Three problem statements, the evidence for each, who is doing what, and where universities fit.
Terence KokExecutive Director, AI Governance & Assurance, Orion Five Engineering
About your speaker
Terence Kok
Executive Director, AI Governance & Assurance Practice, Orion Five Engineering
Former Chief AI and Innovation Officer, Meinhardt Group
Twenty-five years leading digital and AI transformation across Asia and the Middle East: national smart-city programmes, critical infrastructure and enterprise operations. Today I help leadership teams put AI into production with controls they can show a board, an auditor and a regulator.
BuiltA governed AI platform serving more than 6,000 engineers, with zero security incidents
CertifiedISO/IEC 42001 Lead Auditor Professional (BSI) · AIGP (SMU) · Registered Management Consultant
ServingCTCB CPSC Advisory Board · UN ESCAP AI for Developing Countries Forum
AwardsCDO Vision Award 2026 · Singapore's Most Influential Data & AI Leaders 2026 · Google Trailblazer Transformation Award 2024
Independent analysis from public sources. The views in this session are mine.
Case study · Food waste in Singapore
This session
Six parts, from the numbers to a research agenda.
The problem in numbersEleven years of NEA data, and what the statistic cannot see.
Rules and infrastructureThe segregation rules, Tuas Nexus and the households gap.
Predict and preventForecasting and kitchen tracking: what the evidence shows.
Collect and recoverContamination, sensors, routing and surplus matching.
Value and modelNudges, the business case and the carbon question.
Who is doing whatGovernment, industry, community and the universities.
Then: where university work is thin, routes in, a research agenda and questions for discussion.
Case study · Food waste in Singapore
The case in one slide
A steady problem, new rules coming, and gaps a university can fill.
of food waste from commerce and industry; the rest has no published split
Sources: [3] MSE, 7 Jul 2026; [1] [2] NEA statistics, Jun 2026; [13] NEA Green Bond Report, Mar 2026; [5] MSE, 5 Feb 2024.
Case study · Food waste in Singapore
01
The problem in numbers.
Eleven years of official data, and the part the statistic cannot see.
Case study · Food waste in Singapore
Food waste, '000 tonnes a year · NEA
The tonnage holds steady, and so does the 18%.
80313%
2015
79114%
2016
81016%
2017
76317%
2018
74418%
2019
66519%
2020
81719%
2021
81318%
2022
75518%
2023
78418%
2024
79018%
2025
Source: [2] NEA, Waste and recycling statistics 2015 to 2025. Orange: recycled share (NEA rounded rate). 2020 dip coincides with COVID-19.
Case study · Food waste in Singapore
What the statistic cannot see
Prevention and rescued food do not show up in the headline.
~40%
Commerce and industry
The government's share for these sources. No split is published for households, hawkers and small outlets.
Not counted
Rescued food
"Food that is rescued or redistributed is not reflected in our food waste statistics." Minister, July 2026.
2017
Last household study
Food was about half of household waste disposed, and over half of that was avoidable.
Sources: [5] MSE, 5 Feb 2024; [3] MSE, 7 Jul 2026; [26] NEA household waste study, Dec 2017.
Case study · Food waste in Singapore
Targets and landfill
The national targets have not yet moved.
30%
less waste to landfill per person per day by 2030against 2018; Zero Waste Masterplan, 2019
Flat
"about the same" as the 2018 baselineParliament, February 2025
2035
Semakau, the only landfill, expected to fillmore than half full, January 2024
2027
Zero Waste Masterplan review dueannounced June 2026
Sources: [8] Zero Waste Masterplan, 30 Aug 2019; [9] MSE, 5 Feb 2025; [10] MSE, 9 Jan 2024; [7] MSE, 7 Jul 2026.
Case study · Food waste in Singapore
02
Rules and infrastructure.
Segregation for large new premises, a plant that keeps moving, and households outside the rules.
Case study · Food waste in Singapore
Resource Sustainability Act · since 8 March 2024
Large new premises must segregate, treat and report.
Malls and hotels with more than 3,000 sqm of food and function space, food factories with more than 750 sqm of operations, and large multi-user food factories, in buildings planned from 2021.
About 360 premises would be in scope once existing buildings join. They have no date.
Sources: [11] Regulations S 194/2024; [4] NEA, Jun 2026; [12] NEA, 4 Mar 2020.
Case study · Food waste in Singapore
Tuas Nexus food waste line · 400 tonnes a day
The rule for existing premises waits on one plant.
2025
Jan 2019 · NEA target after the Ulu Pandan trial [15]
2024
Mar 2020 · existing premises, plant "operational in 2024" [12]
Mid-2026
May 2024 · construction delays reported [14]
2027+
Mar 2026 · "progressively from 2027 onwards" [13]
No date
Jun 2026 · existing buildings "in tandem with" Tuas Nexus [4]
Food waste is pulped and co-digested with used-water sludge; the Ulu Pandan trial gave about 40% more biogas than digesting the two apart [15].
Case study · Food waste in Singapore
Households and hawker centres
The largest unmeasured share sits outside the rules.
No household mandate
"We do not have plans" to require it; trial participation could not be sustained. October 2025.
RFID charging judged hard
High-rise blocks share a refuse chute, so weighing per household is difficult. July 2026.
Tray return since 2021
A natural point to separate hawker food waste. No published pilot results found.
Sources: [16] MSE, 15 Oct 2025; [17] MSE, 7 Jul 2026; [68] NEA, May 2021; [55] NUS CDE, Nov 2021.
Case study · Food waste in Singapore
Where AI can act · the food waste flow
Three problem statements map onto one flow.
Sources: [3] [35] [13]. The mapping is my reading.
Case study · Food waste in Singapore
03
Predict and prevent.
The strongest evidence of the three, with limits that matter for local food.
Case study · Food waste in Singapore
Reported food waste reduction · by who measured it
Independent results are smaller than the sales figures.
Retail forecasting, 1,300+ storesIndependent pilot
14.8%
Hotel kitchens, year oneProgramme study, 42 sites
21%
IKEA production waste vs 2017Company-reported
54%
Kitchen CV rangeVendor claim (Winnow)
40-70%
Sources: [40] PCFWC, Dec 2022; [42] Champions 12.3, 2018 (restaurants 26% [43]); [39] Ingka, Sep 2022; [31] Winnow via TechNode, Jul 2026 (Lumitics: 30% in six months). Measures differ by row.
Case study · Food waste in Singapore
The best independent result · 2020 to 2022
Forecasting cut store waste by 14.8%.
AI ordering pilots in more than 1,300 stores of two US grocers. Results were visible after about eight weeks.
The main barrier was not the model. Merchandising, finance and operations had conflicting goals.
Source: [40] Pacific Coast Food Waste Commitment, Dec 2022. Vendor marketing figures for the same category run higher.
Case study · Food waste in Singapore
Kitchen tracking in Singapore and beyond
Camera tracking works where staff act on the numbers.
54%
IKEA production waste
Against 2017, over about four years, with Winnow and 20,000 staff trained. Excludes plate waste.
4.1% → 2.3%
SATS waste intensity
FY2021 to FY2023, after AI tracking in its Singapore kitchens. Company-reported.
Jul 2026
Lumitics sold to Winnow
The Singapore vision company, used by Pullman Hill Street and others, now UK-owned.
Black soldier fly blueprint for urban food waste [56] [57]
No
A*STAR, Mottainai
Three CTRL valorisation awards, S$5.3m combined [23]
No
Republic Poly
Okara protein and peptides, licensed in 2026 [60]
No
Case study · Food waste in Singapore
Where university work is thin
Five gaps a new programme could fill.
Forecasting outside retail. Hawkers, caterers, central kitchens and canteens.
Household and hawker measurement. No repeated survey or diary panel; the last NEA study was 2017.
Vision data for local food. No published Singapore set of food waste images.
Collection logistics. Moving segregated food waste from existing premises to Tuas Nexus.
A cost and carbon model. Cost per tonne avoided, for operators and for the state.
My reading of [58] [26] [47] [13] [59]. Valorisation chemistry is the crowded area.
Case study · Food waste in Singapore
Engaging universities
Routes in, and what each suits.
Route
Suits
Pace
TREASURES Node 1
A food waste flow and analytics workstream in a funded centre, to 2030 [48] [49]
Months
NEA CTRL calls
Industry and institutions co-apply [22] [23]
Call cycle
SMU-X
Low-cost analytics pilot on a real data set [52] [62]
One term
SIT work-study
Students placed in an operator for 8 to 12 months [63]
Semester
NUS Zero Waste
Canteen trials before hawker centres [54]
Semester
NTU FRESH
Shelf-life and spoilage trials with retailers [50]
Months
Case study · Food waste in Singapore
International benchmarks
Universities elsewhere own the baseline.
−17%
Wageningen
Runs the Dutch national food waste monitor; waste fell 17% from 2015 to 2023.
Panels
Leeds and Ohio State
Household surveys and food diaries, repeated on a schedule.
Open
ReFED
A public cost-benefit model of 40+ solutions, not just a tonnage.
Sources: [64] Rijksoverheid, Sep 2025; [65] Leeds, Jan 2021; [67] Ohio State; [66] ReFED. Seoul's weight-based fees: 47,000 t less over six years, reported [46].
Case study · Food waste in Singapore
A research agenda
Seven questions to put to the universities.
Baseline. How much food waste do households, hawkers and small outlets produce, and how much is avoidable?
Forecasting. Can a demand model cut overproduction at a central kitchen, against a control group?
Local vision data. Can an open image set lift accuracy on mixed dishes?
Contamination at source. Can a camera or sensor catch it before the digester?
Collection design. Least-cost routes from existing premises to Tuas Nexus.
Nudges at the tray return. Do portions and signage work in hawker centres?
The value model. Cost per tonne avoided for each measure, published openly.
My recommendations, drawn from [65] [67] [40] [47] [35] [13] [45] [66].
Case study · Food waste in Singapore
07
What leaders can take from it.
Seven lessons, a pilot checklist and questions to take back.
Case study · Food waste in Singapore
Lessons 1 and 2 · before the technology
LESSON 01
Measure first.
The national statistic cannot see prevention or rescued food.
ActionSet a weighed baseline before any pilot starts.
LESSON 02
Fix the input before the machine.
Contamination sank IUT Global and holds back Changi's digesters.
ActionFund detection and staff training ahead of treatment capacity.
Lessons are my reading of the record: [3] [37] [35].
Case study · Food waste in Singapore
Lessons 3 and 4 · where to invest
LESSON 03
Back prediction where someone will act.
Every gain came from staff changing what they order or cook.
ActionChoose partners who will change the decision, not only host the tool.
LESSON 04
Ask universities for the thin areas.
Valorisation is well served; measurement and modelling are not.
ActionCommission forecasting, measurement, vision data, logistics and the value model.
Lessons are my reading of the record: [39] [40] [58].
Case study · Food waste in Singapore
Lessons 5 to 7 · making it last
LESSON 05
Use existing institutions.
TREASURES, CTRL and SMU-X give funding, a host and a pace.
ActionAnchor the work where it outlasts one grant.
LESSON 06
Plan for vendor change.
A sale, a restructure and a liquidation in the record.
ActionOwn the data and make sure it can move.
LESSON 07
Count value in Singapore terms.
Food cost, disposal, compliance and landfill life.
ActionDo not plan on carbon credit income.
Lessons are my reading of the record: [48] [22] [31] [37] [24].
Case study · Food waste in Singapore
For government
Count what policy wants to change.
Count prevention
Rescued and prevented food is invisible in the main statistic today.
Use the 2027 review
The masterplan review is the natural moment to change the measure.
Give it a method
A university-run baseline, on the Wageningen model, would make the count credible.
Tuas Nexus food waste line, and the date for existing buildings.
Jun 2027
2026 statistics: whether food waste recycling moves off 18%.
2026 to 2030
TREASURES grant calls, and whether Node 1 takes on food waste.
Ongoing
Results from Singapore sensor and smart-bin pilots.
Case study · Food waste in Singapore
Evidence and methodology
How strong each claim in this case is.
Claim
Rests on
Grade
Food waste generated, recycled and recycling rate, 2015 to 2025
NEA statistics tables, June 2026
Measured, official statistics
Commercial and industrial sources are about 40% of food waste
MSE written reply, February 2024
Official estimate, no method published
Rescued food is not in the statistics
MSE written reply, July 2026
Official statement
RSA dates for new and existing buildings; Tuas Nexus timing
Regulations S 194/2024; NEA pages; Green Bond report
Primary legal and agency documents
Blue-bin contamination about 40%
MSE replies, 2023 and 2026
Official, aggregated sampling
Independent retail forecasting pilots averaged 14.8%
Pacific Coast Food Waste Commitment, 2022
Measured, multi-retailer pilot
IKEA cut production food waste 54% against 2017
Ingka Group, 2022
Company-reported, unaudited
Kitchen vision tracking cuts 30% to 70%
Winnow and Lumitics materials
Vendor claim
Smaller plates cut buffet waste 19.5%
Kallbekken and Sælen, 2013
Peer-reviewed field experiment
Hotels and restaurants return about 7:1
Champions 12.3, 2018 and 2019
Programme study, self-selected sites
University activity by problem area, and the gaps
University and agency releases, read October 2026
My call, from what is published
Engagement routes, research agenda and controls checklist
My reading of the record cited against each
My call
Case study · Food waste in Singapore
Full timeline · 1 of 3
January 2011 to October 2022.
Jan 2011IUT Global, then Singapore’s largest food waste digestion plant, folds. Built for 800 t/day, it collected 120-130 t/day, with up to 40% of intake contaminated.[37]
Dec 2016NEA starts co-digesting food waste with used-water sludge at Ulu Pandan. In January 2019 it reports about 40% more biogas than digesting the two apart.[15]
Dec 2017NEA household waste study: food is about half of household waste disposed, and over half of that is avoidable.[26]
7 Mar 2019NEA says segregation will be mandatory for large commercial and industrial food waste generators from 2024.[72]
30 Aug 2019Zero Waste Masterplan: 30% less waste to landfill per capita per day by 2030, 70% overall recycling, food named a priority stream.[8]
Jun 2021Tray return becomes mandatory at hawker centres, coffeeshops and food courts; enforcement from September.[68]
Nov 2021NUS and NEA pilot a containerised digester at East Coast Lagoon Food Village.[55]
Feb 2022NUS, the Singapore-ETH Centre and NTU start a black soldier fly blueprint for urban food waste.[56]
Oct 2022FairPrice and A*STAR launch BINgo AI/IoT smart bins for recyclables at three sites.[38]
Case study · Food waste in Singapore
Full timeline · 2 of 3
February 2023 to January 2026.
Feb 2023MSE puts blue-bin and chute contamination at about 40%. Olio restructures and its Singapore head leaves.[74] [70]
Jun 2023SMU and FairPrice report a student-built write-off dashboard with predictive analytics.[52]
8 Mar 2024Food waste segregation, treatment and reporting start for new large malls, hotels and food factories.[11] [4]
May 2024Tuas Nexus hit by construction delays; completion reported as mid-2026.[14]
1 Jan 2025Good Samaritan Food Donation Act 2024 comes into force.[25]
5 Feb 2025Waste to landfill per capita per day is “about the same” as the 2018 baseline.[9]
Mar 2025Sembcorp completes the sale of its waste business, rebranded Cora Environment in September 2025.[69] [21]
15 Oct 2025MSE: no plan to require households to segregate food waste; trial participation could not be sustained.[16]
1 Jan 2026Carbon tax rises to S$45 per tonne. TREASURES, the NEA and NTU research centre, starts its term.[24] [48]
Case study · Food waste in Singapore
Full timeline · 3 of 3
February 2026 to July 2026.
24 Feb 2026Blue-bin contamination “remained at about 40%”.[19]
Mar 2026NEA Green Bond report: Tuas Nexus food waste line “progressively from 2027 onwards”.[13]
20 Apr 2026NTU reports an AI model that predicts microbial growth so safe food is not thrown away.[50]
17 Jun 20262025 statistics: 790,000 t of food waste, 18% recycled. Zero Waste Masterplan review announced for 2027; TREASURES launched.[1] [7] [48]
7 Jul 2026Parliament: rescued food is not counted in the statistics; RFID charging judged hard with high-rise chutes.[3] [17]
16 Jul 2026Winnow announces it has acquired Lumitics, the Singapore food waste vision company.[31]
7Ministry of Sustainability and the Environment. (2026, 7 July). Oral reply to parliamentary question on waste statistics and review of the Zero Waste Masterplan.www.mse.gov.sg
8Ministry of the Environment and Water Resources. (2019, 30 August). Launch of Singapore’s Zero Waste Masterplan (press release, National Archives).www.nas.gov.sg
9Ministry of Sustainability and the Environment. (2025, 5 February). Oral reply on progress on waste to landfill and solar deployment.www.mse.gov.sg
10Ministry of Sustainability and the Environment. (2024, 9 January). Written reply to PQ on extension of Semakau Landfill lifespan.www.mse.gov.sg
11Singapore Statutes Online. (2024). Resource Sustainability (Food Waste Segregation, Treatment and Reporting) Regulations 2024, S 194/2024.sso.agc.gov.sg
12National Environment Agency. (2020, 4 March). Businesses required to segregate food waste for treatment under new legislation.www.nea.gov.sg
13National Environment Agency. (2026, March). Green Bond Report.www.nea.gov.sg
14Eco-Business. (2024, 20 May). Singapore’s world-first integrated waste and water treatment facility hit by construction delays.www.eco-business.com
15National Environment Agency. (2019, 29 January). Co-digestion of food waste and used water sludge enhances biogas production for greater energy generation.www.nea.gov.sg
Case study · Food waste in Singapore
Sources · 2 of 5
Agencies, statutes and Parliament.
16Ministry of Sustainability and the Environment. (2025, 15 October). Written reply on food waste segregation for households.www.mse.gov.sg
17Ministry of Sustainability and the Environment. (2026, 7 July). Written reply to parliamentary question on food waste (RFID-based charging).www.mse.gov.sg
18National Environment Agency. (2026). Waste collection systems (public waste collectors and sectors).www.nea.gov.sg
19Ministry of Sustainability and the Environment. (2026, 24 February). Written reply on contamination of recyclables in blue bins.www.mse.gov.sg
20Ministry of Sustainability and the Environment. (2023, 4 July). Written reply to PQ on sensors in recycling bins.www.mse.gov.sg
21Ministry of Sustainability and the Environment. (2025, 11 September). Speech by SMS Zaqy Mohamad at the Cora Environment brand launch.www.mse.gov.sg
35Sustainability Open Innovation Challenge. (n.d.). Helping F&B operators improve food waste segregation and transfer to an on-site treatment facility (Changi Airport Group problem statement).sustainability-industry.innovation-challenge.sg
36Marina Bay Sands. (2026). Waste and circularity.www.marinabaysands.com
38A*STAR. (2022, 6 October). FairPrice Group and A*STAR launch pilot for smart bins (BINgo).www.a-star.edu.sg
39Ingka Group. (2022, 20 September). IKEA stores halve production food waste, saving more than 20 million meals over four years.www.ingka.com
40Pacific Coast Food Waste Commitment. (2022, December). Case study: AI to reduce food waste in retail.pacificcoastcollaborative.org
41Wasteless. (n.d.). Edeka achieves waste reduction and margin gains with AI-powered markdowns (vendor case study).wasteless.com
42Champions 12.3 and WRAP. (2018). The business case for reducing food loss and waste: hotels.champions123.org
43Champions 12.3. (2019, 13 February). Report finds restaurants save significant money from fighting food waste.champions123.org
44WBCSD. (2017, March). New research finds companies saved $14 for every $1 invested in reducing food waste.www.wbcsd.org
45Kallbekken, S., & Sælen, H. (2013). “Nudging” hotel guests to reduce food waste as a win-win environmental measure. Economics Letters, 119, 325-327. Summary via The Decision Lab.thedecisionlab.com
Case study · Food waste in Singapore
Sources · 4 of 5
Universities and research institutes.
46World Economic Forum. (2019, April). South Korea once recycled 2% of its food waste. Now it recycles 95%.www.weforum.org
47Rokhva, S., & Teimourpour, B. (2025, 19 July). Food waste estimation based on computer vision in a university dining hall. arXiv 2507.14662.arxiv.org
48Nanyang Technological University. (2026, 17 June). National research centre to turn trash on landfill to treasure (TREASURES).www.ntu.edu.sg
50NTU School of Chemistry, Chemical Engineering and Biotechnology. (2026, 20 April). AI-powered predictive modelling to enhance food safety and reduce food waste.www.ntu.edu.sg
52Singapore Management University. (2023, 8 June). Singapore Management University and FairPrice Group partner to reduce food waste.news.smu.edu.sg
53Xue, W., & Cao, K. (n.d.). Optimal routing for waste collection: a case study in Singapore. SMU Institutional Knowledge.ink.library.smu.edu.sg
54NUS Zero Waste. (c. 2022). How providing rice level options can reduce rice waste.www.nus.edu.sg
55NUS College of Design and Engineering. (2021, 17 November). Food waste digester turns leftovers into energy and fertiliser.cde.nus.edu.sg
56NUS Faculty of Science. (2022, February). Interdisciplinary team to develop blueprint for sustainable urban food waste management using black soldier flies.www.science.nus.edu.sg
57Singapore Global Network. (2026, 12 February). Meet the German scientist using larvae to turn Singapore’s food waste into gold.singaporeglobalnetwork.gov.sg
58Singapore Government. (n.d.). Research on food waste valorisation carried out by research institutes and institutes of higher learning (PDF).file.go.gov.sg
59A*STAR. (2021, 19 April). As Singapore grapples with a food waste problem, A*STAR’s scientists find ways to manage it.www.a-star.edu.sg
62Singapore Management University. (n.d.). SMU-X: experiential learning through industry partnerships.blog.smu.edu.sg
63Singapore Institute of Technology. (n.d.). Offering students a gateway to the working world (Integrated Work Study Programme).www.singaporetech.edu.sg
64Rijksoverheid. (2025, 4 September). Voedselverspilling gedaald met 17% sinds 2015 (food waste down 17% since 2015).www.rijksoverheid.nl
65Leeds University Business School. (2021, 15 January). Using the lessons of lockdown to reduce household food waste.business.leeds.ac.uk
67The Ohio State University. (n.d.). Food Waste Collaborative.u.osu.edu
68National Environment Agency. (2021, May). Mandatory for diners to return dirty trays, crockery and clean table litter at hawker centres, coffeeshops and food courts.www.nea.gov.sg
69Sembcorp Industries, via SGX. (2025, March). Sembcorp completes sale of Sembcorp Environment to SBT Investment 2.links.sgx.com
70Eco-Business. (2023, 21 February). Food-sharing app Olio’s Singapore founding country head moves on after restructure.www.eco-business.com
71Singapore Food Agency. (2025, 4 November). SFA commits $42 million to drive breakthroughs in future foods and food safety.www.sfa.gov.sg
72National Environment Agency. (2019, 7 March). Food waste segregation for treatment by large commercial and industrial food waste generators to be mandatory from 2024.www.nea.gov.sg
73Food from the Heart. (2026). Annual Report 2025 (figures read from secondary summaries).www.foodfromtheheart.sg
74Ministry of Sustainability and the Environment. (2023, 23 February). Written reply to PQ on contamination rate for centralised recycling chutes.www.mse.gov.sg
Connect on LinkedIn