The F-35 Shipment That Went to Hong Kong: Why Critical Components Need Verifiable Custody

Australia confirmed on 22 September that unserviceable F-35 components shipped from the country for repair in the United States had instead been diverted to Hong Kong. The Australian government is assisting US authorities and Lockheed Martin with the investigation.

ABC reported that the shipment included a canopy incorporating radar-resistant technology and that an intermediary was responsible for transport. Australia’s defence minister, Richard Marles, also said that, according to his understanding, no technologically sensitive equipment had been lost. The reason for the diversion, the present location of every component and any access by a third party have not been publicly confirmed.

That distinction matters. This is not evidence that China obtained F-35 secrets. It is evidence of a custody-control failure serious enough to trigger government and congressional scrutiny.

A manifest records intent, not physical reality

A shipping document can say “United States” while the physical asset travels somewhere else. In a multi-party transport chain, information moves through maintenance teams, freight forwarders, airlines, customs agents and repair facilities. A wrong routing instruction, an unauthorised change or a simple mismatch between systems can survive several handoffs before anyone sees the full picture.

Critical logistics therefore needs independent evidence tied to the component itself—not only to an aircraft, booking or airway bill.

At Sensefinity, we connect the physical unit to its digital identity. An item-level or case-level sensor can report location, movement, opening, light exposure, separation from its assigned container and abnormal loss of communication. Geofences can cover approved routes, airports and jurisdictions. When the component leaves that policy envelope, the platform can create an exception while intervention is still possible.

An alert does not prove hostile intent. It reduces the time between divergence and response.

Learn more about Sensefinity Cargo Safety.

A destination change must become a signed custody event

Rerouting is normal in global logistics. Silent rerouting is not acceptable for controlled components.

Every destination or carrier change should require:

• strong identity for the requesting party;

• dual authorisation for high-risk exceptions;

• an explicit reason and timestamp;

• confirmation through a channel independent of the change request;

• acceptance by the new carrier and consignee;

• reconciliation with the component’s live telemetry.

Each physical handoff should bind the person, organisation, component identifier, seal, time and location. If the transport record says the United States while the sensor indicates Hong Kong, the mismatch should block automatic acceptance and open an investigation.

Blockchain makes later rewriting visible

Sensors show what happened in the physical journey. Blockchain can help preserve who authorised it, what each participant accepted and which record existed at a specific time.

The right design for sensitive supply chains is permissioned. Military or commercially restricted information does not need to be published on a public chain. Hashes, identities, document versions and timestamps can be anchored while operational detail remains off-chain under access control.

This creates a shared, tamper-evident chronology across organisations that may not share one database. An investigator can compare the original manifest, authorised route, sensor events, custody receipts and later changes without trusting a single party’s retrospective export.

Blockchain does not make bad input true, and it does not physically prevent diversion. It is effective when combined with authenticated sensors, strong approvals and an operational response process.

See how our Blockchain solution connects telemetry, identity and chain of custody.

Five controls critical shipments should adopt now

1. Bind identity to the physical unit. Link every component, case and container to its sensor, seal, manifest and authorised destination.

2. Monitor policy, not just position. Alert on prohibited jurisdictions, route deviations, unplanned transfers, opening and abnormal silence.

3. Make changes explicit. Require dual approval and out-of-band confirmation for destination, carrier or consignee changes.

4. Prove every handoff. Record strong identity, time, location, condition and counterparty acceptance at each custody transfer.

5. Preserve the evidence. Keep raw telemetry and anchor approvals and versions in a permissioned, tamper-evident record with defined escalation and recovery procedures.

The lesson is broader than defence. Semiconductor equipment, prototypes, medical products, luxury goods and other high-value assets face the same gap between the route written in a system and the journey taken by the physical object.

Critical cargo should never be “somewhere in the network.” It should have a continuously verifiable identity, route and custodian.

Read the complete Prova analysis, including the confirmed facts and the limits of what is publicly known.

Sources

• ABC News — Australia confirms F-35 jet parts went missing in Hong Kong, sparking China security fears

• CNN — Australia confirms F-35 fighter jet parts were mistakenly diverted to Hong Kong

NB-IoT, LTE-M or Cat 1 bis? Choosing the right cellular link for real-world IoT

Choosing cellular connectivity for an IoT product is not a contest to find the newest acronym. It is a design decision about where the device will travel, how often it reports, how much data it sends, how long it must run and which networks are actually available along the route.

NB-IoT, LTE-M and LTE Cat 1 bis all use licensed cellular spectrum, but they solve different problems. NB-IoT and LTE-M are 3GPP low-power wide-area technologies introduced in Release 13. Cat 1 bis is a simplified version of LTE Cat 1 that uses one receive antenna instead of two.[1][3]

None is the universal winner. The right answer comes from the workload.

Three technologies, not three generations

It is tempting to arrange the options as a ladder: NB-IoT at the bottom, LTE-M in the middle and Cat 1 bis at the top. That misses the point.

NB-IoT is deliberately narrow. It gives small, infrequent messages a path through licensed networks while keeping device complexity and power demand low. LTE-M adds better mobility, responsiveness and throughput while retaining LPWA power-saving features. Cat 1 bis uses the broader LTE footprint and offers much more data capacity, but it is not classified as an LPWA technology by the GSMA.[1]

The comparison is closer to choosing a vehicle. A bicycle, a van and a truck can all move goods, but the payload and route decide which one makes sense.

Where NB-IoT has the advantage

NB-IoT is built for devices that sleep most of the time, wake up, send a small reading and return to sleep. 3GPP designed it for improved indoor coverage, large populations of low-throughput devices, low delay sensitivity, low device cost and low power consumption. It can run inside an LTE carrier, in the guard band or in dedicated spectrum.[2]

That profile fits applications such as:

• meters and tank-level sensors;

• environmental monitoring;

• stationary cold-chain sensors;

• returnable packaging that reports occasionally;

• assets stored in basements, warehouses or other difficult radio locations.

Its narrow 180 kHz carrier and coverage-enhancement modes help when penetration matters more than speed. Power Saving Mode and extended discontinuous reception can keep the modem asleep for long periods, although real battery life still depends on signal quality, reporting frequency, retries, temperature, battery chemistry and operator settings.[1]

The trade-off is responsiveness. NB-IoT is intended for small payloads and delay-tolerant traffic. Classic Release 13 deployments do not provide the seamless handover expected by fast-moving trackers; later releases added mobility improvements, but support depends on the module and network. It is not the first choice for frequent position updates, large firmware downloads or an asset that continuously crosses cell boundaries.

Where LTE-M has the advantage

LTE-M, also called Cat-M1 in its first widely deployed form, keeps many LPWA benefits while behaving more like mobile LTE. It supports handover between cells, lower latency and higher throughput than NB-IoT. This is useful when the device is attached to a pallet, trailer, container, vehicle or piece of equipment that keeps moving.

LTE-M is a good fit for:

• mobile asset and cargo tracking;

• regular GNSS position and sensor updates;

• alarms that need a faster response;

• larger over-the-air firmware updates;

• wearables and safety devices;

• products that may need voice, where the operator has enabled VoLTE for LTE-M.

The GSMA describes LTE-M as an LPWA technology with low device complexity, low power use, extended coverage, low latency and support for high connection density.[1] It also supports PSM and eDRX, so a well-designed tracker can spend most of its life asleep and wake only when movement, a schedule or a sensor event requires a report.

LTE-M's weakness is not the radio specification. It is availability. Operators have deployed NB-IoT and LTE-M unevenly, and roaming support is not identical to ordinary smartphone LTE roaming. The GSMA's launch inventory shows substantial commercial deployment of both technologies, but a country appearing on a list does not prove that the required band, operator, roaming agreement or power-saving feature will work for a specific device.[5]

Where Cat 1 bis has the advantage

LTE Cat 1 bis takes a different route. It keeps Cat 1's data capability and standard LTE operation but removes the second receive chain and antenna. That reduces hardware complexity, board area and cost. The price is a coverage penalty compared with two-antenna Cat 1, particularly near the cell edge.[1][3]

Cat 1 bis can reach the Cat 1 class maximum of about 10 Mbps downlink and 5 Mbps uplink; commercial modules advertise those figures for supported networks.[4] That capacity changes what an IoT device can do:

• transfer larger firmware images more quickly;

• upload richer diagnostics or frequent telemetry;

• support payment terminals and industrial gateways;

• send audio or limited image data where the application requires it;

• provide tracking and telematics across ordinary 4G LTE coverage.

It also supports mobility and standard LTE handover. Because it does not need the network to activate a separate NB-IoT or LTE-M radio feature, Cat 1 bis can be easier to deploy across regions where 4G LTE is present but LPWA coverage is incomplete.[1][3]

The compromise is power. New Cat 1 bis modules may support PSM and eDRX, but Cat 1 bis is not an LPWA category and network support for those modes is not uniform. Its faster transfer can reduce radio-on time for a large payload, yet a small sensor sending a few bytes may still get better energy economics from NB-IoT or LTE-M. The only reliable answer comes from measuring the complete duty cycle on the target networks.

The comparison in one view

NB-IoT

• Primary strength: deep coverage and very low power for small, infrequent messages.

• Mobility: best for stationary or slowly changing deployments; support for later-release mobility features varies.

• Typical data: tiny, delay-tolerant telemetry.

• Power profile: usually strongest for long sleep cycles and sparse reporting.

• Voice: not supported.

• Coverage dependency: requires an operator's NB-IoT deployment and compatible bands.

• Best starting point: static sensors, metering, storage and hard-to-reach locations.

LTE-M

• Primary strength: low-power mobility with moderate throughput and faster response.

• Mobility: seamless handover between cells.

• Typical data: regular telemetry, GNSS positions, alarms and manageable firmware updates.

• Power profile: a strong balance of battery life and mobile operation.

• Voice: possible when the operator supports VoLTE on LTE-M.

• Coverage dependency: requires LTE-M deployment, compatible bands and suitable roaming agreements.

• Best starting point: cargo and asset tracking that moves between cells.

LTE Cat 1 bis

• Primary strength: higher throughput across ordinary LTE infrastructure.

• Mobility: standard LTE mobility and handover.

• Typical data: frequent telemetry, large firmware images, logs, audio or limited image data.

• Power profile: sleep modes are possible, but active and network behaviour need closer power budgeting.

• Voice: possible on some modules and operator profiles.

• Coverage dependency: uses standard LTE, but bands, certification and roaming still need validation.

• Best starting point: higher-data IoT or global products where LTE-M availability is insufficient.

These are design tendencies, not procurement guarantees. A module data sheet cannot tell you whether a roaming SIM will access a feature in a particular port, warehouse or border corridor.

What to test before choosing

A connectivity decision should begin with a route and workload model.

Map the real network footprint. Check every country, operator and LTE band in the planned deployment. Test roaming, not just home-network attachment. An operator may support a technology domestically but restrict access to roaming devices.

Model the complete energy budget. Include network search, registration, TLS setup, retransmissions, GNSS acquisition, sensor warm-up and firmware downloads. The lowest sleep current does not rescue a device that repeatedly searches for a missing network.

Use the real payload. Protocol headers and security handshakes may be larger than the sensor message. Compression, batching and event-driven reporting can change the best radio choice.

Plan firmware updates on day one. A product expected to remain deployed for years will need security and application updates. Estimate the worst-case image size, transfer time and energy cost before fixing the modem architecture.

Test cell edges and movement. Laboratory signal levels do not reproduce a metal container, a refrigerated trailer, a basement or a device moving between countries. Antenna placement and enclosure design can matter as much as the modem category.

What this means for cargo and asset tracking

A stationary temperature logger in a warehouse and a high-value shipment crossing five countries are both IoT devices, but they do not have the same connectivity problem.

NB-IoT can be the right choice when the asset remains within known coverage, messages are small and long battery life has priority. LTE-M is often the stronger fit when a tracker moves, reports position and condition regularly, and must keep sessions alive through cell changes. Cat 1 bis becomes attractive when the route has broad LTE but inconsistent LTE-M availability, or when the device needs larger updates and richer data.

Some products should support more than one radio mode. A multimode design can prefer LTE-M for mobility, use NB-IoT where available for sparse telemetry, or select Cat 1 bis for a separate high-throughput product family. That flexibility has a cost in hardware, certification, firmware and testing, so it should solve a real deployment problem rather than become a feature-list exercise.

At Sensefinity, connectivity is part of the Internet of Cargo, not the whole solution. The useful outcome is verified information about location, temperature, humidity, shock, opening events and custody, delivered with the right balance of coverage, energy and cost. Our NB-IoT trackers and cargo-monitoring work begin with that operational question: what must the customer know, where and how quickly?

The decision rule

Choose NB-IoT when the message is small, the device sleeps for long periods and coverage penetration matters more than mobility or speed.

Choose LTE-M when the asset moves and the application needs LPWA power consumption with handover, lower latency and enough throughput for regular tracking and updates.

Choose Cat 1 bis when higher data rates and the ordinary LTE footprint matter more than achieving the lowest possible power profile.

Then verify the choice on the target operators, bands, routes and enclosures. A technology comparison narrows the field. Field testing makes the decision.

Sources

[1] https://www.gsma.com/solutions-and-impact/technologies/internet-of-things/wp-content/uploads/2024/10/Mobile-IoT-in-a-5G-Future-Final.pdf — Mobile IoT in a 5G Future — GSMA [2] https://3gpp.org/news-events/3gpp-news/niot — NarrowBand IoT — 3GPP [3] https://www.u-blox.com/en/blogs/insights/lte-cat-1bis — LTE Cat 1bis — u-blox [4] https://www.quectel.com/product/lte-cat-1-bis-eg916q-gl — LTE Cat 1 bis EG916Q-GL — Quectel [5] https://www.gsma.com/solutions-and-impact/technologies/internet-of-things/mobile-iot-commercial-launches — Mobile IoT network launches — GSMA

When Cold-Chain Service Cannot Be Proven: The Graciosa Fish Case

Complaints reported on 9 September 2026 about cold handling for fish air cargo travelling to and from Graciosa Island raise a simple but important question: how can an operator prove that the paid refrigeration service was actually delivered at every stage? The allegations still require investigation, but the underlying visibility gap is already clear.

An invoice is not a temperature record

A service charge confirms that refrigeration was purchased. It does not show when a fish lot entered cold storage, its temperature while waiting or transferring, how long it remained outside its permitted range, or who held custody when a deviation occurred.

For perishable cargo, those facts must travel with the lot. Sensefinity's cold-chain monitoring connects temperature, location and time data so teams can receive alerts while intervention is still possible—not discover a broken chain only after delivery.

Each transfer needs evidence

The risk is not limited to the flight. It can appear at the dock, during transport to the airport, while awaiting screening, at an intermediate stop or at final delivery. Every handover needs an accountable party, defined thermal limits and a continuous record.

Sensor data can be linked to a tamper-resistant blockchain audit trail, creating evidence of what happened, where and when. Blockchain does not refrigerate fish; it helps prove whether the cold-chain process was followed.

Read the full Prova analysis

Our partner publication Prova examines the Graciosa case in greater detail, including how real-time alerts, lot-level traceability and a Digital Product Passport can support release, inspection, quarantine and claims decisions.

Read “Fish cold chain in the Azores: proving the service was delivered” on Prova.

The lesson is practical: for temperature-sensitive cargo, availability is not enough. The supply chain must be able to demonstrate, operation by operation, that the service was delivered and the product remained within its agreed conditions.

A New Destination Is a New Custody Act

A freight destination is not just an address field. Changing it alters the authorised route, the intended consignee and the point at which custody passes to another party. When that decision lives only in an email, a phone call or an overwritten TMS field, a fraudulent redirection can look like routine administration.

Replace silent edits with signed events

We believe every destination change should create a new custody act tied to the lot: previous destination, new destination, reason, requesting organisation and person, timestamp, approval policy and a cryptographic reference to the preceding event. The old instruction remains visible. Revocations, refusals and corrections become additional events rather than rewritten history.

Strong authentication must come before signature. For sensitive loads, that means step-up or multi-factor authentication, explicit confirmation of what is being signed and dual approval when value, product or route justifies it. A valid login is not enough; the signer must also be authorised to redirect that specific lot.

Anchor proof, protect commercial data

With Sensefinity's blockchain layer, the signed document can remain access-controlled off-chain while its hash, version and timestamp are anchored in a tamper-resistant history. Authorised partners can verify that the instruction presented is the one that was signed and identify what came before it without publishing sensitive commercial details.

Blockchain cannot prove that a bad instruction was true or prevent a compromised credential. Its value is narrower and more useful: changes, sequence and conflicts become much harder to hide.

Connect authorised intent to physical movement

The signed act defines what should happen. Cargo telemetry shows what is happening. Sensefinity's Cargo Safety capabilities can detect unplanned stops and unauthorised access and support asset-level tracking. Comparing those signals with the latest authorised destination creates an actionable exception when a load leaves its approved route before a valid change exists.

Read the full Prova article

The full Prova analysis explains the event model, authentication and authorisation controls, blockchain anchoring, a seven-step operating workflow and how this approach aligns with the move toward authenticated, auditable freight data.

Read “A destination change is a custody act — and it should be signed” on Prova.

A destination change should never erase the past. It should add a signed, attributable and verifiable decision to the lot's custody history.

The $647,420 Nike Pickup That Looked Legitimate — Until It Wasn’t

A driver arrived at a Memphis facility with documents that appeared sufficient to collect a Dallas-bound shipment of Nike merchandise. The cargo left the dock. But the driver was not completing the authorized movement, the paperwork was fraudulent, and the shipment headed towards the Chicago area instead.

The fictitious pickup happened on 21 August 2026. Nike contacted the Cook County Sheriff’s Police Organized Retail Crime Unit two days later, and investigators ultimately recovered approximately $647,420 in merchandise at a distribution facility in Des Plaines, Illinois, together with a stolen trailer. At the time of the report, no arrests or charges had been announced and the investigation remained active.[1]

This was not a failure of a lock. It was a failure of trust: the wrong person presented documents that looked right long enough for legitimate staff to release valuable cargo.

When a document looks right but custody is wrong

Traditional shipping documents describe an intended transaction. They do not necessarily prove that the person standing at the dock is the person currently authorized to execute it.

A convincing PDF, bill of lading or pickup reference can be copied, altered or generated from compromised information. The Nike incident is especially instructive because the shipment did not first disappear from an unattended parking area. It was handed over at origin after false paperwork made an unauthorized collection appear legitimate.[1]

It also sits within a wider pattern. In a separate Nike case, federal prosecutors alleged that unauthorized UPS labels were used to divert products from the company’s Memphis distribution operation. Reporting on that case describes “ghost labels” covering original shipping labels and redirecting packages to private addresses; those allegations have not been proven in court.[2]

The common weakness is clear: when operational truth lives only in editable labels, emails and documents, whoever can reproduce the appearance of authority may be able to redirect the physical goods.

Blockchain should verify the handover, not archive the fraud

Simply uploading a document to a blockchain is not enough. If false information is accepted at the start, an immutable database only preserves a false statement.

The useful model is different: create a verifiable digital chain of custody and require every physical handover to match its current authorized state. Before a warehouse releases a shipment, the system should be able to answer five questions:

1. Which exact shipment is being collected? A unique digital identity binds the order, cargo unit, pallet or container to the operational record.

2. Who is authorized right now? The approved carrier, driver, vehicle and pickup window are recorded by trusted parties, rather than inferred from a document presented at the gate.

3. Has anything changed? A destination, carrier or collection instruction cannot be silently overwritten. A new authorization becomes a visible, time-stamped event.

4. Is this handover happening in the expected place and time? A one-time release credential can be bound to the shipment, facility and pickup window.

5. What happened after release? Independent IoT data confirms departure, movement and route exceptions instead of relying only on status messages from the party holding the load.

In that model, a counterfeit document is no longer the authority. It is merely a claim that must match the shared, verifiable record.

What the Nike pickup could have looked like

Imagine the same collection with digitally verified custody controls.

At the Memphis gate, the operator scans the shipment identifier and the driver’s one-time pickup credential. The system checks both against the latest authorization recorded for that shipment. If the driver, carrier, vehicle, destination or collection window does not match, the cargo is not released and the discrepancy is escalated through a known channel.

If all checks pass, the handover is signed as a new custody event. A cargo-level tracker then confirms when the load leaves the facility. A Dallas-bound shipment moving towards Chicago crosses a route or geofence rule, generating an alert while intervention may still be possible.

The incident report says investigators recovered the Nike merchandise after the diversion was identified.[1] A verified release process could have moved the decisive control point earlier—from recovery after the theft to refusal at the dock. No technology can guarantee prevention, but forged paperwork becomes much less useful when it cannot satisfy the digital authorization and physical-event checks required for release.

Sensefinity connects the record to the real cargo

The hard part of supply-chain blockchain is connecting digital claims to physical events. Sensefinity addresses that gap with IoT “oracles”: trackers and sensors that collect cargo location and environmental data, register it in the Sensefinity platform and can also write selected information to a logistics blockchain. Authorized partners can access the shared record without requiring direct access to one another’s internal IT systems.[3]

Our Blockchain solution provides the tamper-evident event layer. Our NB-IoT trackers locate assets on land and at sea and can issue alerts when an asset enters or leaves a configured geofence.[4]

Together, those capabilities support a stronger release and custody process:

• digitally signed shipment and pickup authorizations;

• an auditable history of instruction changes;

• verified handovers between shipper, carrier, warehouse and receiver;

• cargo-level location evidence independent of the truck or transport paperwork;

• immediate alerts for unauthorized departure, route deviation or unexpected arrival;

• a provenance record that follows the goods beyond a single logistics provider.

Trust the verified event, not the convincing document

Fictitious pickup succeeds in the gap between what a document says and what the operation can prove. Closing that gap requires more than checking logos, signatures and reference numbers. It requires a shared source of truth, a controlled handover and independent evidence from the cargo itself.

The $647,420 Nike recovery had a positive outcome. The more important objective for the next shipment is to make forged paperwork fail before the doors close and the truck leaves.

Talk to Sensefinity about combining IoT visibility, geofencing and blockchain-backed chain of custody for high-value cargo.

Sources

1. $647K Nike cargo recovered near Chicago after fraudulent Memphis pickup — FreightWaves

2. How Nike insiders were charged in a lucrative sneaker theft conspiracy — Los Angeles Times

3. Sensefinity Logistics Blockchain

4. Sensefinity NB-IoT Trackers

EPCIS in 2026: The State of Supply Chain Event Data

Supply chains have no shortage of data. They have a shortage of data that different companies can interpret in the same way.

That is the problem EPCIS was built to solve. GS1 describes EPCIS as its flagship standard for sharing the what, when, where, why and how of products and assets across organisations. EPCIS 2.0 extends that event model to sensor readings, certifications, JSON/JSON-LD, REST interfaces and GS1 Digital Link identifiers.[1]

This report assesses where the standard stands in 2026, what is driving implementation, and where projects still break down. It is a desk-research snapshot, not a vendor adoption survey. Our evidence comes from current GS1 specifications and public regulatory material.

The 2026 snapshot

Signal What the evidence says Industry implication Standard maturity EPCIS 2.0 was ratified in June 2022; its implementation guideline followed in March 2023 and the GS1 EPCIS Sandbox launched in February 2024.[1] The technical foundation is established. The main risk has moved from specification maturity to implementation discipline. Condition data EPCIS 2.0 can carry timestamped sensor data, including readings used in cold chains and industrial IoT.[1] Location events and temperature evidence can travel in one interoperable event stream. Certification data The standard supports certification details associated with products, organisations, locations, harvests and shipments.[1] Compliance evidence can be linked to the event where it matters rather than stored in an isolated document repository. API accessibility JSON/JSON-LD and REST capture/query interfaces are part of EPCIS 2.0.[1] Integration no longer has to start with XML-heavy, batch-only architecture. Regulatory pull The FDA Food Traceability Rule requires covered actors to retain Key Data Elements linked to Critical Tracking Events and provide requested information to FDA within 24 hours.[2] Regulated traceability is becoming an event-data problem, even where the law does not mandate EPCIS by name. DPP convergence The EU Digital Product Passport is being introduced progressively and will carry lifecycle, origin, material and environmental information for selected product groups.[11] Product master data and supply chain events will increasingly need a common identity layer.

Our finding: EPCIS covers five evidence layers, but governance remains outside the standard

We reviewed EPCIS 2.0 against five practical evidence layers required by modern cargo and product programmes.

Evidence layer EPCIS 2.0 coverage What still has to be designed Identity Native support through GS1 identifiers and Digital Link URI syntax.[1] Identifier ownership, granularity and partner onboarding. Business events Native event model for status, movement, transformation, aggregation and chain of custody.[1] A shared event vocabulary and rules for late or corrected events. Physical condition Native sensor-data support.[1] Device calibration, sampling frequency, alert thresholds and proof that a sensor remained attached to the cargo. Claims and certifications Native certification fields.[1] Who may issue a claim, how it expires and how it is revoked. Exchange JSON/JSON-LD and REST interfaces.[1] Access control, commercial permissions, retention and cross-company service levels.

The conclusion is useful because it separates a standards question from an operating-model question. EPCIS can express all five layers. It cannot decide which partner is trusted, how often a sensor should report, who pays for data retention or which event wins when two systems disagree.

Regulation is pushing companies toward event-level traceability

The strongest implementation pressure is no longer a generic promise of visibility. It is the need to reconstruct specific product histories quickly.

The FDA's Food Traceability Rule applies additional recordkeeping to foods on the Food Traceability List. Covered organisations must associate Key Data Elements with Critical Tracking Events and be able to provide the information to FDA within 24 hours or another agreed period.[2] GS1's own food-safety guidance maps GTIN, GLN and EPCIS event data to this need for product, location and movement records.[18]

The compliance date was originally January 20, 2026. FDA subsequently proposed a 30-month extension to July 20, 2028, and Congress directed the agency not to enforce the rule before that date.[2] That extension is preparation time, not a reason to postpone architecture. Partner identifiers, event semantics and exception workflows usually take longer than the API connection.

Europe is creating a second source of pressure. Under the Ecodesign for Sustainable Products Regulation, Digital Product Passports will be introduced through product-specific rules. The Commission lists batteries first, followed by product groups such as textiles, iron and steel, construction products and others.[8][11] A passport tells stakeholders what a product is and what must be known about it. EPCIS can supply the time-ordered operational evidence of what happened to it.

The adoption gap is not capture. It is continuity.

Most pilots can generate a shipping event. Far fewer can maintain a trustworthy history through repacking, consolidation, subcontracted transport and handover to another platform.

Four gaps appear repeatedly:

1. Identity breaks at aggregation. A pallet identifier is recorded, but the link between item, case, pallet and container is incomplete.

2. Condition data lacks business context. A temperature reading exists, but the system cannot say which shipment leg, custody holder or product lot it belongs to.

3. Partners use different event meanings. "Received" may mean arrival at the gate, unloading, quality acceptance or ERP posting.

4. Corrections are not governed. Event histories need a controlled way to handle duplicates, delayed data and amended records without erasing the audit trail.

EPCIS has the structures needed to address these problems. Implementers still need to agree on the operating rules.

A practical 90-day EPCIS readiness test

A company does not need a multi-year transformation programme to learn whether its data is ready. A useful first test follows one real shipment and asks five questions:

• Can every tracked object and logistics unit be identified consistently?

• Can the business record packing, shipping, receiving and transformation events using shared vocabulary?

• Can sensor readings be tied to the correct object, place and time?

• Can one external partner query only the events it is authorised to see?

• Can the team reconstruct the shipment history without manually joining spreadsheets?

If any answer is no, the pilot has identified a concrete interoperability gap. That is more valuable than a polished dashboard built on ambiguous data.

Where Sensefinity fits

Sensefinity already supports EPCIS supply chain events and can combine them with location, temperature and humidity data. Our NB-IoT trackers create observations from the physical journey; EPCIS gives those observations a shared business context.

The result is not simply another track-and-trace screen. It is an event history that can be exchanged with customers, suppliers and compliance systems without forcing every participant into the same application.

What to watch next

During the next implementation cycle, three developments deserve attention:

• convergence between EPCIS event histories and Digital Product Passport records;

• practical use of sensor and certification fields beyond proof-of-concept projects;

• partner governance, especially access rights, event correction and long-term availability.

The standard is ready enough. The differentiator in 2026 is whether companies can keep identity, condition and custody evidence connected after cargo leaves their own system.

Methodology and limitations

This report was prepared on September 5, 2026 from public GS1, FDA and EU sources. The readiness matrix is Sensefinity's analysis of features documented in EPCIS 2.0; it is not an adoption-rate survey. We found no authoritative global count of production EPCIS 2.0 deployments and have not invented one.

Sources

[1] https://www.gs1.org/standards/epcis — EPCIS & CBV | GS1 [2] https://www.fda.gov/food/food-safety-modernization-act-fsma/fsma-final-rule-requirements-additional-traceability-records-certain-foods — FSMA Food Traceability Rule | FDA [8] https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32024R1781 — Regulation (EU) 2024/1781 (ESPR) [11] https://single-market-economy.ec.europa.eu/single-market/digital-product-passport_en — Digital Product Passport | European Commission [18] https://gs1.org/public-policy/leveraging-GS1-standards-to-meet-key-food-safety-challenges — Leveraging GS1 standards for food safety | GS1

Fewer Thefts, Twice the Losses: The $304.6 Million Visibility Gap

Cargo theft incidents fell in the second quarter of 2026. The financial damage did not.

Verisk CargoNet documented 677 supply-chain theft incidents across the United States and Canada in Q2 2026, down 26% from the same quarter a year earlier. Yet estimated losses rose from $135.7 million to $304.6 million. Among incidents with a reported commodity value, the average reached $564,009.

This is not a contradiction. It is a warning: organized groups do not need to steal more freight when they can identify and divert the right freight.

For shippers of enterprise technology, metals, pharmaceuticals, food and other high-value goods, the new risk is concentration. A single compromised shipment can now erase the benefit of dozens of successful deliveries.

The shipment has become the target

CargoNet’s analysis points to continued targeting of enterprise computer equipment, networking components, cryptocurrency-mining hardware and industrial metals. These goods often travel through conventional logistics networks even when the value inside a trailer reaches several million dollars.

That creates a dangerous mismatch: extraordinary cargo protected by ordinary visibility.

A transport-management system may confirm that a truck was assigned. A carrier portal may show that a delivery is in progress. Neither necessarily proves that the valuable pallet is still with the expected vehicle, following the authorized route or remaining under the correct custody.

The distinction matters because modern cargo theft is increasingly strategic. CargoNet reports that business-email compromise and shipment misdirection remain important access points. Criminals can use a compromised account to identify valuable loads, impersonate trusted parties and alter shipment instructions while appearing legitimate.

In that environment, tracking the journey only at dispatch and delivery leaves too much time—and too many handovers—unobserved.

From vehicle visibility to cargo visibility

The answer is not one more status call. It is an independent data layer attached to the asset that matters.

Sensefinity’s NB-IoT trackers and sensors are designed to locate assets on land and at sea, record environmental conditions and generate alerts when an asset enters or leaves a configured geofence. Depending on the use case, the monitored asset can be a container, a pallet, a crate or the cargo itself.

This changes the security model in four practical ways:

• Independent location: the shipment can report its own position instead of relying only on the tractor, trailer or driver’s phone.

• Geofence alerts: an unexpected departure from an authorized corridor, yard or destination can be flagged while intervention is still possible.

• Condition monitoring: temperature and humidity thresholds help reveal whether sensitive cargo remained within specification during a diversion or delay.

• Datalogging and auditability: a continuous history gives operations, insurers and investigators evidence of where the asset moved and under which conditions.

No single technology prevents every theft. Carrier verification, access controls, secure communications and trained people remain essential. Cargo-level IoT complements those controls by reducing the period in which a diverted shipment is invisible.

Why response time determines recovery

High-value cargo moves quickly after a successful theft. Loads may be transferred, split, relabelled or routed into established resale channels. The longer a shipper waits to discover a deviation, the less useful yesterday’s location becomes.

Real-time alerts turn detection into an operational event rather than a reconciliation problem. If an asset exits a planned geofence or fails to arrive where expected, the shipper can investigate immediately, preserve the movement history and share actionable information with security partners.

The goal is not simply to know that a loss occurred. It is to know soon enough to change the outcome.

Proof across every handover

Location answers “where?” High-risk supply chains must also answer “what happened?” and “who had custody?”

Sensefinity’s IoT data can also be registered in a logistics blockchain, creating a traceable record of location and environmental conditions such as temperature, humidity and shock. This combination links physical events to a shared digital history.

For insurers and cargo owners, that evidence can support faster incident reconstruction. For customers, it can strengthen confidence that a shipment followed the expected route and remained within agreed conditions. For regulated or sensitive goods, it helps replace assumptions with verifiable data.

The lesson behind $304.6 million

The Q2 figures show why incident counts alone can be misleading. Fewer thefts did not produce lower risk because criminals concentrated on more valuable cargo and more sophisticated methods.

Supply-chain security must respond in the same way: focus protection on the shipment, not only on the vehicle; detect deviations as they happen, not after delivery fails; and preserve evidence across every handover.

The next loss may not come from more crime. It may come from one carefully selected load moving through one avoidable blind spot.

Sensefinity turns containers, pallets and cargo into connected assets—helping companies see where their goods are, understand their condition and act when the journey no longer matches the plan.

Talk to Sensefinity about cargo-level visibility.

Sources

• Verisk CargoNet, “Cargo Theft Losses More Than Double to $304 Million in Q2 Despite a Drop in Thefts,” 6 August 2026: https://www.globenewswire.com/news-release/2026/08/06/3340253/0/en/cargo-theft-losses-more-than-double-to-304-million-in-q2-despite-a-drop-in-thefts-driven-by-high-value-metals-and-technology-heists.html

• Sensefinity, “NB-IoT Trackers”: https://www.sensefinity.com/nbiot-trackers

• Sensefinity, “Blockchain”: https://www.sensefinity.com/blockchain

When the target is AI hardware: how real-time visibility stops theft and speeds recovery

By Sensefinity — August 29, 2026

Over the past few weeks, cargo theft in the United States has gained a new and extremely expensive target: artificial intelligence hardware. U.S. press reports describe thieves ramming security escorts and hijacking data-center equipment shipments in California, with the so-called "bump and run" tactic costing around $111 million in data-center freight. On another front, rail cargo theft runs at roughly $200 million a year. In Germany, thefts at highway rest areas (A7, A11) and "Planenschlitzer" — tarpaulin slashers — keep the logistics sector on alert.

The pattern is clear: as the value per pallet rises, so does the sophistication of those who steal it. A single pallet of AI accelerators or servers can be worth more than an entire truckload of conventional goods. And once diverted, the equipment disappears into the grey market within hours.

The problem isn't just theft — it's invisibility

The thief's biggest advantage is the supply chain's blindness. Between leaving the warehouse and reaching its destination, cargo passes through hands, trucks, rest areas and depots where almost no one knows exactly where it is, what temperature it is at, or whether it has been opened. When something goes wrong, the first question — "where was the cargo at 3 a.m.?" — often has no answer.

This is where Sensefinity's approach comes in.

Prevent: make theft visible before it happens

Sensefinity uses NB-IoT trackers and smart sensors together with 4G/LTE-M connectivity and AI monitoring to give continuous visibility of what is being moved — containers, pallets or high-value equipment.

• Real-time location. A tracker fixed to the cargo or pallet reports position independently of the driver and the truck. If the cargo leaves the planned route, or stops at an unauthorized rest area, the system raises an immediate alert.

• Opening and tamper detection. Door, shock and motion sensors detect when a container or box is opened outside an authorized stop — often before the theft is completed.

• Geofencing and routes. The AI compares the actual route with the planned one and flags deviations, anomalous stops, or "bump and run" (where the trailer is separated from the tractor).

• Condition as evidence. Temperature, shock and humidity are recorded continuously, building forensic proof of what happened to the cargo in transit.

The deterrent effect is simple: cargo that "sees and talks" stops being an easy target.

Recover: cut response time from days to minutes

If theft occurs, response speed decides whether the equipment is recovered or resold. Sensefinity trackers stay active even after diversion:

• Post-theft tracking. As long as the tracker has cellular coverage (NB-IoT/LTE-M), location keeps being reported — often inside the buyer's vehicle or warehouse, not the owner's.

• Tamper-proof history. Movement, opening and stop events form a timeline that authorities can use to reconstruct the path of the theft.

• Response integration. Alerts can be routed to security teams or law enforcement, shrinking the time between theft and action.

In a market where AI hardware can be resold within hours, those minutes matter.

Why NB-IoT and LTE-M matter in the field

Low-power, wide-coverage technologies like NB-IoT and LTE-M are ideal for long-haul cargo: they work where ordinary phones fail (underground warehouses, rest areas, rural areas) and draw little battery, enabling weeks of monitoring without recharging. It is exactly the coverage a pallet of servers needs as it crosses a continent.

Conclusion

Theft of AI hardware is not an isolated "physical security" problem — it is an information problem. Whoever knows where the cargo is, in real time, and can prove what happened to it, holds the advantage. Sensefinity turns invisible cargo into trackable cargo, deterring theft at the source and accelerating recovery when something goes wrong.

Because in the new game of high-value logistics, to see is to protect.

Sensefinity — Internet of Cargo. AI- and Blockchain-powered supply chain visibility, from origin to destination.

The Trailer Was Found. The Servers Were Gone: Why High-Value Cargo Needs Item-Level Visibility

A stolen trailer carrying several pallets of server equipment was recovered in Germany within minutes of its original location. The GPS tracker had worked. The trailer had been found.

But the cargo — reportedly worth millions of euros — was already gone.

The incident, reported by VerkehrsRundschau on 2 September 2026, exposes a critical weakness in conventional cargo security: knowing where the vehicle or trailer is does not necessarily tell you where the goods are.

For high-value logistics, visibility must travel with the cargo itself.

What happened in Herzberg am Harz

According to the report, a semi-trailer containing several pallets of server technology was parked in the Aue industrial area of Herzberg am Harz, in Germany’s Göttingen district, on Friday afternoon. The equipment was destined for a data centre and was scheduled for delivery on Monday.

At approximately 21:50 on Sunday, 30 August, thieves allegedly connected the trailer to an unidentified tractor unit and drove it away. When the driver returned early on Monday, the trailer had disappeared.

The transport company used a GPS transmitter to locate it near a recycling facility on Kreisstraße 409. The journey appears to have taken only a few minutes. By the time the trailer was recovered, however, it was empty.

Police found a pallet truck and the trailer’s registration plate near the recovery location. Investigators believe the volume and weight of the missing server equipment may have required several people and one or more additional vehicles.

The estimated loss runs into the millions of euros.

Trailer tracking solved only half the problem

This case is not evidence that GPS tracking is useless. On the contrary, the tracker helped recover the stolen trailer quickly.

It does show the limit of tracking only the transport asset.

A trailer and its load can separate in minutes. Once pallets are transferred to another vehicle, a tracker fixed to the trailer continues to report the location of an empty metal box. The logistics operator may know where the equipment was last transported, but not where it went next.

That distinction matters whenever the cargo is more valuable than the vehicle carrying it. Servers, semiconductors, electronics, pharmaceuticals and other high-value goods require a security model centred on the shipment — not only on the truck or trailer.

Cargo-level visibility changes the response

At Sensefinity, we believe the monitoring architecture should match the value and risk profile of the cargo. For a shipment such as this one, visibility can be layered across the trailer, pallets and, where justified, individual high-value assets.

Smart trackers and sensors can provide:

• Independent pallet-level location: selected pallets remain visible after being removed from the original trailer.

• Geofence alerts: an unexpected departure from a loading area, route or approved delivery zone can trigger an immediate warning.

• Movement and handling events: sensor data can identify unauthorised movement or a transfer occurring outside the planned operation.

• Environmental monitoring: temperature, humidity and shock data can protect sensitive equipment against damage as well as theft.

• Exception-based operations: teams receive actionable alerts instead of having to watch dots on a map continuously.

The objective is not simply to collect more data. It is to shorten the interval between an unauthorised event and an operational response.

In a theft completed within minutes, that interval is decisive.

From location history to verifiable chain of custody

Location data becomes more useful when it is connected to shipment identity and custody events.

Who released the cargo? Which vehicle collected it? Was the collection expected? When did custody change? Did the pallet follow the authorised route? Was it delivered to the correct consignee?

Sensefinity combines IoT tracking with a blockchain layer that can register selected logistics events in a tamper-resistant record. Trackers and sensors act as oracles, connecting physical events to the digital ledger. Authorised partners can verify key information without requiring unrestricted access to one another’s internal systems.

For high-value cargo, this creates a stronger evidence trail across shippers, carriers, warehouses and recipients. It can support investigations, claims, compliance and dispute resolution — while making unauthorised substitutions or unexplained custody gaps easier to detect.

A Digital Product Passport can extend that verified identity throughout the asset lifecycle, connecting origin, logistics operators and other product information to a persistent digital record.

A practical layered-security model

No single device can eliminate cargo theft. Effective protection combines physical, operational and digital controls.

For high-value technology shipments, operators should consider:

1. Tracking the trailer and selected pallets independently.

2. Creating geofences around loading sites, authorised stops and delivery locations.

3. Triggering alerts for movement outside approved time windows.

4. Linking sensor identities to shipment and custody records.

5. Escalating exceptions to named responders with a documented playbook.

6. Reviewing where cargo can be transferred quickly and without observation.

7. Preserving verifiable event data for investigations and insurance claims.

The right configuration depends on shipment value, route, dwell time and threat profile. The important point is that the monitoring plan must follow the risk all the way down to the level at which the cargo can disappear.

Track what creates the value

The trailer in Herzberg was recovered. The multimillion-euro load was not.

That difference captures the challenge facing high-value logistics: asset tracking can locate the container of value without locating the value itself.

Sensefinity’s NB-IoT trackers and smart sensors help organisations monitor cargo on land and at sea, configure geofence and environmental alerts, and connect physical logistics events to trustworthy digital records.

If your organisation transports servers, electronics or other high-value goods, talk to Sensefinity about designing a cargo-level visibility and chain-of-custody strategy.

Primary source: Thomas Burgert, “Unbekannte stehlen Sattelauflieger mit teurer Servertechnik,” VerkehrsRundschau, 2 September 2026. The source reports that the theft occurred on 30 August 2026.

"The Worst I’ve Ever Seen": Cargo Theft Turns Violent in Pursuit of AI Hardware

A cargo investigator with 25 years of experience called it "the worst I’ve ever seen." The words refer to two alleged attacks on high-value technology shipments in California, reported by WIRED on August 12, 2026. In both cases, criminals appear to have targeted the security escort first. One escort vehicle was rear-ended; another was forced into a spin. Once the escorts were immobilised, the trucks continued away from their planned destinations and millions of dollars in data-centre equipment disappeared.

WIRED could not independently confirm every detail because the investigations were still active and identifying information was withheld. Two other cargo-security sources did, however, corroborate elements of the incidents. The reported method matters because it exposes a weakness in conventional transport security: if visibility depends on the vehicle, the driver or an escort, attackers only need to compromise one of them.

AI hardware has changed the economics of cargo theft

Servers, accelerators and other data-centre components concentrate enormous value in a small space. A few pallets can be worth millions and the goods can move quickly through grey markets. That combination attracts organised groups with the resources to conduct surveillance, obtain inside information, impersonate legitimate carriers and coordinate an attack.

Physical escorts remain useful, but these cases show their limits. An escort protects what it can see. It cannot continuously verify the identity of the driver, confirm that each pallet remains inside the trailer or detect that the cargo has been separated from the authorised vehicle.

The security model therefore has to follow the cargo itself.

Track the load, not only the truck

Sensefinity’s Internet of Cargo platform attaches visibility to the shipment through connected trackers and sensors placed on containers, pallets or high-value equipment. The telemetry remains independent of the truck’s onboard system and of the driver’s phone.

If an escort is forced to stop but the shipment keeps moving, the platform can still report the cargo’s position. If the tractor, trailer and pallets separate, cargo-level devices make that separation visible. This closes the blind spot that criminals exploit when they neutralise the people or systems around a load.

Detect the attack while it is happening

Prevention depends on recognising abnormal behaviour early enough to intervene. Sensefinity combines NB-IoT trackers and smart sensors with rules and alerts that can identify events such as:

• departure from an authorised route;

• an unscheduled stop or entry into a high-risk area;

• unexpected motion after a vehicle should be stationary;

• shock or impact consistent with rough handling or an attack;

• opening or tampering outside an approved location;

• separation between tracked cargo and its assigned vehicle.

A security team does not have to wait for a missed delivery or an end-of-shift phone call. It can receive the alert as the deviation happens, check the shipment and escalate to the carrier or law enforcement according to an agreed response plan.

No tracking system can guarantee that a violent attack will not occur. What it can do is remove the anonymity and time advantage on which cargo theft depends. Faster detection makes a theft harder to complete and improves the chance of recovery before the goods are broken up or resold.

Verify who had custody

The WIRED report also raises the possibility of driver involvement. That risk cannot be solved by location data alone. High-value transport needs a verifiable chain of custody: who collected the load, which vehicle was assigned, when custody changed and whether the route and handover matched the plan.

Sensefinity can record shipment events in a blockchain-backed logistics trail. This creates an auditable history that is harder to alter after an incident. Combined with authorised pickup identities and cargo-level sensor data, it helps operators detect a fictitious collection, an unauthorised handover or a route that no longer matches the shipment’s instructions.

Build security around response time

For high-value AI hardware, a useful security programme should define more than a tracker installation. It should specify:

1. which assets are tracked at trailer, pallet and item level;

2. the approved route, stops and handover points;

3. alert thresholds for route deviation, opening, shock and separation;

4. who receives each alert and how quickly they must respond;

5. how location and custody evidence is shared with investigators.

Sensefinity brings those signals into one cargo-monitoring view. Operators can see where the shipment is, whether it has followed the authorised journey and whether anything happened to it in transit. The goal is practical: identify the first sign of interference, act before the cargo disappears and preserve reliable evidence if an incident still occurs.

The attacks described by WIRED show that guarding the vehicle is no longer enough. When criminals plan around the escort, security must stay with the cargo.

Sources

• WIRED: “The Worst I’ve Ever Seen”: Cargo Thefts Have Turned Violent in Pursuit of AI Hardware

• Sensefinity Internet of Cargo

• Sensefinity NB-IoT Trackers

• Sensefinity Blockchain