SINGAPORE, 8 AUG 2026 — A Member of Parliament asked the Finance Ministry three things about Temasek and eFishery: whether it invested, how much, and whether the government would hold Temasek's management accountable.
The reply addressed accountability, but not the amount invested.
What was asked and what came back
The question, from Mr Fadli Fawzi on 4 August, was in three parts. Whether the government is aware that Temasek Holdings had invested in eFishery. If so, whether it has information on the amount invested and on measures Temasek has taken to prevent similar investments. And whether it intends to hold Temasek's management accountable for the approval and oversight of that investment.
The answer, from Senior Minister of State for Finance Mr Jeffrey Siow, opened by noting that there are public reports regarding the investment and that Temasek has publicly commented on the matter. It then set out the constitutional position: the government does not scrutinise or direct Temasek's individual investment decisions, which are commercial decisions made independently.
On accountability it was specific. The government's role is to ensure Temasek has a sound governance framework; it holds the Temasek Board accountable for the company's overall long-term performance, including governance and risk management; and it assesses performance on an overall portfolio basis over the long term rather than on the outcome of any individual investment.
The sum invested does not appear. Neither do the measures the question asked about.
What eFishery was
The absence matters because of what the investment was in.
eFishery was an Indonesian agritech company that reached unicorn status, backed by investors including SoftBank Group and Temasek Holdings. It sold automated fish feeders to aquaculture farmers and was among Indonesia's most celebrated startups.
| Reported for Jan–Sep 2024 | Found by forensic audit |
|---|---|
| US$752 million revenue | US$157 million |
| US$16 million profit | US$35.4 million net loss |
⚠️ These figures come from reporting of the investigation and the subsequent court case, not from a document RECATOOLS has read. A separate whistleblower estimate reported by Bloomberg and cited by CNBC put the revenue inflation at almost US$600 million over the nine months to September 2024. We report both because they are different measures from different stages of the same inquiry.
The founder and former chief executive, Gibran Huzaifah, was sentenced by the Bandung District Court in April 2026 to nine years for embezzlement and money laundering. That is the figure most coverage still carries, and it is no longer current: on appeal, the Bandung High Court reduced the sentence to six years, in a ruling reported on 20 July 2026. The same court reduced the sentence of a former vice-president in a separate ruling.
The doctrine in the answer is not new
The reply might sound evasive, but it is a restatement of a long-standing arrangement.
Temasek is a commercial investment company owned by the government, and the separation between owner and manager is the design. The government appoints the board and holds it to account; the board runs the company. A minister who began commenting on individual holdings would be dismantling that boundary, not enforcing it.
The reply also makes an argument about portfolio construction. Startups are described as inherently higher risk and a relatively small part of the overall portfolio, and performance is assessed across the portfolio and over the long term. On that framing, a single failed investment is not evidence of a governance failure, and demanding accountability for one is a category error.
Even so, the amount invested is a figure the public has a right to know.
Why the missing number is the story
The distinction the reply relies on is between individual decisions, which are Temasek's, and the governance framework, which is the government's. Disclosure of a sum invested does not obviously fall on the commercial side of that line.
Knowing the loss is not about second-guessing the investment. It is about the governance framework the government claims to oversee — specifically, whether the exposure to a single startup was within the board's risk-management limits.
Temasek has commented publicly on the matter, as the reply notes. The reply does not clarify whether the government has the figure and is withholding it, or simply does not have it.
What this says about regional exposure
Singapore's institutional investors are among the largest sources of growth capital in Southeast Asia, and the region's startup ecosystem has spent several years absorbing the consequences of valuations built on numbers that were not audited to public-market standards. eFishery is the most severe case so far, with a founder imprisoned and a forensic audit describing revenue overstated by a multiple.
An investor judging performance by the whole portfolio over the long term will naturally absorb some failures without disclosing them. It is a defensible model, but it means the public cost of regional venture failures comes from court documents and reporting, not from the investor itself.
What the question was really testing
Parliamentary questions about Temasek and GIC follow a well-worn pattern; they are interesting for which part of the boundary they probe.
A question about whether to invest in something is outside the boundary, and gets the commercial-independence answer. A question about the governance framework is inside it, and gets a substantive answer. This question did both: parts (a) and (b) asked for facts, part (c) asked about accountability.
The reply engaged fully with (c) and set out the doctrine at length. The factual questions went unanswered — the reverse of what the government's framing would predict. Disclosure of a figure requires no judgement about a commercial decision at all.
Nothing in the separation between owner and manager prevents a government from saying what it knows.
The comparison that is not available
There is a specific reason the missing figure is hard to substitute for.
Temasek reports a net portfolio value and a total shareholder return, both at portfolio level. It does not publish position sizes for individual holdings, and there is no filing regime that would force it to for a private Indonesian company. So the exposure cannot be reconstructed from the outside.
The only routes to the number run through the failed company, not the investor — an Indonesian court process, an administrator's report, or investor litigation, none of them certain.
The precedent this sets for the next one
Regional venture failures are not going to stop, and the handling of this question establishes how the next will go.
An investor assessed on portfolio outcomes over long horizons has a coherent reason never to itemise a loss, and a government that respects commercial independence has a coherent reason not to compel it. Both positions are defensible in isolation. Together they mean that no single failure will ever produce a figure.
That is a choice about disclosure rather than about governance. The alternative — publishing an amount once an investment has failed publicly and been through a criminal court — would not require the government to second-guess anything.
What to watch
Two things.
The first is whether the amount ever emerges. It could come from Temasek's own disclosures, from Indonesian court proceedings that enumerate creditors and shareholders, or from litigation among investors. None of those routes is the Singapore Parliament.
The second is whether the framework question is asked directly. The reply answers a question about a decision. A question about whether the risk-management framework allowed for the right concentration limits in early-stage venture would be different, and would fall squarely within the government's stated remit.
