Delfi Limited (P34.SI) - Deep Dive Research Report

Compiled 2026-05-28. Ticker P34.SI, Singapore Exchange Mainboard. Reporting cadence: half-yearly (December year-end), with supplementary trading updates.

1. What the company does

Delfi makes chocolate, mostly for Indonesians. SilverQueen is the bar most Indonesian households know. Ceres is the chocolate sprinkle the same households put on bread for breakfast. Delfi Top is the moulded chocolate eaten at Lebaran. The company is the dominant branded chocolate confectioner in Indonesia and runs a smaller branded business in the Philippines, plus a third-party distribution business across Singapore, Malaysia and Indonesia for outside brands like Toblerone, Pringles and Kellogg's.

The deeper story is more interesting. The business now called Delfi was, until 2013, one of the largest cocoa bean processors in the world. Founded in 1984 by John Chuang and his brothers as Petra Foods, it built a global cocoa ingredients operation supplying Cadbury, Mars, Meiji and Nestlé from plants in Indonesia, Malaysia, Thailand, the Philippines, Brazil and Mexico. In 2012 the family sold the cocoa ingredients division to Barry Callebaut for around US$950 million, completing in 2013, and retained the branded consumer business. In 2016 they renamed the listed entity Delfi Limited to make the strategic split clear: this is now a downstream chocolate brand owner, not a cocoa grinder, and the distinction matters because the two businesses behave very differently when cocoa prices move.

The core value proposition to consumers is simple. Indonesia has roughly 280 million people, an underdeveloped cold chain, a fragmented retail landscape with hundreds of thousands of warungs (kiosks), and a population whose per-capita chocolate consumption is a small fraction of what is seen in Singapore or Japan. Reaching that consumer with a recognised, affordable, shelf-stable chocolate product is hard. Delfi has spent six decades doing exactly that. The Ceres factory in Bandung dates back to 1950. SilverQueen was launched in the 1950s. The brand recall is intergenerational - four generations of Indonesian families have grown up with these products - and the distribution that supports it reaches several hundred thousand points-of-sale, an installed-base that is genuinely hard to replicate in a country where modern trade is still the minority channel outside the top cities.

A concrete example of what the business does, end to end: a 65g SilverQueen Cashew bar leaves the Bekasi or Bandung factory in Java, moves through one of dozens of company stock points operated by PT Nirwana Lestari (Delfi's Indonesian distribution arm), reaches a warung in a small city like Cirebon or Banjarmasin via a sub-distributor truck route that runs weekly, sits on a shelf at room temperature (the chocolate is formulated to survive Indonesia's heat without a refrigerated truck), and sells to a consumer at a price point of a few thousand rupiah per pack. The whole process - formulation for tropical heat tolerance, micro-distribution across an archipelago, brand investment that has compounded for half a century - is the moat. The product itself is commodity chocolate; the system that gets it onto 400,000 shelves in a country with no proper cold chain is not.

2. Business segments

Delfi reports along two geographic axes and two brand axes. The geographies are Indonesia and Regional Markets (Philippines, Malaysia, Singapore). The brand cut is Own Brands versus Agency Brands. Together these create the four lenses management uses to talk about the business.

Indonesia (own brands + agency)

Indonesia is the company. It accounts for roughly 60% of group revenue in a typical year. The own-brand portfolio here - SilverQueen, Ceres, Delfi, Selamat, Cha Cha, Top - is what makes the segment unique. Most of these brands have been on Indonesian shelves for fifty years or more, and the company has spent decades building a distribution arm, PT Nirwana Lestari, that reaches the modern trade (supermarkets, minimarkets like Alfamart and Indomaret) and the traditional trade (warungs) simultaneously. The competitive position here is strongest: Delfi is consistently the largest branded chocolate player in Indonesia, with Euromonitor data placing share at around 34% in 2024, well ahead of the next-largest player, Mayora Indah. Indonesia is also where the agency distribution business has the most scale, carrying third-party brands like Toblerone and Guylian into local retail. This is the cash engine and the strategic priority - management has been investing capex into automation and capacity here for the better part of a decade. The two factories at Bandung and Bekasi run by PT Perusahaan Industri Ceres are the workhorse manufacturing footprint.

Regional Markets (Philippines, Malaysia, Singapore)

Regional Markets is the growth bet, structured as three distinct sub-businesses. Philippines is the most substantial of the three: a manufacturing presence in Marikina City that came in via the 2006 acquisition of Nestlé Philippines' confectionery business, bringing local brands Goya and Knick Knacks into the portfolio. Malaysia is largely a distribution business carrying own-brand SilverQueen and the agency portfolio. Singapore is small in chocolate terms but is where the corporate centre, listing and treasury sit. The competitive position in Regional Markets is weaker - in the Philippines, Delfi competes with locals like Universal Robina (Cloud 9) and the multinationals (Cadbury, Mars); in Malaysia and Singapore it competes against full-scale multinational shelf presence. Regional Markets has been the better-growing segment in recent halves: management called out 7.2% revenue growth in Regional Markets in 1H 2025 against a decline in Indonesia (1H 2025 results, August 2025).

Own Brands vs Agency Brands (cuts across both geographies)

The brand cut matters because the unit economics are very different. Own brands - SilverQueen, Ceres, Delfi, Van Houten (licensed for SE Asia distribution since 2018), Goya, Knick Knacks, Cha Cha - carry significantly higher margins because Delfi controls the formula, the marketing and the price ladder. Agency brands - Toblerone, Guylian, Loacker, Kellogg's, Pringles, Fisherman's Friend - are pass-through distribution: lower margin, lower capital intensity, but useful for filling out the distribution truck and earning route economics. Management has been steering the mix toward own brands deliberately: own-brand sales grew 4.9% in FY2025 while agency-brand sales fell 7.4%, partly because management let an unspecified agency contract terminate. Excluding that agency termination, group sales would have grown roughly 6% in FY2025 (FY2025 press release, February 24, 2026).

3. Products and business detail

The product catalogue is broader than the headline brands suggest. The own-brand portfolio in Indonesia includes SilverQueen (milk chocolate bars, cashew variants, the everyday large-format bar), Ceres (chocolate sprinkles - the iconic meses for bread), Delfi (a European-style moulded chocolate sub-brand, including the Delfi Top tin gifted at Lebaran), Selamat (wafer-based chocolates), Cha Cha (chocolate buttons), and a long tail of seasonal and SKU-level variants. The licensed portfolio adds Van Houten across SE Asia (the rights were acquired in 2018 for around US$13 million). In the Philippines, the Goya and Knick Knacks brands, picked up from Nestlé Philippines in 2006, are the local equivalents of SilverQueen. Across the group, the company has historically run somewhere between 200 and 500 SKUs at any given time; management has been rationalising downward as modern retail consolidates and shelf-space pressure forces a tighter assortment.

Manufacturing sits in two clusters. In Indonesia, PT Perusahaan Industri Ceres operates two plants in the Greater Jakarta-Bandung corridor: Bandung (the original 1950 site) and Bekasi (the newer, more automated facility). Together these have a combined annual production capacity in the order of 100,000 tonnes across moulded chocolate, enrobed wafers, panned products and biscuits. The capacity has been progressively modernised since 2014 through an automation programme that has cost north of US$100 million cumulatively, including SAP ERP rollout and line automation. In the Philippines, Delfi Foods Inc. operates a facility in Marikina City (Metro Manila) inherited from Nestlé in 2006. The Indonesian formulation work - particularly heat tolerance and tropical-stable formulations - is the most distinctive piece of process knowledge in the group, because most multinational competitors design product for temperate markets and have to retro-engineer it for the tropics.

Geographies are concentrated. Indonesia is by far the largest market. Philippines is the second-largest. Malaysia and Singapore round out the home region. Outside SE Asia the company has essentially no consumer presence today - a deliberate choice after the 2013 Barry Callebaut divestment, which was the moment the global cocoa-ingredients geography (Brazil, Mexico, Thailand) left the perimeter.

Key milestones, with dates: Ceres Bandung factory opened 1950; PT General Food Industries founded 1969; Petra Foods incorporated by John Chuang 1984; began third-party brand distribution 1987; international cocoa ingredients expansion 1988; Meiji Seika confectionery joint venture in Indonesia 2001; Malaysia stake (Ceres Sime) 2002; Nestlé Mexico cocoa facility 2003; cocoa grind capacity ~200,000 mt by 2003; Petra Foods listed on SGX 2004; acquired Nestlé Philippines confectionery (Goya, Knick Knacks) 2006; cocoa ingredients division sold to Barry Callebaut 2012-2013; rebranded to Delfi Limited 2016; Van Houten Asian rights licensed 2018; Yuraku Confectionery joint venture in Indonesia 2019; Indonesia automation programme ongoing through the 2020s.

4. Customers

The buying relationship in this business has two layers - distributors and end consumers - and Delfi sells to both.

The proximate customer is the trade. In Indonesia, the company sells primarily through its own distribution arm, PT Nirwana Lestari, which serves modern trade chains (Alfamart, Indomaret, Hypermart, hypermarkets) and traditional trade (independent warungs, kiosks, wet markets) via dozens of stock points and a sub-distributor network. The buying decision at modern-trade chains is made by category buyers who care about turn rate, gross margin contribution, on-shelf availability, and promotional support. The buying decision at warung level is essentially the route salesman's relationship with the shopkeeper - what the truck dropped last week, what is moving, what is in stock. The criteria here are not glamorous: reliable supply, predictable margins, brands consumers ask for by name. Delfi wins on all three. In the Philippines, the analogous structure runs through Delfi Marketing Inc.

The end consumer matters more for the long-term franchise. SilverQueen is the chocolate Indonesian teenagers grow up with; Ceres meses is what children get on their breakfast bread; Delfi Top is the assortment tin given as a Lebaran or Christmas gift. These are habit purchases sitting at price points (a few thousand to tens of thousands of rupiah) that survive most macro cycles. The buying decision is repeat consumption built on decades of brand exposure.

Switching costs are asymmetric. For a modern-trade buyer, switching out SilverQueen would mean explaining to consumers why the shelf they expect is empty - a non-starter for a category buyer who is judged on category sales. For a warung, switching means losing access to the broader Nirwana distribution route, which includes the agency portfolio. For consumers, taste preference is locked in early and Indonesian chocolate palates lean toward sweeter, milkier formulations that Delfi has spent fifty years optimising for. None of this is contractual lock-in but the cumulative friction is real.

Concentration on the customer side is low - hundreds of thousands of points-of-sale, no single retail chain that approaches double-digit revenue share. On the agency side, there is concentration risk: a single agency contract termination cost the group around 7.4% of agency-brand revenue in FY2025. Management called it out plainly: an unspecified agency account terminated and the comparison will distort the next 12 months of growth optics.

Contract structures are largely informal trade terms. There are no long-term consumer supply agreements analogous to industrial or semis customer relationships; revenue is essentially recurring repeat purchase via the trade channel, with seasonality concentrated around Lebaran (Ramadan/Eid), Christmas and Chinese New Year.

5. Competitive landscape

The Indonesian branded chocolate market is structurally a two-horse race with a long tail.

Mayora Indah (IDX: MYOR) is the most direct competitor. Mayora is a much larger diversified Indonesian consumer goods company (coffee, biscuits, candies, beverages) where chocolate confectionery is one of many product lines. In chocolate-adjacent confectionery, Mayora's Beng Beng wafer bar and Choki Choki chocolate paste are direct shelf competitors to Delfi products. Per Euromonitor data referenced in independent coverage, Mayora holds around 35% Indonesian chocolate share against Delfi's roughly 34% (down from a high of ~45% several years earlier). Where Delfi wins: stronger brand equity in moulded/bar chocolate, deeper distribution reach in non-urban Indonesia, a more focused chocolate-only operation. Where Delfi loses: Mayora has much more financial firepower, runs aggressive promotional pricing, and benefits from a multi-category sales force that can amortise route costs across more SKUs.

The multinationals - Mondelez (Cadbury, Toblerone in non-Delfi-distributed markets), Mars (Snickers, Dove), Nestlé (KitKat), Ferrero, Hershey, Lindt - are present in Indonesia but mostly in modern-trade urban formats. Their products typically sit at higher price points and target a more aspirational consumer. They compete with Delfi at the top of the Indonesian price ladder (premium gifts, imports) but cede most of the mid-price everyday consumption to Delfi and Mayora. The cost structure works against multinationals in Indonesia: they ship product in or produce at smaller local scale, neither of which lets them match Delfi's unit economics in the traditional trade.

A second tier of Indonesian local players - Orang Tua Group, Garudafood, Nabati - competes selectively in chocolate-adjacent products (wafers, biscuits, chocolate-coated snacks) without directly challenging Delfi's moulded-chocolate stronghold.

Barriers to entry are partly capital, partly time. The factory capex to build a chocolate plant of meaningful Indonesian scale is in the low-to-mid hundreds of millions USD. The harder barrier is the distribution build - reaching 400,000+ points-of-sale across an archipelago of 17,000 islands without modern cold chain is a multi-decade project. The third barrier is brand: SilverQueen has been on shelves since the 1950s, and building a competing brand to that recall is far more expensive than the chocolate factory.

In the Regional Markets, the competitive position is more challenged. In the Philippines, Universal Robina (Cloud 9, Tootsy) is the local heavyweight and the multinationals have full-scale shelf presence. In Malaysia and Singapore, modern trade is much more developed and the multinationals dominate; Delfi is a niche player at the value end.

Structural shifts to watch: modern-trade consolidation in Indonesia (Alfamart/Indomaret format expansion) is gradually compressing the traditional-trade channel where Delfi has its biggest relative advantage, and is forcing SKU rationalisation (management cut from roughly 500 SKUs to under 300 in recent years per independent coverage). Mayora has been investing aggressively in marketing, which has forced Delfi to step up promotional spending in 2024-2025 to defend share.

6. Industry

The Indonesian chocolate confectionery market is the industry that matters for Delfi. It is structurally under-penetrated. Per-capita chocolate consumption in Indonesia sits at roughly 20-25% of Singapore or Japan levels (independent coverage citing Euromonitor), against a population over 280 million, growing roughly 1% annually, urbanising, and with a slowly rising middle class. The demand drivers are essentially demographic and income-led: more urban consumers, more cash income, more modern-trade exposure means more chocolate consumption per capita over time.

Industry size is not enormous in global terms but is meaningful regionally. Euromonitor-style estimates put the Indonesian chocolate confectionery market in the low single-digit billions of USD annually, growing in the mid-single digits in volume terms in normal years and faster in value terms when cocoa pricing translates into shelf price rises (with a lag).

Supply chain context: Indonesia is itself a significant cocoa bean producer (third largest globally, after Ivory Coast and Ghana), which would in principle be a structural advantage. In practice, most Indonesian cocoa is exported as beans and the high-grade cocoa butter and liquor used for chocolate manufacturing is sourced globally. Delfi sources processed cocoa ingredients (no longer doing its own grinding since 2013), which means it sits at the mercy of the international cocoa price cycle rather than the local farmgate cycle. This was the painful realisation of 2024-2025: cocoa prices spiked to record levels (peaking in early 2025 at multiples of historical levels) on West African crop disease and weather; Delfi hedged forward 6-12 months on contracts and 18 months on inventory, so prices fell ~60-75% from the peak through 2026 but the benefit will reach the P&L only gradually as hedged contracts roll off. Independent coverage from UOB and RHB (cited in nextinsight, March 2026) expects margin benefit to be more visible in FY27 than FY26.

Import dynamics are not a meaningful threat to Delfi's everyday-consumption business; finished imported chocolate is largely a premium-segment story. Regulation is light-touch at the product level (halal certification is the most important compliance gate for Indonesia, and Delfi has long had it) but the macro overlay matters: Indonesian government fuel and electricity subsidies, IDR-USD currency policy, and consumer-facing fiscal stimulus all flow into demand. Management cited Indonesian government stimulus as having supported own-brand sales in FY2024.

Cyclicality is mild relative to industrials. Chocolate is a consumer staple with strong seasonality (Lebaran is the single biggest gifting season for Delfi's Top assortment tins, Christmas and Chinese New Year follow). The 2020 COVID year showed the staple property: revenue was relatively resilient. The cycle Delfi actually faces is a commodity cycle (cocoa) layered on top of a currency cycle (IDR/USD), more than a consumer-spending cycle.

Tailwinds at the industry level: chocolate penetration in Indonesia has decades of runway; modern-trade format expansion is increasing chocolate's per-capita reach; cocoa prices are off the peak and trending toward normalised levels. Headwinds: cocoa price volatility remains structurally elevated due to West African supply uncertainty; IDR weakness compounds USD-denominated raw material costs; competitive intensity from Mayora has been a multi-year pressure on promotional spending; modern-trade SKU rationalisation continues to thin the assortment tail.

7. Growth triggers

Drawn from the four reporting periods (1H 2024 release, FY 2024 release Feb 26 2025, 1H 2025 release Aug 2025, FY 2025 release Feb 24 2026) plus the April 23 2026 AGM Q&A.

• Cocoa price normalisation flowing into margins from FY27 onward. Cocoa prices fell roughly 75% from the 2025 peak, but Delfi hedges 6-12 months forward and holds inventory up to 18 months, so the benefit will hit reported margins gradually. Management explicitly told the AGM the recovery will be "not immediate." (FY 2025 press release, Feb 24 2026; AGM, Apr 23 2026; UOB/RHB analyst notes cited in nextinsight)

• Own-brand mix shift accelerating. Management let an agency account terminate during FY2025 and explicitly steered the portfolio toward higher-margin own brands; own brands grew 4.9% in FY2025 while agency declined 7.4%. The agency-termination drag rolls off in 2026. (FY 2025 press release, Feb 24 2026)

• Regional Markets outperforming Indonesia. Regional Markets grew 7.2% in 1H 2025 versus a decline in Indonesia, suggesting Philippines (where own brands Goya and Knick Knacks live) and Malaysia are pulling weight. (1H 2025 press release, Aug 2025)

  Management on 1H 2025 own-brand performance: "Proprietary brands grew 3.1% year-on-year, or 5.9% in constant currency... necessary to maintain brand visibility and consumer engagement." (1H 2025 press release)

• Indonesian government consumer stimulus supporting demand. Management cited government stimulus packages as having supported own-brand sales in Indonesia despite the Rupiah weakness. (FY 2025 press release, Feb 24 2026)

• Capex programme shifting from capacity to productivity. Capex stepped up from US$16.3M (FY23) to US$27.5M (FY24) for the automation programme, then moderated. Management told the AGM that the heavy capacity-expansion phase is largely complete and future capex will focus on "production automation, labour efficiency, and food quality." Translation: lower capex intensity going forward, more cash available for dividends. (AGM, Apr 23 2026)

• Distribution capability expansion. Management committed to "reinforce market leadership through brand investment and product innovation while expanding distribution capabilities in 2026." (FY 2025 press release, Feb 24 2026)

• CEO succession bench in place. John Chuang transitioned to Executive Chairman, with the company stating it has "decentralised decision-making and cultivated a strong leadership bench"; Chin Koon Yew was appointed Lead Independent Director. No named successor yet but the structure is moving. (AGM, Apr 23 2026)

• Cocoa fall versus Indonesian Rupiah trade-off. Management framed entering 2026 with "guarded optimism" on cocoa, partially offset by ongoing Rupiah weakness (depreciated ~3.8% in FY2025 against USD). (FY 2025 press release, Feb 24 2026)

8. Key risks

Cocoa price relapse before hedges roll off. Delfi's hedge book is the thing standing between cocoa volatility and the P&L. The current situation is that high-cost hedges are still being worked through; if cocoa prices fall further the company is essentially holding expensive inventory while competitors who hedge shorter benefit faster. The mechanism is straightforward: gross margin recovers more slowly than the spot cocoa price would suggest. Management has been explicit about this in successive disclosures.

Management at the 2026 AGM: "Margin recovery will not be immediate due to industry hedging practices and inventory depletion cycles... gradual margin recovery is expected as existing forward contracts expire and inventories are depleted." (AGM, Apr 23 2026)

Indonesian Rupiah depreciation. Most raw material is USD-denominated; most revenue is IDR. The mechanism is mechanical: every 1% IDR depreciation gives the gross margin a meaningful headwind. UOB analysts (cited in nextinsight) estimated that 3.9% IDR depreciation in 2024 contributed roughly 110bps of gross margin compression. The Rupiah depreciated a further ~3.8% in FY2025. If IDR weakness continues into 2026, it will offset some of the cocoa benefit when it arrives.

Competitive intensity from Mayora forcing sustained promotional spend. Delfi's market share has drifted from around 45% several years ago to roughly 34% in 2024 per Euromonitor. The mechanism is well understood: Mayora's larger sales force and broader category footprint let them outspend on promotions, and Delfi has to match or lose shelf. Management increased promotional spending materially in 2024-2025 to defend share. This is not a catastrophic risk but it is a structural margin drag that may not fade with cocoa.

Agency portfolio further attrition. Agency brands are someone else's products; the principal can terminate. FY2025 already saw one unspecified agency account terminate, costing roughly 7.4% of agency revenue. If a larger agency relationship (Toblerone, Pringles, Kellogg's) were to end, the impact would be larger. Lower probability per relationship but the cumulative tail risk across the portfolio is real.

Family/holding-company governance overhang. The Chuang family (via Aerodrome International and Johnsonville Trust) holds around 52% of the company. John Chuang transitioned to Executive Chairman in 2025 with no named successor. The mechanism: a family-controlled SGX small-cap with concentrated voting power and no clear succession plan carries governance risk that institutional investors typically discount.

Modern-trade consolidation eroding the traditional-trade advantage. Delfi's biggest relative edge versus multinationals is reach into the 400,000+ traditional-trade points-of-sale. As Alfamart/Indomaret-style minimarkets consolidate share away from warungs, the relative advantage narrows. The mechanism is slow but steady: each year modern trade takes another point or two of Indonesian retail share, and on modern trade Delfi competes more directly against Mondelez/Mars/Nestlé on their preferred terrain.

Geopolitical/logistics shocks. Management cited Middle East conflict driving energy and logistics costs at the 2026 AGM. The mechanism is mostly indirect (input costs, freight) but the company also flagged that government fuel subsidies in Indonesia and Malaysia have been buffering the impact, which means the buffer could disappear if subsidies are cut.

9. Walk the talk

The four reporting periods reviewed are: 1H 2024 release (August 2024), FY 2024 release (February 26, 2025), 1H 2025 release (August 2025), and FY 2025 release (February 24, 2026). The most recent is within ~90 days of today (2026-05-28). The April 23, 2026 AGM Q&A provides additional management commentary.

Starting with 1H 2024. Management's framing then was that cocoa prices, having spiked through 2023 and 2024, would continue to pressure margins through the hedging cycle, and that competitive intensity in Indonesia was forcing elevated promotional spending. Delivery: this played out exactly as management framed. PATMI for the full year FY2024 came in at US$33.9 million, a 27% YoY decline, with gross margin compressing from 28.5% to 27.4% on the cocoa and currency drag. They did not sugar-coat what was coming; they delivered exactly what they said was coming.

Moving to FY 2024 (announced February 26, 2025). Management framed the FY2025 outlook around continued cocoa pressure, ongoing Rupiah weakness, and a focus on own-brand growth. Their explicit guidance for FY2025 was that elevated cocoa would persist and that own-brand sales growth was the priority. Delivery: own-brand sales grew 4.9% in FY2025, in line with the strategic direction; agency declined 7.4% because they let an account go (consistent with the stated own-brand pivot); gross margin compressed a further 90bps to 26.5%, also consistent with their pre-warning that cocoa would remain a drag. The full-year dividend held at 2.72 US cents (vs 3.24 US cents in FY2024 - a cut, see Section 10), which was conservative but not surprising given the earnings pressure they had flagged.

By 1H 2025 (August 2025), management's language had shifted slightly more cautious. They noted that "a challenging operating environment for the remainder of 2025 and into 2026" should be expected, citing geopolitical tensions, inflation, and cocoa. They also flagged that proprietary brands were growing 3.1% YoY (5.9% constant currency), and were explicit that "increased promotional spending was necessary to maintain brand visibility." This was a careful set-up for the full year. Delivery against the 1H signal: FY2025 came in roughly where the 1H trajectory pointed - own brands grew, agency declined, margins compressed but only modestly, EBITDA fell 1.9% (much less severe than the 26% 1H EBITDA decline, because the year-on-year compare improved in 2H as the 2H 2024 base was already cocoa-impacted).

By FY 2025 (February 24, 2026) and the AGM in April 2026, John Chuang's tone shifted to "guarded optimism" on entering 2026, with cocoa down 75% from the peak but margin recovery flagged as "not immediate." The honest framing was: we expect benefit but it takes quarters, not weeks, to flow through.

John Chuang at the 2026 AGM: "We are very focused on maintaining a high level of situational awareness, to stay agile." (FY 2025 release language, Feb 24 2026)

The credibility read: this management team consistently underpromises slightly and explains the mechanism. They flagged the cocoa cycle, the hedge lag, the IDR headwind, and the agency-portfolio attrition all before they hit the P&L, and the outcomes matched the framing. They were less successful in defending market share against Mayora (the long-term share drift from ~45% to ~34% is the longer-run flag), and they have not yet named a CEO successor despite the chairman transition - those are the credibility gaps. On near-term operational guidance and capital allocation discipline, however, they have done what they said.

Net: a "does what they say" management team, with the known gaps being long-run share defence and CEO succession clarity.

10. Shareholder friendliness index

Dividends. Delfi pays half-yearly (interim in September, final in May after the AGM). DPS in US cents per share over the last three financial years: FY2023 4.41 US cents (8.06 US cents including a special dividend that year per Asian Century Stocks reference), FY2024 3.24 US cents, FY2025 2.72 US cents. The trajectory is a clear step-down: FY2024 down ~27% on FY2023 ordinary DPS, FY2025 down a further ~16%. The payout ratio policy remained anchored at roughly 50% of PATMI, which means the DPS cut tracks the earnings decline rather than reflecting any change in capital-return philosophy. Special dividends have been paid in better-earning years (e.g., FY2023) but not in FY2024 or FY2025.

Buybacks and dilution. The MoatMap MOATMAP DATABASE block at the top of this report shows zero share-buyback filings for P34.SI over the last 3 years; data freshness is flagged as "unknown (no scrape recorded)" so MoatMap may not have tracked this ticker. Cross-checked against the company's IR insider-transactions page, there is no evidence of an active share buyback programme over the period - all recorded transactions are substantial-shareholder filings and trust-restructuring entries, not company repurchases. Share count has been broadly stable around 611 million shares; no material option-driven dilution. Capital return is therefore done entirely through dividends, with no buyback mechanism in active use.

Verdict: Returns Capital - dividend-only, payout-ratio-anchored. Management runs a clean ~50% payout ratio with no buybacks; when earnings fell, the dividend fell proportionally rather than being defended by leverage or by drawing on cash.

11. Insider activities

Data source note. The MOATMAP DATABASE block at the top of this report shows 0 insider transactions for P34.SI with data freshness marked as "unknown (no scrape recorded)" - this indicates MoatMap has not yet scraped this ticker rather than confirming there has been no activity. To avoid a silent omission, the disclosures below are taken from Delfi's own IR insider transactions page (delfilimited.listedcompany.com/stock_insider.html), which is explicitly named in the report instructions as the canonical fallback when the primary regulatory feed is not available and which aggregates the underlying SGXNet filings.

Material transactions over the last 12 months (June 2025 - May 2026):

All transactions are cited (SGXNet via Delfi IR Insider Trades). The two large 317.7 million-share entries in March 2026 and December 2025 are the same block of shares moving through a family trust restructuring - they are deemed-interest disclosures, not open-market trades. They do not represent the Chuang family adding to or reducing economic ownership; the family retains effective control through Aerodrome International / Johnsonville structures, with the consolidated stake hovering at ~52%.

Buys - reading the signal. The only genuine open-market buy in the window is the Fidelity (FMR LLC) accumulation of 177,000 shares in March 2026. Fidelity is a substantial-shareholder filer rather than an insider in the management sense, so this is more accurately read as institutional accumulation than as insider conviction. There were no open-market buys from John Chuang, board members or named executives over the 12-month window. Absence of management open-market buying is mildly notable given that the stock had been depressed by the cocoa cycle through 2024-2025 - if management were strongly convicted that the cocoa-cycle recovery was about to break through to the P&L, this would be a plausible window for personal buying, and it did not happen.

Sells - working out the why. The FPA / First Pacific Advisors / Steven Romick entity has been a long-term SSH on the register and has been trimming gradually since 2023 (the IR records show similar small disposals in January, March and May 2023, plus the April 2026 sale). The pattern is consistent with portfolio management at a US value fund rather than any negative read on Delfi specifically. No insider footnote on the sells indicates anything other than market disposal. Reason not directly disclosed in the filings; the multi-year cadence suggests scheduled rebalancing.

Net assessment. The 12-month picture is dominated by trust-restructuring deemed-interest filings (no economic insider action), small SSH trimming by FPA (long-standing pattern), and modest SSH accumulation by Fidelity. No management or director open-market buying; no management or director open-market selling either. Read: neutral - no signal either way from insiders, with mild concern attached to the absence of conviction buying during a depressed earnings window.

12. Scenarios

Bull case. Cocoa stays off its 2025 peak through 2026 and 2027. As Delfi's high-cost hedges roll off through 2026, gross margins start visibly recovering in the 2H 2026 reporting period, with the full benefit landing in FY2027. The Indonesian Rupiah stabilises against the USD as Bank Indonesia tightens or Fed cuts narrow the rate differential. Indonesian consumer demand remains supported by government fiscal stimulus and a continuing modern-trade build-out. Mayora's promotional intensity eases as the cocoa-cost relief also relaxes their need to defend share. Own-brand sales continue compounding mid-single-digit; the agency drag from the FY2025 contract termination annualises out by mid-2026 and growth optics normalise. The Philippines business compounds high-single-digit on Goya and Knick Knacks. The capex programme moderates as guided, freeing cash for dividend restoration toward the FY2023 ordinary payout level. CEO succession gets resolved with a named internal successor and the governance discount narrows.

Base case. Cocoa prices drift sideways at a level meaningfully below the 2025 peak but above pre-spike norms. Margin recovery happens but is partially offset by ongoing IDR weakness, exactly as the AGM 2026 guidance suggested. Own brands grow mid-single-digit, agency stabilises at the post-termination level, and reported revenue growth ticks up modestly in FY2026 vs FY2025. Gross margin recovers by maybe 50-100bps over 2026-2027 rather than a full snap-back. Dividend held flat or slightly recovered as payout-ratio policy translates the modest earnings recovery into modest DPS growth. Competitive intensity from Mayora remains structural, requiring sustained promotional spending. CEO transition continues without major incident, but no clear succession announcement.

Bear case. Cocoa spikes again on a fresh West African crop shock, catching Delfi's hedge book on the wrong side and extending the margin pressure into 2027. Indonesian Rupiah depreciation accelerates if global rate dynamics turn against EM Asia, compounding the input-cost drag. A second meaningful agency contract terminates and the agency revenue base steps down again. Mayora intensifies promotional spend further on the back of its broader category cash flow, forcing Delfi into a deeper marketing investment that compresses operating margin even after gross margin recovers. Modern-trade consolidation accelerates in Indonesia, eroding the traditional-trade reach advantage faster than expected. Family succession remains unresolved into 2027-2028 and institutional investors continue to apply a governance discount.

Sources:

• Delfi FY2025 Financial Results - Minichart
• Delfi FY2025 results announcement PDF (SGX)
• Delfi 1H2025 Financial Results - My Sweet Retirement
• Delfi 1H2025 Press Release (SGX)
• Delfi 1H2024 Press Release (SGX)
• Delfi 2026 AGM Shareholder Q&A - Minichart
• Delfi (DELFI SP) 2025 update - Asian Century Stocks (paid)
• DELFI: Cheap Cocoa Looks Sweet (UOB / RHB cited) - nextinsight
• DBS Vickers analyst note on Delfi (Feb 2026)
• DBS Vickers analyst note on Delfi (Feb 2025)
• Delfi Limited Competitive Strengths - IR
• Delfi Limited History - IR
• Delfi Limited Insider Trades - IR
• Delfi Limited Wikipedia
• Delfi Limited 3QFY2024 Results - The Edge Singapore
• Indonesia Chocolate Confectionery Market - StrategyHelix
• Demand for chocolate products on the rise in Indonesia - S-GE

Report complete. Delfi's four most recent reporting periods (1H 2024, FY 2024, 1H 2025, FY 2025) plus the April 2026 AGM Q&A were used as the primary management-commentary sources, with cross-references to DBS, UOB and RHB analyst notes and a paid Asian Century Stocks update. Section 13 omitted - SemiAnalysis, Stratechery and MBI Deep Dives do not cover this name.