the bilateral list, and a price at which authorized credits clear to a Singapore taxpayer. Neither can be settled here and both are cheap to ask inside stage one. It is worth saying that this is the only shape in this chapter that pays for the digesters the memo says nobody funds.

The shape. Treat carbon dioxide as the product and biomethane as the offtake that pays for the plant. Upgrading strips 174 tonnes a day of biogenic carbon dioxide at the 100 tonne-per-day scale, on the memo's own figure. SIG is an industrial gas company. The molecule the memo books at zero and then tests as a marginal add-on is, at merchant prices, a business of the same order as the gas.

The arithmetic, from the memo's own numbers. At 280 effective days a year, 174 tonnes a day is 48,720 tonnes. At the $238/t Japanese and $335/t Indian merchant prices the memo cites, that is $11.6M to $16.3M of gross revenue, against gas revenue at the same scale of about $25.9M. So the carbon dioxide stream is 45% to 63% the size of the gas revenue. That is an upper bound and should be read as one: those are import-market prices in two other countries, not a netback at a Songkhla fence, and capture, liquefaction and storage cost an estimated $16.87M of capital before a tonne moves. But the memo's test of this stream assumes a $50/t netback which it declares an estimate, against merchant prices five to seven times higher, and the entire question is what sits in that gap. The gap is distribution, and distribution is SIG's parent's business.

What would have to be true. Merchant carbon dioxide demand within economic haul of the hub, which means food and beverage, welding and greenhouse customers in Southern Thailand or a route to a port. A netback from SIG's parent well above $50/t. And food-grade certification of biogenic carbon dioxide from palm oil mill effluent, which is a genuine technical question given the sulfur load the memo measures at 943 mg/Nm3 in POME biogas.

Does it look true here. The technology is routine and commercial. Pentair sells a combined upgrading and recovery system specifically to add food-grade carbon dioxide as a second revenue stream (Pentair), Bright Biomethane sells the same recovery step as an extra source of revenue for the plant owner (BiogasWorld), and simultaneous production of biomethane and food-grade carbon dioxide has been documented as an industrial case study (Energy & Environmental Science). What was true in every one of those cases is a merchant carbon dioxide market within reach of the plant and a customer willing to certify a biogenic source. Europe has both. Southern Thailand's merchant carbon dioxide demand is not established anywhere in this case and we did not search for it.

Verdict: needs something we do not have, and what it needs is one quotation. A netback from SIG's parent, on a product the joint venture partner already sells for a living, is the cheapest test in this chapter and it is not on the memo's stage-one list. At a stream this size relative to the gas, its position in the sequence deserves revisiting.

The shape. Do not build a delivered business at all. Sell compressed biomethane free-on-board at the plant gate or at the Malaysian border to a party that already owns a Singapore contract and a multi-country portfolio. Three candidates are already named in the memo's own competitive section. BAC Renewable Energy signed the first physical deal under the sandbox with YTL PowerSeraya and states an intent to aggregate across ASEAN. Straits Bio-LNG at Muar buys compressed biomethane from established West Malaysian producers by dedicated long-tube truck and exports through its own jetty (Straits Bio-LNG). And Kestrel, in which Salerno is already a shareholder, holds a truck-based Malaysia-to-Singapore bio-LNG supply contract that the brief records and this case has never used.

Why it changes the answer. It deletes the route from the business. The gating condition on the Trans Thai-Malaysia line stops being Salerno's problem, and so does the $1.65/MMBtu route cost, the injection station, the certification charge, the $1.15 of unsourced estimate inside the delivered stack, and the cross-border biomethane precedent the memo searched for and could not find. It also softens the condition most likely to fail: an aggregator holding a portfolio across three countries can carry swing volume that a single-source supplier at 60% minimum take cannot.

What it costs. Price, and possibly a lot of it. The memo's plant-gate cost is $10.33/MMBtu against $12.33 delivered, so the route is worth $2.00 in cost and $7.67 in margin. Selling at the fence surrenders the delivered margin for whatever an aggregator pays. Salerno already reaches only $5.79M of a $10M target on the delivered business, so this shape has to be argued on capital avoided rather than margin earned, and it is not obvious that it survives that argument.

Does it look true here. The shape is operating, with somebody else in the producer's seat: Straits buys compressed biomethane at the fence from West Malaysian producers today, so a fence price exists in this region. It is not published, and no search here was spent finding it, so whether it clears $10.33 plus a return is unknown. That is the one number this what-if turns on.

Verdict: worth a look. The memo already half-describes it as an escape hatch and declines to price it. It also makes the free phone call the memo flags as unmade a good deal more interesting: if Kestrel's existing Singapore channel will take Thai molecules, this shape is the only one here with a proven route to a Singapore buyer already inside the shareholder group.

The shape. Salerno's stated role is deal-maker, curator and investor. The plan in front of it spends several years and $32.37M of first-phase capital becoming an operator. The alternative is to take a minority position in a platform that already operates palm oil mill effluent biomethane at scale and already sells to the buyers this case is trying to reach, contributing the Thai feedstock relationships and the Singapore access as the thing Salerno brings to it.

Does it look true here. The precedent is close enough to be uncomfortable. In November 2025 Mitsubishi Corporation acquired a minority stake in KIS Group's Indonesian operations, its first entry into the biogas market, with a stated ambition of $1 billion across renewable gas in Southeast Asia and India by 2030 (Bioenergy Insight; Business Standard). KIS operates palm oil mill effluent bio-compressed natural gas plants in Sumatra and Kalimantan, has completed more than 80 projects in eleven countries, holds long-term supply agreements with Unilever and Shell, and the partnership names Thailand among its expansion markets and a bio-LNG export hub in Indonesia by 2026 (Feature Asia).

What was true in that case and would have to be true here: a trading house buying into an operator and contributing a network rather than engineering, which is Salerno's exact self-description. What differs: Mitsubishi bought a platform that was already running. Salerno would need a target, and every candidate this case has identified so far has been identified as a competitor rather than approached as a counterparty. Asia Biogas is Thai and already in Krabi. Cenergi, Straits and KIS itself are regional. None has been contacted.

The second reading, which is less comfortable. Mitsubishi is a company of Salerno's own type that faced this question in this region in the same year, and it did not build. And KIS naming Thailand as an expansion target means the Southern Thai position this case is studying may be contested by a better-capitalized operator inside the brief's own end-2026 decision window.

Against it. Buying a minority stake does not obviously reach $10M a year either, and it gives up the strategic aim behind the whole exercise, which is a durable owned position in a Thai value chain. It also cannot be evaluated at all without a target and a price, and this chapter has neither.

Verdict: worth a look, with the caveat that it is the only shape here Salerno's stated capabilities fit without a stretch, and the only one whose precedent is a company of Salerno's own kind making the same decision in the same market in the same year.

Four more were considered and set aside, with the reason.

Kestrel builds and never owns. A pure engineering and operations business selling upgrading trains to mills that keep their own molecules. Real, low-risk, and too small: it is a contracting margin on a few tens of millions of capital, and the memo already shows that engineering is one-off and cannot close a recurring $4.21M gap.

The mill cooperative. Mills contribute gas rights for equity. Already in the memo, and correctly placed there as a structuring choice rather than a shape.

The pure regional trading book. Buy Thai, Malaysian and Indonesian molecules and sell one delivered offer into Singapore. The memo identifies this as the winning position and declines to evaluate it because it breaches both the feedstock and the corridor constraints. Nothing here changes that, but what-if 5 above is its junior version and does not breach either.

Selling the attribute to Europe or Japan rather than Singapore. The memo names it, prices nothing, and is right that scope forbids it. What-if 1 is the version of the same idea that stays in scope, because the attribute goes to Singapore and only the molecule stays home.

Four of these six remove a condition rather than improve one. The first is the one to test soonest, because it costs a paragraph in a letter the memo already plans to send, and because the memo's own reasoning about why PTT and PETRONAS will refuse is an argument about a shape rather than about those two companies. The fourth is the cheapest test of any kind in this case: one netback quotation from a partner who sells that product for a living, on a stream the memo has valued at an assumed number five to seven times below the market it cites.