SMR Financing in Southeast Asia: Singapore Won't Build the Reactors, It Will Decide Who Can Afford To

If you are building a hard technology and waiting for Southeast Asia to become a market, you are probably watching the wrong variable. You are watching whether the technology works. The people who decide whether it gets deployed are watching something else entirely.

We convened four of them — a reactor developer, an energy market analyst, an infrastructure investor, and a deep-tech venture investor — for a session on Singapore's role in SMR financing in Southeast Asia. What emerged was less a conversation about nuclear than a conversation about who absorbs risk, and in what order. Founders in any capital-intensive category should read it that way.

The bottleneck is not the technology, and the people financing it will tell you so

We opened by asking what is actually broken in Southeast Asia's utility and grid structure that makes first-of-a-kind projects so hard to finance. The answer from Lathika Chandra Mouli of Aurora Energy Research was unambiguous:

"It's not a technology issue and the system as well is fundamentally sound, but it's really about how risk is allocated in the early stages and how that's misaligned with traditional financing structures and how power assets are generally commissioned. Commercial banks are structured to be conservative — well-suited for low-risk, long-duration assets with stable cash flow."

She identified two structural problems. The first is the region's prevailing single-buyer model, which cuts both ways: it "mirrors how most nuclear fleets were historically built in France and Korea," but where state-led utilities carry visible public financial strain, "banks are a bit more risk-averse when it comes to funding projects that have offtake only with these state utilities."

The second is a regulatory vacuum. Every ASEAN market maintains different licences, different requirements, different safety regimes. Absent harmonisation, "it will constantly be a piecewise approach." Her conclusion reframes the whole problem: the core financing challenge is not market design but "how to bridge that regulatory risk and that construction risk into an investment landscape that's generally more risk-averse and traditional."

Emily Bolon, commercialising Blykalla's reactors, confirmed it from the developer's side of the table — and the specific thing she said should be pinned above the desk of every deep-tech founder eyeing this region:

"If you are talking about a small country that has an obscure licensing pathway that's very complex and unproven, it's more challenging than the technology."

Demand, she noted, is not the constraint. "We get calls every week from ministers of economy." The first question her team asks in response is "how the market potential squares with the ecosystem setup from a regulatory and financing perspective." Enthusiasm from a ministry is not a market. It is a signal to go and check whether the licensing pathway exists.

You are being priced against the last generation's failures

Bolon's second point was about perception, and it generalises well beyond nuclear.

"Many are traumatized by the giga-projects of the past that come in a decade delayed and billions over budget. People still have that old mentality."

Her counter-framing is that a modular reactor is not a mega-project at all. A better analogy is developing a new car model: "you might be some months late to market, and maybe the first units cost a little more before the learning curves kick in, but then you have industrial mass production. The project risk fades in comparison to the economics once you're in the NOAK side of things."

The lesson for founders is that a category's historical failure mode gets priced into your cost of capital whether or not it applies to your architecture. Part of the job is actively dismantling the analogy investors are reaching for. Bolon's response is to make the difference concrete rather than rhetorical: Blykalla's reactors are 55 megawatts, sized to fit a flatbed truck on standard roads, built as eight functional blocks with predefined interfaces and pricing, serially produced independently of site configuration, and assembled on site — which is what permits "fixed-price contracts with guaranteed timelines." Their target is 24 months from construction to commissioning. As she put it: "We're not talking about seven years."

If you cannot state a comparable structural reason why your economics differ from your category's worst historical case, you will be underwritten as that case.

The sequencing problem at the heart of SMR financing in Southeast Asia

The mechanics of how public capital unlocks private capital were the most transferable part of the discussion. Andrew Wong of TRIREC described Singapore's Green Investment Partnership and FAST-P programme as "a fantastic first move," and explained why the structure works:

"You have junior type investors that are willing to take a first loss. Usually, that's more the government side or DFIs who have a strategic interest in seeing this built up. Afterwards, you could have the senior tranche for the more financially interested investors."

He added two modifications nuclear would require: longer loan tenors, and some form of backstop or subsidy against construction overruns, which "more often than not do happen."

Three conditions for SMR financing in Southeast Asia

Suchitra Narayan of SGInnovate set out three conditions for blended finance to work in this asset class. First, public-private co-investment — anchoring projects on development banks or sovereign funds "sends a very strong signal on policy commitment and reduces perceived political and regulatory risk." Second, genuinely patient capital: "We're not talking a fund life of 7 to 10 years; we're talking 20, 30, or 40 years to match the nuclear asset development and operational timelines." Third, ecosystem construction — regulatory capacity, supply chain, workforce.

Bolon supplied the working proof from Sweden, where the government offers 80% debt financing at sovereign rates for eligible projects plus a 40-year contract for difference, with returns addressed across investor classes. That structure, she said, is what enabled Blykalla to announce a first park of six reactors in central Sweden, in an area dense with data centres. "With that financial backbone, we can execute."

For a founder, the actionable version is this: the jurisdiction that offers you a de-risking structure is more valuable than the jurisdiction that offers you a large addressable market. Sweden is a small market. It is where the reactors are getting built.

Split the company in two, because investors are not one audience

The most structurally useful idea in the session was the BuildCo–OpCo separation — splitting the risky construction phase from the stable operating phase. Wong explained why it works by analogy to renewables:

"Investors in the building phase are accepting higher risk with construction and timelines. Their expectation on returns is higher — they are looking for equity-type upside. If you separate them, you have one set of investors primed for that risk. For the OpCo side, it's really about cash flow yield. That type of investor is looking for lower but more stable returns."

He was candid that this is not yet how the market operates: today's deals are "all-in-one," with hyperscalers investing directly into startups while signing future PPAs. The separation makes sense "eventually, when things mature."

Narayan argued Singapore is unusually well-positioned to sit on both sides — regulatory expertise and risk assessment capability for the operating entity, deep capital pools for the construction entity. That dual capability, she said, "means both venture and catalytic investors can absorb high development risk while infrastructure funds step in once projects are operational and revenue-generating."

Any founder whose product has a build phase and an operate phase should be asking whether they are trying to raise a single round from investors who want two different risk profiles.

The demand side has a creditworthiness problem, and data centres are the answer to it

Chandra Mouli identified the question underneath ASEAN's bankability gap: "who is a creditworthy long-term offtaker?" Developers cannot find counterparties willing to sign 20- or 30-year contracts, and without those contracts, banks decline.

Data centres resolve this, because they have "credibility and bankability" and because they "last 20 to 40 years" — matching the asset life of the generation they would consume. That is the structural reason the nuclear story and the AI story belong in the same sentence rather than merely sharing a conference agenda.

She also described the shift to what she calls a Bring Your Own Power era, driven by a straightforward speed differential: "We're seeing multi-megawatt data centers coming online in 16 to 20 months — way faster than grid reinforcement." Hyperscalers are already acting on it, acquiring companies and signing PPAs directly. But nuclear, she cautioned, remains categorically different: "It involves policy concerns like waste and safety. It will always need government backing."

Three things need to change for private ownership models to scale regionally: direct PPAs must be permitted, since in many ASEAN markets vertical integration forces data centres to transact through state utilities; licensing must harmonise across markets; and governments must provide revenue-stability mechanisms — she pointed to Sweden's package and to the UK's strike price floor, which guarantees minimum revenue if a PPA defaults.

Underwrite milestones, not revenue

Asked how investors should approach a 10-to-15-year asset with a 5-to-7-year return expectation, Wong rejected the framing that this is unprecedented:

"You need to shift your mindset toward milestone-based outcomes. In pharmaceuticals or biotech, companies aren't earning revenue when investors first come in. They go through clinical study milestones and regulatory approvals that uplift valuation. For VC investors with 10-year fund lives, it's not too far off because you see meaningful uplift by the time the fund winds down."

The milestones he named for his own portfolio are concrete: validation of reactor designs, signed agreements with counterparties such as the Tennessee Valley Authority, or reaching a Final Investment Decision.

This is the single most useful reframe in the session for founders with long development cycles. Your job is not to shorten your timeline to fit a fund's. It is to identify the intermediate events that create defensible valuation uplift, and then to hit them on schedule.

The honest gaps

Two constraints were named plainly, and we would rather publish them than leave them out.

On workforce, Narayan was blunt: "we are nowhere near ready to operate nuclear plants in ASEAN. We need more nuclear physicists and engineers. It's work in progress, and while it's factored into financing, the components are currently disparate." On supply chain, she described a chicken-and-egg problem — invest now, or wait for proof of concept — which investors currently approach with caution.

Chandra Mouli was similarly measured on competitiveness. Aurora's analysis, she said, shows nuclear can provide competitive baseload power and stabilise price swings through commodity cycles. But the conditional matters: "If managed properly, nuclear is competitive; if not, it can increase system costs. It depends on market design and regulatory support."

Where the bet actually sits

Asked whether Singapore's long-term position would be pioneer, coordinator, or financier, Chandra Mouli was careful to mark her answer as personal rather than institutional:

"These are my own views and not necessarily Singapore's official priorities. But if I were betting, it would be on Singapore being that 'financial architect' and developing enabling infrastructure — grid modernization, interconnectors, nuclear safety studies, and regulatory capacity building. Singapore has an edge in financial and regulatory expertise."

The supporting evidence she cited is institutional: a nuclear energy office already established within the Energy Market Authority, and a mature green bond programme targeting $35 billion by 2030 — capital that could back regional projects in Johor or Batam, where multi-gigawatt data centre announcements are concentrating along the SIJORI corridor.

Bolon closed with the most commercially direct advice of the session, and it is worth ending on because it inverts the usual assumption about who is scarce:

"Systemic investors should seek priority in the order book. There won't be enough reactors to fulfill global demand. There's going to be a race to be at the top of the order book."

That is a supply-constrained market, not a demand-constrained one. The capital that moves earliest does not merely get better terms; it gets delivery slots. For founders, the mirror image applies: the jurisdictions and partners that structure risk properly will get your first units, and everyone else will wait.

Southeast Asia will not be first to deploy advanced nuclear. It does not need to be. But the region that solves risk allocation before its neighbours will be the one where the reactors — and the data centres they power — actually land.

---

This article draws on Earth VC's webinar "Singapore as the Financial Architect of Advanced Baseload for ASEAN," part of the Nuclear-AI Nexus series following our March 2026 report, Nuclear Energy Fuels AI Boom in Southeast Asia Data Centers.

Download the full report.