Connect with us

Non classé

AI at the Edge: Why On-Device Intelligence Changes the Game for Supply Chains

Published

on

Ai At The Edge: Why On Device Intelligence Changes The Game For Supply Chains

Artificial intelligence is entering a new phase of deployment. For most of the past decade, enterprise AI systems have relied heavily on centralized cloud infrastructure. Data collected at the edge of operations, such as warehouse scans, transportation events, and shipment documentation, has typically been transmitted to centralized systems for processing and analysis.

That architecture is now beginning to shift.

Advances in semiconductor design, particularly the integration of AI accelerators into mobile and embedded processors, are enabling increasingly capable models to run directly on edge devices. Smartphones, industrial handhelds, vehicles, robotics systems, and infrastructure sensors are becoming capable of performing complex inference locally.

For supply chain and logistics operations, this shift is significant. Edge AI distributes intelligence throughout the operational network, allowing analysis and decision support to occur much closer to where events actually take place.

Moving Intelligence Closer to the Point of Action

Traditional enterprise software concentrates intelligence in centralized systems. Operational data flows into ERP platforms, warehouse management systems, transportation management systems, and planning applications where analysis occurs.

While this model remains essential, it introduces limitations. Data often arrives after events have already occurred, many operational signals never enter enterprise systems at all, and decision cycles can be slowed by communication delays.

Edge AI alters this model by placing intelligence directly within operational environments.

A warehouse worker’s handheld device can analyze a photo of damaged goods immediately. A driver’s mobile device can interpret delivery instructions or flag route risks in real time. A field technician’s device can diagnose equipment issues using image recognition and contextual guidance.

In effect, intelligence moves closer to the point of action, where it can influence operational outcomes more quickly.

Reducing Latency in Operational Decisions

One of the most practical benefits of edge AI is the reduction of latency in decision support.

Cloud based AI systems require data transmission, processing in remote infrastructure, and delivery of responses back to the user or system. Even under good network conditions, this introduces delay.

When AI models run locally, inference happens almost instantly.

For logistics operations, where timing frequently matters, this improvement can be meaningful. Warehouse workers can verify pick accuracy immediately. Drivers can receive routing guidance without relying on connectivity. Inspection processes can identify defects at the moment they are observed.

Across thousands of operational events each day, these small reductions in delay accumulate into meaningful improvements in responsiveness.

Expanding the Supply Chain’s Sensing Layer

Perhaps the most transformative aspect of edge AI is its ability to expand the supply chain’s sensing capabilities.

Modern logistics networks already rely on a range of sensing technologies such as RFID, telematics, IoT devices, and robotics sensors. Edge AI extends this sensing layer by enabling everyday devices to interpret information from the physical environment.

Images, voice interactions, documents, and environmental observations can all be converted into structured operational signals.

A driver can dictate a delivery exception that is automatically transcribed and categorized. A warehouse employee can photograph damaged packaging and have AI classify the issue. A technician can capture images of equipment components that trigger automated diagnostics.

These signals enrich operational data streams and provide a more detailed view of what is happening across the network.

Enabling AI Assisted Frontline Work

Edge AI also changes how frontline personnel interact with digital systems.

Historically, operational workers have been required to manually record events by scanning barcodes, filling out forms, and entering structured data into mobile applications. These tasks are necessary but often introduce friction into operational workflows.

AI enabled devices allow interactions to become more natural. Workers can speak to devices, capture images, or request assistance through conversational prompts. AI systems interpret these inputs and translate them into structured records for enterprise systems.

The result is less time spent on data entry and more time focused on operational tasks.

Supporting Human in the Loop Operations

Despite the growing capabilities of artificial intelligence, supply chains remain environments where human judgment is critical. The most effective deployments of AI maintain a human in the loop model where technology augments decision making rather than replacing it.

Edge AI reinforces this approach.

A planner may ask a device to summarize shipment delays across a region. A warehouse supervisor may request prioritization recommendations for inbound trailers. A driver may receive suggestions for alternative routes when disruptions occur.

In each case, AI provides analysis and recommendations while humans remain responsible for final decisions.

This balance is essential for building trust in AI enabled systems.

Creating the Conditions for Distributed Intelligence

As edge AI capabilities spread across logistics networks, a broader architectural change begins to emerge.

Instead of intelligence being concentrated solely in enterprise platforms, it becomes distributed across many nodes in the operational network. Devices used by workers, vehicles, and automated systems all participate in generating insights and responding to events.

This distributed intelligence model allows operational signals to be interpreted immediately and shared with other systems in the network. Over time, it enables more coordinated and responsive supply chain operations.

In combination with advances in AI coordination architectures, this distributed intelligence can support more adaptive logistics networks capable of responding dynamically to changing conditions.

A New Layer in the Supply Chain Technology Architecture

Edge AI does not replace existing enterprise platforms. Systems such as ERP, WMS, TMS, and planning applications remain foundational elements of supply chain infrastructure.

Instead, edge AI introduces a new layer within the technology architecture. This layer sits between physical operations and digital enterprise systems.

It captures signals from the physical world, interprets them locally using AI models, and feeds structured insights into enterprise platforms.

Over time, this architecture enables supply chains to operate with far greater situational awareness and responsiveness.

The Strategic Implication

For supply chain leaders, the rise of edge AI represents more than an incremental improvement in mobile computing. It signals a structural shift in how logistics networks perceive and respond to operational events.

As intelligence becomes embedded in the devices used by drivers, warehouse operators, technicians, and automated systems, supply chains gain a richer and more immediate understanding of what is happening across the network.

Organizations that integrate edge AI into their operational workflows will be able to detect disruptions earlier, respond faster, and operate with a higher level of situational awareness.

In an environment defined by volatility, complexity, and increasing expectations for speed and transparency, these capabilities will help define the next generation of intelligent, adaptive supply chain networks.

The post AI at the Edge: Why On-Device Intelligence Changes the Game for Supply Chains appeared first on Logistics Viewpoints.

Continue Reading

Non classé

Electric Aircraft Pilot Program Opens a New Logistics Frontier

Published

on

By

Electric Aircraft Pilot Program Opens A New Logistics Frontier

The U.S. Department of Transportation and the Federal Aviation Administration recently selected eight pilot projects to begin limited cargo and passenger operations using electric aircraft under the eVTOL Integration Pilot Program (eIPP). Most headlines will focus on the promise of air taxis. Supply chain leaders should focus on something else.

This initiative represents the early formation of a new logistics layer.

Several of the selected projects will begin with freight-oriented missions. According to New York, Florida among the states selected for electric aircraft pilot program, Florida’s program will start with cargo delivery before expanding to medical response and eventually passenger operations.

Other pilots focus on logistics use cases such as offshore support flights, regional cargo transport, and emergency response mobility.

These are not speculative concepts. They are operational logistics pilots.

Logistics Often Leads Transportation Innovation

Historically, freight networks have been the proving ground for new transportation technologies. Passenger mobility systems usually follow after reliability, economics, and operational models have been validated.

Cargo networks tolerate early-stage technology better than consumer transportation. Routes are predictable, operational windows are flexible, and the value of time-sensitive delivery is easier to quantify.

The new pilot program reflects that reality. As reported in DOT announces projects selected for U.S. eVTOL Integration Pilot Program, the program will test cargo logistics networks, medical transport missions, regional passenger service, and autonomous aircraft operations.

Freight operations are likely to be the first area where the economics begin to make sense.

Filling the Gap Between Trucks and Aircraft

Electric aircraft and eVTOL systems occupy a potential middle ground between trucking and traditional aviation.

Trucks offer flexibility but are constrained by congestion, distance, and road infrastructure. Conventional air freight provides speed but requires airports, runways, and higher operating costs.

Electric aircraft introduce the possibility of short-range aerial logistics corridors connecting regional distribution points without the infrastructure requirements of commercial aviation.

For supply chains, this creates a new operational option: rapid regional freight movements for urgent inventory, medical supplies, and high-value shipments.

Instead of replacing trucks or cargo aircraft, these systems could fill a transportation gap that has historically been inefficient.

Florida as an Early Testbed

Florida’s inclusion in the pilot program is particularly notable.

The state’s geography and dispersed metropolitan pattern make it well suited for regional aerial logistics. Medical supply delivery, disaster response logistics, and time-critical freight movements are all realistic early use cases.

The program also aligns with Florida’s broader aerospace ecosystem anchored by Kennedy Space Center, aviation manufacturing activity, and established aerospace testing infrastructure.

If these pilots demonstrate operational value, Florida could become an early example of how advanced air mobility integrates into logistics networks.

Certification and Infrastructure Still Ahead

Despite the momentum around advanced air mobility, several major hurdles remain.

None of the participating aircraft manufacturers have yet received full FAA type certification. Infrastructure such as charging networks, vertiports, maintenance capabilities, and air traffic integration must also mature before large-scale deployment becomes feasible.

Pilot programs like eIPP are designed to generate the operational data regulators and operators need to build those frameworks.

What This Means for Supply Chain Strategy

Modern supply chains are increasingly multi-modal and digitally orchestrated. AI-driven planning systems already evaluate routing options across truck, rail, ocean, and traditional air freight.

Electric aircraft introduce another potential variable in that optimization process.

If these systems become operationally viable, future logistics platforms will evaluate them alongside existing transport modes when planning urgent shipments or responding to disruptions.

Over time, electric aircraft could become part of a broader shift toward adaptive, AI-enabled logistics networks, where transportation decisions are dynamically optimized across multiple infrastructure layers.

A Small Step Toward a Different Logistics Network

The eIPP program should be viewed as infrastructure experimentation rather than immediate transformation.

Initial deployments will remain limited and focused on specialized missions such as medical logistics, offshore support, and high-value cargo routes.

But transportation networks evolve gradually. Pilots establish operational viability, standards follow, and infrastructure develops around proven use cases.

Electric aircraft have now entered that early operational phase.

For supply chain leaders, the message is simple: a potential new transport layer is beginning to move from concept into real-world deployment.

The post Electric Aircraft Pilot Program Opens a New Logistics Frontier appeared first on Logistics Viewpoints.

Continue Reading

Non classé

Air rates continue to climb on Mid-East closures – March 10, 2026 Update

Published

on

By

Air rates continue to climb on Mid-East closures – March 10, 2026 Update

Discover Freightos Enterprise

Published: March 10, 2026

Blog

Weekly highlights

Ocean rates – Freightos Baltic Index

Asia-US West Coast prices (FBX01 Weekly) increased 10%.

Asia-US East Coast prices (FBX03 Weekly) decreased 9%.

Asia-N. Europe prices (FBX11 Weekly) increased 6%.

Asia-Mediterranean prices (FBX13 Weekly) increased 2%.

Air rates – Freightos Air Index

China – N. America weekly prices increased 11%.

China – N. Europe weekly prices increased 2%.

N. Europe – N. America weekly prices decreased 9%.

Analysis

Very few vessels have passed through the Strait of Hormuz since the start of the war in Iran a week and a half ago. The closure is of global concern due to its stifling of oil tanker movements which is already slowing production and pushing oil prices up.

The US International Development Finance Corporation announced a plan to insure vessel transits despite the security risk, though few carriers will likely take up the offer without naval protection as well. And the government is much more likely to devote these resources to oil tankers than to container flows.

For the container market, disruptions from the strait’s closure are so far limited to containers already en route to (or stuck in) the Gulf ports, with some knock-on congestion elsewhere. Ports in countries like India and Bangladesh – significant exporters to the Gulf states – are reporting backlogs. And yard utilization levels are increasing at transhipment hubs in the Far East where some Gulf-bound containers are now being diverted. Yard density at those ports could increase somewhat in the coming days as shippers who so far decided to wait and see may choose to divert Gulf-bound containers there as the Strait remains closed.

Carriers are rolling out contingency plans for Gulf volumes via alternative ports in the region, with shipments then moving on to their destinations by road. Carriers are applying significant surcharges for those containers already in transit on these lanes, and offering alternatives like storage, return to origin and change of destination, which will also incur additional costs.

For the broader container market though, Iran’s closure of the Strait of Hormuz has not caused disruptions or price increases yet. And while congestion at Far East tranship hubs could cause delays for non-Gulf volumes in the near term, these backlogs should be much less significant than those seen at the outset of the Red Sea crisis, since the volumes involved are much lower and since carriers have already suspended new bookings to those destinations.

But one way global container shipping could be impacted is by rising fuel costs. The climbing price of oil has already led some carriers like CMA CGM and Hapag-Lloyd to announce emergency fuel surcharges for all lanes at $70-75/TEU for regional transits and $150/TEU for long haul voyages starting March 23rd, with others announcing fuel increases only on certain lanes for now. If oil prices remain elevated, more carriers are likely to introduce emergency surcharges, though standard fuel rates – adjusted on a quarterly basis and already set for Q2 – could only increase to start Q3.

Freightos Baltic Index transpacific container rates increased about $200 or 10% to $2,022/FEU last week, with daily rates to the East Coast showing a similar climb to about $3,000/FEU so far this week. Asia-Europe rates increased 6% last week to about $2,600/FEU and 2% to $3,700/FEU to the Mediterranean, with prices continuing to climb so far this week.

These rate increases are not likely related to the war in Iran, and instead reflect attempts by carriers to bump prices up now that we’re in the post-Chinese New Year period when it is not uncommon for demand to increase for a couple weeks. Emergency fuel surcharges, though, could start showing up as a component of rate levels once implemented.

In air cargo, a more broad range of lanes have been impacted by airspace closures in the Gulf, with rates climbing significantly on a list of lanes.

Gulf carriers Qatar Airways and Emirates Skycargo are two of the top three largest cargo carriers by capacity, and together with Etihad make up about 13% of global capacity. Their hubs serve as a major east-west connection point, including providing a significant share of South Asia and South East Asia capacity to Europe and N. America.

As a result Freightos Air Index data show air cargo rates have increased by about 50% since the start of the war from South Asia to N. America and Europe, with rates now at about $6.00/kg and $4.00/kg respectively. SEA – Europe prices are up 20% to more than $4.00/kg. China – US rates have climbed 20% to more than $7.00/kg too, though this increase may be mostly due to post-LNY demand. That these disruptions are coinciding with the post-LNY rush could be adding some upward pressure to ex-China rates as volumes that normally would go via the Gulf compete for long haul space.

Most Gulf airports closed completely for several days at the start of the war, but over the last few days some started to reopen partially. The UAE has reportedly opened safe air corridors enabling 48 flights per hour to depart. Emirates Airways reports it is now operating a reduced but stable flight schedule, and are now estimated to be flying more than 50% of their scheduled flights. Etihad has also resumed some flights, though Qatar Airways Cargo operations through Doha remain suspended. With these flights and capacity returning to the cargo market, we may see pressure on rates ease somewhat in the coming days. Like in ocean, some forwarders are working to fly cargo with final destinations in the Gulf to alternate airports in the region, especially Saudi Arabia, and move goods on by road.

In trade war developments, following the Supreme Court’s decision invalidating President Trump’s IEEPA-based tariffs, the US Court of International Trade ordered the government to start refunding the billions of dollars in IEEPA tariffs paid over the past year.

Some experts were surprised by the speed at which this ruling was issued, and that it relies on a single case to resolve refunds for everyone even as more than 2,000 individual suits had been filed. Customs and Border Protection responded that these hundreds of thousands of payments can’t be made immediately and asked for 45 days to set up an automated system. There are still a lot of questions as to when refunds will actually start going out and who will receive them, but these developments are a big step forward.

In the meantime, the administration has put 10% global tariffs into place relying on Section 122 – even as two dozen states are challenging the administration’s use of this law – and the USTR says Section 301 investigations that will be used as the basis for country-specific tariffs replacing EEPA duties are underway and will be concluded before the Section 122 tariffs expire in late July. The president has said that an additional executive order will introduce Section 122 tariffs at 15% for some countries, though so far no such order has been issued.

The current Section 122 tariffs mean lower-than-IEEPA duty levels for some companies, leading to reports of some companies starting to frontload before the July deadline, while others are not taking action just yet. The latest National Retail Federation US ocean import volume projections through June are about even with those from just before the SCOTUS decision, suggesting that most companies are remaining cautious, and not pulling volumes forward. The report expects H1 volumes to be 2.5% lower than in 2025.

Discover Freightos Enterprise

Freightos Terminal: Real-time pricing dashboards to benchmark rates and track market trends.

Procure: Streamlined procurement and cost savings with digital rate management and automated workflows.

Rate, Book, & Manage: Real-time rate comparison, instant booking, and easy tracking at every shipment stage.

Judah Levine

Head of Research, Freightos Group

Judah is an experienced market research manager, using data-driven analytics to deliver market-based insights. Judah produces the Freightos Group’s FBX Weekly Freight Update and other research on what’s happening in the industry from shipper behaviors to the latest in logistics technology and digitization.

Put the Data in Data-Backed Decision Making

Freightos Terminal helps tens of thousands of freight pros stay informed across all their ports and lanes

The post Air rates continue to climb on Mid-East closures – March 10, 2026 Update appeared first on Freightos.

Continue Reading

Non classé

Anthropic and the Pentagon: A New Debate Over AI Supply Chain Risk

Published

on

By

Anthropic And The Pentagon: A New Debate Over Ai Supply Chain Risk

Artificial intelligence is moving rapidly from the research frontier into the operational backbone of modern organizations. As this transition accelerates, governments are beginning to examine AI through a new lens. The question is no longer simply what these systems can do. The question is how resilient the infrastructure behind them really is.

The recent dispute between AI developer Anthropic and the U.S. Department of Defense illustrates how quickly this shift is unfolding. The company has challenged a Pentagon assessment suggesting that elements of the AI technology stack could present supply chain risks for government users. While the details of the classification remain technical, the broader issue is clear. Federal agencies are beginning to evaluate AI systems in the same way they evaluate other strategic technologies.

That change in perspective is significant. It signals that artificial intelligence is no longer viewed solely as software innovation. It is increasingly treated as infrastructure.

For supply chain leaders, that distinction matters.

Artificial Intelligence as a Technology Stack

The modern AI ecosystem is built on a technology stack that resembles a complex industrial supply chain more than a traditional software market.

At its foundation sits semiconductor manufacturing. Training advanced AI models requires specialized accelerators and high performance graphics processors produced by a relatively small group of global suppliers. Many of these chips depend on fabrication capacity concentrated in a limited number of advanced facilities.

Above that hardware layer sits hyperscale compute infrastructure. AI training and deployment rely on enormous data center clusters that require high bandwidth networking, specialized cooling systems, and increasingly large amounts of electrical power. These environments are operated primarily by large cloud platforms that provide the computational backbone for model development and deployment.

The next layer involves the organizations building the models themselves. These firms operate complex research and engineering pipelines that rely on extensive datasets, software frameworks, and global collaboration networks.

Once developed, the models move into the application layer where they are integrated into enterprise systems, industrial platforms, logistics networks, and national security tools.

This layered structure is precisely why governments have begun to analyze artificial intelligence as an infrastructure ecosystem rather than as a single technology product.

Why Governments Are Examining AI Supply Chains

From the perspective of defense planners, the rationale is straightforward. AI capabilities are increasingly used to support activities such as logistics planning, intelligence analysis, cyber defense, and operational decision support.

When these capabilities become embedded in mission critical systems, the resilience of the infrastructure supporting them becomes a strategic concern.

In practice, that means evaluating the same types of supply chain questions that arise in other critical industries. Where are the key components produced? How concentrated are the suppliers providing essential inputs? What geographic dependencies exist across the infrastructure stack? And how vulnerable might those dependencies be to disruption, whether from geopolitical tensions, export controls, or industrial bottlenecks?

These are not new questions for the supply chain community. What is new is that they are now being applied to artificial intelligence.

Anthropic’s Perspective on the Issue

Anthropic’s response to the Pentagon’s position reflects a different interpretation of the same risk.

The company has argued that characterizing AI systems as supply chain vulnerabilities may misrepresent how the technology actually operates. Modern models often run on distributed cloud infrastructure that provides redundancy and geographic diversity.

From that perspective, the resilience of AI capabilities should be evaluated at the level of the broader infrastructure platform rather than at the level of individual model developers.

The disagreement highlights an emerging policy challenge. Artificial intelligence systems are built on deeply interconnected technology layers that span multiple industries and geographies. Evaluating risk within that environment requires governments to understand the full ecosystem, not just the organizations producing the models.

The Structural Issue: Infrastructure Concentration

For observers of technology supply chains, the deeper issue may lie elsewhere.

The global AI ecosystem currently depends on a relatively small number of critical infrastructure providers. Advanced semiconductors are produced by a limited group of manufacturers, and large scale training environments rely heavily on hyperscale cloud platforms.

This concentration is not unique to artificial intelligence. Similar patterns exist in sectors such as aerospace, telecommunications, and energy infrastructure.

What makes the situation different is the speed with which AI capabilities are expanding. As adoption accelerates across industries, the infrastructure supporting these systems becomes more strategically important.

Artificial Intelligence as an Operational Layer

Artificial intelligence is increasingly functioning as a decision layer across enterprise operations.

In supply chain environments, these systems already support activities such as demand forecasting, transportation routing, inventory balancing, and risk monitoring. As these capabilities mature, they are evolving into intelligence layers that connect planning, execution, and exception management across logistics networks.

Research in this area has emphasized that the next generation of supply chain systems will rely on interconnected intelligence frameworks capable of coordinating information across networks of suppliers, logistics providers, and enterprise platforms. AI in the Supply Chain-sp

When that intelligence layer becomes critical to operations, the reliability of the infrastructure supporting it becomes a strategic issue.

A Preview of Future AI Governance

The current dispute between Anthropic and the Pentagon is likely a preview of broader developments.

Governments around the world are beginning to treat AI infrastructure in much the same way they treat other critical technology sectors. This process will likely involve greater transparency around infrastructure dependencies, closer examination of semiconductor supply chains, and more structured approaches to evaluating platform resilience.

For organizations deploying AI capabilities, the implications are clear. Adopting these systems means connecting operations to a global infrastructure network that includes specialized hardware, large scale compute environments, and complex software ecosystems.

As adoption accelerates, the conversation will increasingly shift from capability to resilience.

The Bottom Line

Artificial intelligence is entering the same phase that many industrial technologies eventually reach. Once a capability becomes central to economic and national systems, attention inevitably turns to the reliability of the supply chains supporting it.

The dispute between Anthropic and the Pentagon illustrates that this transition has already begun.

The next phase of AI adoption will not be defined solely by model capability.

It will be defined by the resilience of the infrastructure that makes those capabilities possible.

The post Anthropic and the Pentagon: A New Debate Over AI Supply Chain Risk appeared first on Logistics Viewpoints.

Continue Reading

Trending