The best-known blockchain supply chain story begins on 24 September 2018, when Walmart told the suppliers of its fresh leafy greens that within a year they would have to trace their produce back to the farm on a blockchain. The retailer said its pilot with IBM had cut the time needed to trace a product from a store shelf to its farm from seven days to 2.2 seconds. Four years later, on 29 November 2022, Maersk and the same IBM announced the end of TradeLens, their blockchain platform for container shipping. Same vendor, same family of technology, opposite fates. That contrast is the most useful thing a beginner can learn about the subject.
This lesson follows our overview of what blockchains are used for, where supply chains took one paragraph. Here we go case by case: food with IBM Food Trust, Walmart and Carrefour; diamonds with Everledger and De Beers’ Tracr; luxury goods with the Aura consortium; shipping with TradeLens. Then comes the limit that no ledger removes, and what US and EU rules actually require. Facts are as of October 2026 unless a date says otherwise.
What a blockchain supply chain actually records
A supply chain is a relay. A farm harvests, a packer packs, a trucker ships, a distribution centre receives, a store sells. Each company keeps its own records, on paper or in its own software, and passes on what the next one needs. When something goes wrong, an investigator has to collect those records company by company, which is slow.
A supply-chain blockchain replaces that relay of private files with one shared record. Each company writes its own events to it: harvested on this date at this field, packed into this lot, shipped in this truck, received at this warehouse. Each entry is timestamped and linked to the previous ones, so an entry that someone changes later no longer matches what came after it. Our guide to why a blockchain is hard to alter explains that mechanism; how a blockchain works covers blocks and nodes.
These systems look little like Bitcoin. Most are permissioned: only named, vetted companies may write, and sometimes only they may read, a difference laid out in our lesson on public and private blockchains. There is usually no coin and nothing to trade, so the SEC, the CFTC and the FCA have no particular role in them; these are business databases with a shared design, not crypto-assets. And the record points to a physical object through an identifier printed on it, often a QR code or a lot number, sometimes a chip.
Companies pitched this for three jobs. Recalls, because a shared record can show in seconds which lots came from a contaminated farm. Provenance, because a buyer can see where a diamond, an orange or a handbag comes from. Paperwork, because shipping still runs on documents that pass through many hands. Each case below went after one of those jobs. In every one, adoption decided how far it got, more than the technology did.
IBM Food Trust: from Walmart’s mangoes to Carrefour’s oranges
IBM Food Trust is the best-known food traceability network built on a blockchain. Its story starts with a test on fruit.
Walmart’s mango test: about seven days down to 2.2 seconds
In October 2016, Walmart and IBM announced two proofs of concept: mangoes sold in the United States and pork sold in China. According to the case study published by the Linux Foundation’s decentralized trust project, the Walmart team first tried to trace a package of mangoes the old way and got an answer almost seven days later. With the pilot system, the same trace took 2.2 seconds.
You will often read a far more precise figure for that first trace, down to the hour and minute. We could not find it in Walmart’s or the Linux Foundation’s own documents, so we use the rounded figure those sources give.
The software was Hyperledger Fabric, an open-source framework hosted by the Linux Foundation. Walmart compared it with Ethereum and other options. Karl Bedwell, one of the Walmart engineers on the project, explained the choice in the same case study: “It is an enterprise-grade blockchain technology, and it is permissioned.” That last word matters. Walmart did not want anyone on the internet writing to its food records; it wanted its suppliers to.
The leafy greens order of September 2018
The pilot became a rule after an outbreak. In spring 2018, E. coli linked to romaine lettuce made people sick across the United States, and investigators struggled to say quickly which farms the lettuce came from. On 24 September 2018, Walmart and Sam’s Club told their leafy greens suppliers to set up real-time, end-to-end traceability back to the farm by the same time the following year, September 2019.
The release said Walmart had worked for more than a year with IBM and 11 other food companies on a network built on open-source technology. Frank Yiannas, then Walmart’s vice president of food safety, said the usual model, where each company knows only its direct supplier and its direct customer, was “outdated for the 21st century”. The release also quoted Robert Tauxe of the US Centers for Disease Control and Prevention (CDC) on how better tracing helps outbreak investigations.
The release left one thing unsaid. Walmart could impose a blockchain because it is the buyer: suppliers who wanted to keep selling to the largest US grocer had to join. That buying power is the hidden ingredient of the food cases.
Carrefour: a European founding member from 2018
In France, Carrefour moved at the same time. On 8 October 2018, Carrefour announced it was joining IBM Food Trust as a founding member and said it would apply blockchain gradually to its Carrefour Quality Line products, the own-label range it builds with selected farmers. The move was part of its Act for Food programme.
On 12 April 2022, Carrefour put the technology on its organic range. The first product was a Spanish dessert orange sold in trays of four. A QR code on the pack showed the producer’s name, the location of the field, the harvest date, analysis results and the organic certificate. The company’s release called Carrefour “the first retailer to use blockchain technology with its own-brand organic products”, and said it had been the first European retailer to use blockchain for food traceability in 2018, then extended it to its TEX clothing line in 2021. Those are Carrefour’s claims about itself.
Carrefour has not, in what we found, published a recent count of the products it tracks this way. Older targets were reported in the trade press, but we could not confirm whether they were met.
Where Food Trust stands in October 2026
No source we found announces a shutdown of Food Trust. IBM still sells it, under the name IBM Supply Chain Intelligence Suite: Food Trust, and its product page still describes a network that companies can join, built on IBM Blockchain, covering growers, processors, distributors and retailers. The page carries no date and no member figures.
The positioning has shifted towards compliance. On 18 September 2023, IBM and iFoodDS launched a joint product for the FDA’s food traceability rule, with Food Trust as the layer where companies share data. That release does not use the word blockchain at all. IBM has not published a recent member count. So the honest status is this: still on sale, still described as blockchain-based, with no public figures on how many companies write to it today.
Diamonds: Everledger collapsed, De Beers’ Tracr kept growing
Diamonds look like the ideal case. Each stone is valuable, can be graded and photographed, and buyers care where it was mined. Two projects took opposite paths.
Everledger: administration in April 2023
Everledger, led by its founder Leanne Kemp, was one of the first companies to put diamonds on a blockchain ledger, then gems and other goods. Its money ran out in 2023. On 24 April 2023, Steven Staatz and Ashley Leslie of Vincents Chartered Accountants were appointed administrators of the Australian company, according to Jeweller Magazine, after Everledger failed to secure a second round of funding. The UK-based parent, Foreverhold Limited, was placed in liquidation, with liquidators from Alvarez & Marsal appointed on 11 May 2023.
That was not quite the end. In August 2023, Kemp told JCK that a restructured Everledger was fully back in her hands, with 15 employees against 55 before, and that an investor’s failure to honour its agreement had caused the collapse, a dispute she said was in court. The new company works with Ernest Jones, the UK jewellery chain owned by Signet, on a collection whose diamonds are traced on Everledger’s ledger and which comes with a digital “jewellery passport”. We found no later report on the company’s activity, so we cannot say how it is doing in 2026.
Tracr: owned at the source, now opening to GIA
De Beers took another road. It started building its own platform, Tracr, in 2018 and opened it to the wider industry in 2023. On 29 May 2026, the Gemological Institute of America (GIA) announced a definitive agreement to take a 30% stake in Tracr. The release gives the scale: more than five million rough diamonds registered, about two-thirds of De Beers’ rough production by value. Since January 2025, the platform has shown a single country of origin for De Beers diamonds, and every newly sourced De Beers rough diamond of one carat or more is registered on it.
“Consumers deserve to know where their diamonds come from,” De Beers chief executive Al Cook said in the release, adding that buyers should feel more confident about each stone’s source. The deal builds on a 2023 initiative that put Tracr provenance information on eligible GIA grading reports. The release does not give a price or a closing date; some copies of the same release circulated with a date of 9 June 2026.
The difference between the two is the lesson. Tracr belongs to the miner that registers the stone at the moment it leaves the ground, and it now gains the lab that grades the stones. Everledger had to persuade others to send it data and to pay for the service, and it depended on outside investors to keep going.
Luxury goods: the Aura consortium, not LVMH alone
The luxury case is often reported as “LVMH’s blockchain”. It started that way and became something wider.
On 20 April 2021, LVMH, Prada Group and Cartier, part of Richemont, launched the Aura Blockchain Consortium. FashionUnited reported that it was open to all luxury brands, that each brand would keep full ownership of its own data, and that it ran on a private, multi-node blockchain using technology from ConsenSys and Microsoft. The point is to give each product a digital certificate: proof that the brand made it, where its materials came from and, later, who owns it.
As of 26 October 2026, Aura’s own site presents it as a non-profit founded by LVMH, OTB Group, Prada Group and Richemont, with more than 50 members and more than 80 million products registered, on what it calls the Aura Blockchain. Its 2026 news gives a sense of the current work:
- Bvlgari extended its digital passport to its watches (15 April);
- Prada put a digital product passport in every piece of its Re-Nylon for SEA BEYOND line (14 May);
- Aura launched a Diamond Traceability Initiative with a “Diamond Passport” that follows a stone from origin and certification through each change of owner (15 June);
- the Italian furniture maker Poltrona Frau launched digital passports for its best-known collections (25 June);
- Tiffany & Co. joined the consortium (29 June).
What Aura does well is authenticity and resale. A buyer of a second-hand bag can check that the brand registered that item. What it cannot do on its own is prove that the bag in your hands is the item behind the certificate. That depends on the physical link: a tag, a chip or a code that is hard to copy and hard to move from a real product to a fake one.
Why TradeLens shut down after four years
TradeLens was the most ambitious project of all, because it tried to bring a whole industry onto one record.
Maersk, the Danish container line, and IBM announced in January 2018 that they planned a joint venture to digitise global trade on a blockchain. By the launch of TradeLens on 9 August 2018, the plan had changed: according to the companies’ launch release, feedback from shippers and carriers led them to drop the joint-venture structure and deliver the platform through an extension of their existing collaboration agreement. At launch, 94 organisations were taking part or had agreed to. The platform was developed by IBM and GTD Solution, a Maersk division. Shipping is a good candidate on paper: a single container can involve a shipper, a forwarder, two ports, customs in two countries, a carrier, a bank and an insurer, all exchanging documents. By 2022 the platform reported more than 300 members, including ocean carriers such as CMA CGM, MSC, Hamburg Süd and Zim, plus terminals, inland depots and customs authorities. It said it handled information exchange for more than 65% of containerised trade, a platform claim we could not check independently. Its electronic bill of lading was approved by the International Group of P&I Clubs in 2021, according to Global Trade Review.
On 29 November 2022, Maersk and IBM said they would withdraw TradeLens, with the platform going offline by the end of the first quarter of 2023. Rotem Hershko, Maersk’s head of business platforms, gave the reason: “while we successfully developed a viable platform, the need for full global industry collaboration has not been achieved.” He added, as Computerworld reported, that TradeLens “has not reached the level of commercial viability necessary to continue work and meet the financial expectations” of an independent business.
Read those two sentences closely. Neither says the technology failed. The platform worked; not enough of the industry used it, and it did not earn enough money. Lars Jensen, chief executive of the consultancy Vespucci Maritime, told Supply Chain Dive that TradeLens was “born out of the blockchain hype” and that commercial use, not technical sophistication, decides the fate of such projects. The same article noted that shipping data is often confidential and hard to share.
What follows is Coindaro’s reading, built on those facts rather than taken from a source. A shared record looks neutral only if no single participant controls it. TradeLens was built by Maersk, one of the world’s largest container lines, with IBM, and it asked Maersk’s competitors to put commercially sensitive data on it. Some carriers did, as the member list shows; the industry as a whole did not, in Maersk’s own words. Walmart could make its suppliers join. Maersk could not make its rivals join.
The limit: a blockchain cannot check the physical goods
Every case above shares one weakness, and it is the most important point of this lesson. A blockchain proves that a record has not been changed since it was written. It does not prove that the record was true when it was written.
If a packer scans the wrong lot, the wrong lot is on the ledger, permanently. If a supplier relabels conventional oranges as organic before the first scan, the ledger faithfully records organic oranges. If a fake handbag carries a tag copied from a real one, the certificate checks out. Engineers call this “garbage in, garbage out”. In blockchain vocabulary it is the oracle problem: the chain only knows what someone, or some sensor, tells it about the outside world.
Researchers say the same thing in plain terms. A 2021 book chapter by Gowri Sankar Ramachandran and colleagues from Queensland University of Technology, CSIRO and the University of New South Wales, “Blockchain in Supply Chain: Opportunities and Design Considerations”, puts it in a figure caption: “Storing data in a blockchain ledger alone does not guarantee authenticity and trust.” The authors add that a blockchain has no mechanism of its own to check whether the data a user submits is correct, so the fixes lie outside the chain.
Those fixes exist, and none of them is a ledger. Sensors can record temperature in a refrigerated truck without a person typing it. Labs and auditors can test samples and sign their results. Tamper-evident seals, chips and hard-to-copy tags can tie a record to one object. In diamonds, laser inscriptions and grading reports do part of that job, which is why the GIA stake in Tracr matters. In crypto, networks such as Chainlink try to bring outside data on chain for smart contracts, but they face the same question: who vouches for the data at the source?
So the honest promise of a blockchain supply chain is narrower than the marketing. Once data is written, nobody can quietly rewrite it, and everyone with access sees the same version. That helps after a problem, when investigators want to know who wrote what and when. It does not stop a problem at the point where goods and data first meet.
US and EU rules ask for records, not for a blockchain
Regulators are pushing traceability hard. Neither of the two main rules for readers in the US and Europe requires a blockchain.
In the United States, the FDA’s food traceability rule under section 204 of the Food Safety Modernization Act (FSMA 204) asks companies that handle foods on the FDA’s Food Traceability List, which includes fresh and fresh-cut leafy greens, to keep records of seven critical tracking events: harvesting, cooling, initial packing, first land-based receiving from a fishing vessel, shipping, receiving and transformation. They must hand those records to the FDA within 24 hours of a request. The original compliance date was 20 January 2026. According to the FDA’s rule page, the agency proposed a 30-month extension to 20 July 2028 and Congress directed it not to enforce the rule before then; some sections of that page still show the January 2026 date. The rule accepts paper or electronic records. It does not mention blockchain.
In the European Union, the Ecodesign for Sustainable Products Regulation, Regulation (EU) 2024/1781 (ESPR), creates the digital product passport: a record of a product’s materials, origin, repair and recycling information, reached through a data carrier such as a QR code. The European Commission’s timetable, which it calls indicative, sets these steps:
| Step (EU digital product passport) | Date given by the Commission |
|---|---|
| DPP registry operational | 20 July 2026 |
| Decision on the remaining two DPP standards | November 2026 |
| Passport mandatory for certain batteries (EV, light transport, industrial) | 18 February 2027 |
| Delegated acts for textiles, aluminium and tyres | Q3-Q4 2027 |
| Delegated act for furniture | 2028 |
| Delegated acts for mattresses and recycled content | 2029 |
| Transition after each delegated act | at least 18 months |
Two details matter for this lesson. The EU registry holds product identifiers and registration data, while the full product data stays with the company or a passport service provider. And the Commission page does not mention blockchain. A passport can run on a blockchain, as Aura’s members show, or on an ordinary database. The law asks for the passport, not for the technology behind it.
Blockchain or a database? What the cases teach
Put the cases side by side and a pattern appears. The checklist below is Coindaro’s reading of the dated facts above, not a finding from a study or a regulator.
A shared ledger earned its place when four conditions held together:
- several companies that do not fully trust each other needed to write to the same record;
- no neutral owner was available, or the owner had the power to make the others join, as Walmart did with its suppliers;
- a shared, hard-to-alter history was worth paying for, as in a recall or a resale;
- the data could be trusted where it entered, through a scan, a sensor, a lab or a grading report.
When one company runs the system and the others only send it data, an ordinary database run by that company does the same job more cheaply. When the competitors who must contribute data will not do it, no technology fixes that, as TradeLens showed. The test of whether a project needs a blockchain at all is the subject of a later lesson in this path.
For a reader, a few questions are enough to judge the next announcement. Who writes to this ledger, and who decides who may write? What is checked at the moment goods and data first meet? Is the data visible to the buyer, or only to the companies? And is anyone outside the founding company actually using it? Our guides to what a blockchain is and to the three pillars of blockchain give the vocabulary to ask those questions.
The record so far, case by case
The table puts the dated facts of this lesson in one place, with what is still live as of October 2026.
| Project | Key date | Figure or fact | Status, October 2026 |
|---|---|---|---|
| Walmart mango pilot (IBM) | announced October 2016 | trace from about 7 days to 2.2 seconds | became IBM Food Trust |
| Walmart leafy greens order | 24 September 2018 | farm-level traceability by September 2019 | Food Trust still sold by IBM |
| Carrefour joins IBM Food Trust | 8 October 2018 | Carrefour Quality Line products | no recent product count published |
| Carrefour organic range | 12 April 2022 | Spanish dessert orange, QR code on pack | company claim |
| TradeLens (Maersk, IBM) | closure announced 29 November 2022 | 300+ members; offline by end of Q1 2023 | closed |
| Everledger (Australia) | administration 24 April 2023 | 15 staff after restructuring, 55 before (August 2023) | no later report found |
| Tracr (De Beers) | GIA deal announced 29 May 2026 | 5 million+ rough diamonds; GIA to take 30% | live; closing not confirmed |
| Aura consortium | launched 20 April 2021 | 50+ members, 80 million+ products (26 October 2026) | live |
Three of the five projects are alive in some form. Food Trust survives as a compliance product, Tracr gained a new shareholder in 2026 and Aura keeps adding brands. TradeLens closed, and the original Everledger company went into administration. The next date to watch is 18 February 2027, when the EU battery passport becomes mandatory: it will show whether companies build passports on blockchains or on plain databases.
