Non classé

The AI Trade Just Ran Into Its Unit-Economics Problem

Published

4 semaines ago

9 septembre 2026

The artificial intelligence debate is changing.

For several years, the central question was whether AI worked. Then it became whether the technology could scale.

Increasingly, the question is becoming more difficult:

Who actually captures the value?

That distinction matters because AI adoption can accelerate, token consumption can soar and the economics of individual participants can still deteriorate.

This is the point at which a technology story becomes an operating-model story.

And for logistics companies deciding where AI creates economic advantage, that may be much more useful than watching the daily valuation of another AI stock.

More Tokens Does Not Automatically Mean More Revenue

The basic AI economic equation has generally seemed straightforward.

Models become better. Better models create more useful applications. Applications increase demand. Demand consumes more tokens. More tokens create more revenue.

The complication is price.

AI inference is becoming cheaper at a remarkable rate. Competition among model providers is increasing, open-weight models are improving, specialized models are proliferating and hardware efficiency continues to advance.

That is excellent for customers.

It is less obviously excellent for every company selling tokens.

Man Group’s bearish analysis of the AI investment cycle describes the problem starkly: if token prices decline faster than inference demand expands, enormous growth in usage does not necessarily produce the revenue required to justify an equally enormous infrastructure buildout.

Goldman Sachs Research is more constructive. Its work points to sharply increasing token consumption – potentially a 24-fold increase by 2030 as agentic AI expands – while declining unit compute costs could improve hyperscaler margins.

Those positions are not actually contradictory.

They describe the variable that matters.

The economic outcome depends on the relationship among volume, price and cost.

That is unit economics.

The De-Rating Is Telling Us Something

The financial market has already become more discriminating about AI exposure.

A recent Goldman basket analysis reportedly showed an all-inclusive AI pair roughly 46 percent below its previous highs, even as underlying demand for compute remains strong. Goldman’s argument is not that AI is ending; rather, the opportunity may be broadening toward businesses with clearer monetization, embedded workflows and defensible economic positions.

That distinction is important.

A broad technology transition does not guarantee that every participant in the technology stack earns extraordinary returns.

The internet changed the world. Many internet companies disappeared.

Containerization transformed global trade. That did not mean every shipping line earned exceptional margins.

Cloud computing became foundational infrastructure. The economics nevertheless concentrated in particular layers of the stack.

AI is likely to behave similarly.

The technology can be revolutionary while the value migrates.

Open Models Are Accelerating the Pressure

The open-model transition makes this more visible.

Vercel’s AI Gateway data recently showed open-weight models reaching approximately 62 percent of token volume on August 22, up from 28.4 percent roughly two months earlier and about 11 percent in April. These are figures from one platform rather than the entire AI industry, but the speed of the shift is notable.

Customers are learning something logistics managers learned long ago.

Not every movement requires premium service.

You do not ship every load by air. You do not give every SKU the same inventory policy. You do not assign the most expensive resource to every task simply because it is technically capable of performing it.

The same logic is beginning to apply to AI.

Enterprises can route difficult tasks to expensive frontier models while sending repetitive, lower-value or highly specialized workloads to cheaper models.

This is good architecture.

It is also price pressure.

As switching becomes easier, the model itself can become one component within a larger decision architecture.

That changes where the economic moat resides.

The Value May Move Up the Stack

Consider a logistics company using AI to manage exceptions.

The model may analyze late shipments, weather, inventory availability, customer commitments and transportation alternatives. It may recommend that an order be reallocated, a shipment expedited or a customer promise changed.

Suppose the model call costs 20 cents today and two cents three years from now.

The model provider has experienced severe unit-price compression.

The logistics operator has not necessarily lost anything.

In fact, the logistics operator may gain.

If the AI avoids a $5,000 expedite, prevents a stockout or allows one planner to manage twice as many exceptions, the economic value exists primarily in the operating outcome – not in the token.

This is where the AI economics discussion becomes particularly relevant to enterprise logistics.

Falling model prices could transfer economic value away from model providers and toward companies capable of embedding inexpensive intelligence into valuable workflows.

The model becomes cheaper.

The decision becomes more valuable.

Architecture Becomes the Moat

This suggests that enterprises should be careful about defining an “AI strategy” around access to a particular model.

Model leadership can change.

Prices can change even faster.

An enterprise architecture built tightly around a single provider may eventually look like a transportation network designed around one carrier regardless of lane, service requirement or price.

The more durable architecture is likely to separate the business problem from the intelligence resource used to solve it.

Understand the decision.

Assemble the required context.

Define the constraints.

Determine the acceptable action.

Then route the problem to the model – or combination of models – capable of solving it at the appropriate cost.

That is not simply AI adoption.

It is AI orchestration.

And it looks very similar to other logistics optimization problems.

Intelligence Is Becoming a Variable Cost

The larger economic change may be that machine intelligence is becoming a purchasable input whose price declines rapidly.

That is extraordinary.

For most of industrial history, intelligence has been expensive. Analytical capacity was constrained by the number of skilled people available to perform the work.

AI begins to relax that constraint.

If the cost of a useful unit of machine reasoning continues falling, companies can economically apply intelligence to thousands of decisions that previously could not justify human analysis.

Shipment prioritization.

Carrier selection.

Inventory rebalancing.

Appointment scheduling.

Exception resolution.

Warehouse labor planning.

Supplier-risk analysis.

Customer-response generation.

Every one of these can potentially consume enormous quantities of tokens while producing economic value far larger than the cost of those tokens.

This is why collapsing AI prices are not necessarily bearish for AI adoption.

They may be extraordinarily bullish for the users of AI.

The Second Half of the AI Trade

The first phase of the AI boom rewarded scarcity.

There were too few advanced GPUs. Too little compute. Too few frontier models. Too little capacity.

Scarcity created pricing power.

The next phase may reward orchestration.

Models will proliferate. Intelligence will become cheaper. Enterprises will learn to switch among providers. Open-weight systems will handle an increasing share of routine workloads. Infrastructure costs will continue to matter, but customers will become much more disciplined about what they are willing to pay for a unit of intelligence.

That does not mean the AI boom is over.

It means the economics are maturing.

The winners in logistics will probably not be the companies that consume the most AI.

They will be the companies that turn increasingly inexpensive intelligence into better operating decisions.

That is a very different metric.

Tokens are an input.

The decision is the product.

And the economic value ultimately belongs to whoever can make that decision improve the system.

The post The AI Trade Just Ran Into Its Unit-Economics Problem appeared first on Logistics Viewpoints.

Non classé

The New Economics of Logistics Visibility

Published

16 heures ago

5 octobre 2026

Knowing that a shipment will arrive six hours late is information. Knowing it early enough to reschedule labor, protect a customer commitment, avoid detention, or change an inventory decision is economic value.

That distinction is becoming central to the logistics visibility market. The earlier argument that exceptions are becoming the real unit of work explains why visibility economics depend less on event volume than on whether the organization can convert important events into timely resolution.

The first era of visibility was largely about answering a basic question: Where is my shipment? The next era is about a harder question: What should I do because its state has changed?

Visibility Is Not the Outcome

Location and status data can be valuable, but they are intermediate products. A business does not earn a return because a dot moved across a map more accurately. The return appears when information changes an operational decision. A useful way to think about visibility is as a chain: signal -> interpretation -> decision -> intervention -> economic outcome. If any link is missing, much of the potential value disappears.

A Signal Has to Arrive Inside the Decision Window

Timing matters.

A delay discovered after the customer has already missed production is history. The same delay identified early enough to expedite an alternate shipment may be actionable. An ETA update received after warehouse labor has reported for a shift may have less value than the same update received while the schedule can still be changed.

This means visibility quality is not only about accuracy. It is about whether the signal arrives with enough lead time to support an intervention.

Not Every Exception Deserves Attention

As visibility improves, organizations often discover a new problem: too many exceptions. A network with thousands of shipments will always contain delays, deviations, missed scans, changing ETAs, and incomplete data. If every deviation creates an alert, planners become the bottleneck. The more important capability is prioritization.

Which late shipment threatens a high-value order? Which delay creates a stockout? Which container risks demurrage? Which arrival change will disrupt a dock schedule? Which event is likely to self-correct without intervention?

Visibility becomes intelligence when the system can distinguish operational consequence from mere deviation.

ETA Is a Decision Input

Estimated time of arrival is a good example of how the economics are changing. ETA was once primarily a customer-service or tracking metric. Increasingly it can influence warehouse scheduling, yard planning, labor, inventory, customer promises, and downstream transportation. That makes ETA a shared operating variable.

The value increases when the prediction is connected to the systems that can respond. A changing ETA that remains trapped in a visibility dashboard creates less value than one that can trigger a workflow or decision elsewhere.

Dwell, Detention, and Demurrage Make the Economics Visible

Some visibility use cases have direct financial consequences. Better awareness of arrival, dwell, free-time windows, and container status can help organizations manage detention and demurrage exposure. Yard visibility can reduce unnecessary trailer search and moves. Earlier exception detection can protect delivery appointments and reduce costly service recovery. These cases make an important point: visibility value is often realized outside the visibility platform itself.

More Visibility Can Increase Work

This is the uncomfortable side of digital transparency. If a company exposes ten times as many events but does not improve prioritization or workflow, it may create ten times as many things for people to inspect. The result can be an expensive monitoring layer sitting on top of the same manual decision process.

That is why visibility and autonomous exception management are converging. The system must increasingly help decide which events require action, assemble context, recommend a response, and automate routine resolution where appropriate.

Measure Intervention, Not Just Coverage

Visibility programs are often measured by tracking coverage, data completeness, ETA accuracy, or number of connected carriers. Those are necessary operating metrics, but they do not fully describe business value.

Organizations should also ask: How many material exceptions were identified early enough to act? How quickly were they resolved? How often did intervention protect service or avoid cost? How many alerts required no useful action? How much planner time was consumed per exception?

Those measures connect visibility to economics.

The Market Is Moving Toward Action

This shift has strategic implications for technology providers. Pure visibility is becoming less differentiated as location and event data become more widely available. The higher-value layer is interpretation and action: understanding what an event means to a specific operation and helping execute the appropriate response.

That pushes visibility platforms toward orchestration, workflow, decision intelligence, and AI. It also pushes TMS, WMS, and other execution systems toward richer external event awareness.

The Bottleneck Moves

For years, logistics organizations complained that they could not make better decisions because they could not see what was happening. Increasingly, they can see more.

The bottleneck is moving.

When a network can identify exceptions continuously, the constraint becomes the speed and quality with which the organization can interpret and resolve them. That is precisely the environment in which AI agents become interesting—not because logistics needs another conversational interface, but because it needs more capacity to do operational work.

Related Logistics Viewpoints research

The New Architecture of Logistics
Systems Engineering in Logistics
2026 Autonomous Exception Management Market Map
The Economics of Decision Latency
Previous in this series: Transportation Is Becoming Computational

Request The New Architecture of Logistics Client Edition

If your organization is assessing connected execution, orchestration, AI, observability, decision velocity, or selective autonomy, I would be glad to provide the complete client edition and discuss the implications for your logistics operating model and technology architecture.

Request the client edition

The post The New Economics of Logistics Visibility appeared first on Logistics Viewpoints.

Non classé

o9 Solutions Uses a Knowledge Graph to Connect Supply Chain Decisions

Published

16 heures ago

5 octobre 2026

Supply chain decision-making is difficult partly because the relevant information is distributed across products, locations, suppliers, orders, capacities, policies, and external events. A system may have access to all of those records and still struggle to understand the relationships among them quickly enough to support a consequential decision.

o9 Solutions addresses that problem through its Digital Brain architecture, including an enterprise knowledge graph and in-memory modeling designed to connect demand, supply, inventory, planning, and operating context. That semantic layer is important because it gives analytics and AI a structured representation of how supply chain entities relate to one another rather than treating the environment as a collection of independent tables and documents.

The company combines this architecture with integrated planning, optimization, scenario modeling, machine learning, and human oversight. The strategic direction is toward a continuous decision environment where changes can be interpreted quickly, alternatives can be modeled, and recommendations can be traced back to the assumptions, events, and constraints that produced them.

As with any broad planning and intelligence platform, the value depends on implementation quality. Knowledge models need strong data governance, entity resolution, process ownership, and clear decision rights. A sophisticated model of the supply chain is useful only if the organization can keep it current and use it consistently in real operating workflows.

o9 Solutions appears in the Logistics Viewpoints Supply Chain Decision Intelligence MarketMap and Autonomous Exception Management MarketMap. The two MarketMaps highlight the relationship between integrated decision intelligence and the faster exception-response capabilities now developing around it.

The post o9 Solutions Uses a Knowledge Graph to Connect Supply Chain Decisions appeared first on Logistics Viewpoints.

Non classé

AI Will Not Transform the Supply Chain Until the Architecture Around It Catches Up

Published

19 heures ago

5 octobre 2026

Executive thesis. The model is becoming the least durable layer of the AI stack. Sustainable advantage will come from the operating architecture around AI: authoritative context, governed tools, permissions, observability, and connection to enterprise workflows.

The model is only one component

Supply chain AI discussions often begin with model capability: prediction accuracy, reasoning quality, computer vision performance, or the fluency of a generative system. Those capabilities matter, but operational value depends on everything around the model. The system still needs authoritative context, enterprise tools, permissions, workflow, observability, and a reliable path from recommendation to action.

Different AI patterns solve different problems

Prediction, optimization, generative AI, vision, and agents should not be treated as interchangeable technologies. Forecasting demand, selecting a route, extracting information from a document, interpreting an image, and executing a multi-step workflow require different evidence, control, and performance measures. A mature architecture starts with the decision or task and chooses the AI pattern that fits it.

Context is the operating fuel

An AI system can produce a plausible answer while using stale or incomplete operational context. In logistics, that can be dangerous because the truth may reside across orders, inventory, rates, carrier status, warehouse state, supplier records, and policy documents. Retrieval, master data, identity resolution, and system access therefore become part of the AI architecture, not secondary data-engineering concerns.

Tool access turns intelligence into consequence

The moment an AI system can create a shipment, change an order, contact a carrier, release inventory, or approve an exception, governance becomes an operational requirement. Tool permissions, financial limits, approval gates, idempotency, retries, and rollback are the mechanisms that separate an interesting demonstration from a dependable production workflow.

Measure workflow performance

A fluent response is not the right success metric for operational AI. Supply chain leaders should measure decision latency, manual context gathering, exception closure, override behavior, error recovery, tool failure, and the business outcome being improved. That measurement discipline also creates a rational basis for expanding autonomy as evidence accumulates.

The Logistics Viewpoints AI in Logistics: Use Cases, Architecture, and Implementation Guide separates the major AI patterns and connects them to data, tools, governance, workflow, observability, and ROI—the architecture required to move from model capability to operating value.

Executive implication

AI strategy should separate model selection from control architecture and measure value through workflow performance, decision quality, and operational outcomes.

Go deeper: provides the durable buyer, architecture, and implementation reference for this topic. AI & Advanced Analytics connects this analysis to the broader Logistics Viewpoints research architecture.