Ford Motor Company is building a smaller vehicle lineup around products customers feel strongly about and products on which the company believes it can earn durable returns. That means more emphasis on the F-150, Bronco, Mustang, Maverick, Explorer, Expedition, and commercial vehicles, and less interest in maintaining a broad lineup of sedans, hatchbacks, compact crossovers, and other vehicles that compete largely on price.

Ford executives have described this as the company’s most “passionate” lineup. That works as a marketing description, but the more important story is operational. Ford is not simply deciding which vehicles it wants to sell. It is reshaping the company around fewer platforms, lower complexity, more profitable customization, and a different manufacturing model for its next generation of electric vehicles.

Fewer Vehicles, Better Economics

For decades, major automakers tried to participate in nearly every meaningful vehicle segment. A broad portfolio helped dealers serve first-time buyers, commuters, families, enthusiasts, commercial customers, and luxury consumers. It also created enormous complexity.

Every additional vehicle program required engineering resources, tooling, supplier capacity, regulatory approval, service parts, marketing support, and dealer training. Every powertrain, trim level, electronics package, and option combination added another layer.

That complexity could be justified when a vehicle generated strong volume and acceptable margins. It became much harder to defend when a model needed substantial discounts and incentives to remain competitive. Ford concluded that several mainstream vehicles did not produce attractive enough returns. Models such as the Focus, Fusion, Edge, and Escape gave Ford broad market coverage, but they operated in crowded segments where differentiation was difficult and pricing power was limited.

CEO Jim Farley has pushed Ford in a different direction. The company is concentrating on categories where it has a stronger identity and a more defensible position: trucks, SUVs, off-road vehicles, performance cars, and commercial vehicles. The sales results suggest that the strategy has traction. Ford’s U.S. sales reached their highest level in six years in 2025, with the F-Series remaining the foundation of the business and the Bronco, Maverick, and Mustang reinforcing the strength of distinctive, brand-driven products.

This is not a strategy designed to maximize the number of vehicles Ford sells. It is designed to improve the return on every dollar the company invests.

A narrower product portfolio can create benefits far beyond marketing. Fewer vehicle programs can reduce tooling requirements, consolidate purchasing volume, simplify production planning, and lower the amount of service inventory required over the life of a vehicle. It can also reduce demand fragmentation.

An automaker with a large number of models, trims, engines, and option packages must forecast demand across thousands of possible configurations. When those forecasts are wrong, the result is excess inventory, dealer discounting, emergency schedule changes, and obsolete parts. A more concentrated portfolio allows Ford to focus volume around fewer platforms and component families, potentially improving purchasing leverage, capacity utilization, and forecast accuracy.

The F-Series is the clearest example. Its scale allows Ford to spread engineering, manufacturing, and supplier investments across hundreds of thousands of vehicles. That is difficult to replicate with lower-volume products in highly contested segments.

Simplification, however, comes with a trade-off. When a company depends more heavily on a smaller number of profitable vehicle families, disruptions affecting those products become more consequential. A supplier failure, labor disruption, quality problem, or component shortage involving the F-Series or another core platform can have an outsized financial effect.

Ford may be reducing portfolio complexity, but it is increasing the importance of resilience around the products that remain. Supplier visibility, dual sourcing, quality control, capacity planning, and contingency management become even more important.

Customization Becomes Part of the Operating Model

The most interesting element of Ford’s reinvention may be customization.

Ford recently showcased modified versions of the Bronco, F-150, Maverick, and Mustang, with upgrades ranging from decals and storage systems to suspension packages, off-road equipment, and engine enhancements. Ford says roughly half of its customers already purchase some form of accessory or vehicle upgrade, and the company now wants to design more of those opportunities into the vehicle from the beginning rather than treat accessories as an afterthought.

That changes customization from an aftermarket activity into a product architecture and fulfillment strategy.

A customer can select a vehicle, add factory-backed upgrades, and potentially finance the complete package as part of the original transaction. Ford and its dealers capture more revenue, while the customer receives a more personalized vehicle without having to assemble the solution independently. The margins on accessories and performance packages can also be attractive.

But customization creates a different supply chain challenge: deciding where and when the final configuration should be completed. Some options may be installed at the assembly plant. Others may be added at a regional modification center, a dealership, or by the customer after delivery.

This is essentially a postponement model. Ford can produce a relatively standardized base vehicle at scale and add selected features later in the fulfillment process. That reduces the need to forecast every possible finished configuration months in advance and gives the company more variety without burdening the main assembly line with excessive complexity.

The challenge is synchronization. Ford must know whether the accessory is available, whether it is compatible with the vehicle, where it should be installed, and whether installation capacity is available. The vehicle, hardware, installer, financing, and delivery schedule all need to align.

A missing cargo system or suspension package may not stop production of the vehicle itself, but it can still delay delivery of the vehicle the customer actually ordered. The finished product is no longer necessarily the vehicle that leaves the factory. It may be the vehicle plus a coordinated package of accessories, software, dealer services, and financing.

That makes the order-to-delivery process considerably broader than it was in the past.

Affordable EVs Require a Different Production System

Ford’s strategy also contains an obvious tension. The company has eliminated several lower-priced vehicles, leaving fewer options for entry-level buyers. At the same time, Ford says it plans to introduce five vehicles priced below $40,000 before the end of the decade, beginning with an electric pickup expected to cost around $30,000.

Ford cannot achieve that goal by simply removing features or accepting lower margins. It needs a fundamentally different cost structure.

That is the purpose of Ford’s Universal EV Platform and the manufacturing system being developed around it. Ford says the new platform will use roughly 20% fewer parts, 25% fewer fasteners, and 40% fewer assembly workstations than a conventional vehicle program. The company also expects assembly time to be about 15% faster.

Those are not minor engineering changes. They go directly to the economics of affordable electric vehicles.

Battery cost matters, but so do labor content, parts count, logistics touches, manufacturing space, quality failure points, and capital investment. Fewer components can simplify sourcing, lower inbound logistics requirements, reduce assembly work, and decrease the number of things that can go wrong.

Ford is also rethinking the assembly process itself. Major sections of the vehicle will be built separately and then brought together, rather than moving through a purely traditional linear assembly sequence. The objective is to make the process faster, simpler, and less capital intensive.

Ford’s first generation of electric vehicles demonstrated that generating demand is not enough. The company also has to build EVs profitably. That is why the new platform matters more than any individual model launch.

Ford is trying to create a manufacturing system that can compete with companies that began with newer architectures, fewer legacy constraints, and lower-cost production models.

The Risks of a Smaller Ford

The financial logic behind Ford’s strategy is clear, but so are the risks.

The first is affordability. Ford’s least expensive vehicles now begin near $30,000, making it harder for the company to attract first-time buyers and customers looking for basic transportation. Those buyers may eventually move into higher-priced trucks and SUVs, but if their first vehicle comes from another manufacturer, Ford may lose the opportunity to build that long-term relationship.

The second risk is cyclicality. Trucks, large SUVs, off-road vehicles, and performance cars can produce strong margins, but they may also be more exposed when fuel prices rise, credit tightens, or consumers become more cautious.

The third risk is execution. Ford must improve quality, control warranty costs, protect production of its core franchises, scale customization, and launch an entirely new EV manufacturing system. A narrower portfolio reduces complexity, but it also leaves less room for operational failure.

Ford’s “passionate” reinvention is often described as a decision to stop building vehicles customers do not care enough about. That is only part of the story.

Ford is reducing low-return product complexity, concentrating volume around platforms where it has brand strength and manufacturing scale, designing customization into the customer order and fulfillment process, and developing a simpler production system intended to make affordable EVs economically viable.

The company is betting that it can earn more from a smaller number of differentiated vehicles than from trying to offer something for every buyer. That bet will not be won by branding alone. It will be won, or lost, through manufacturing discipline, supplier execution, configuration management, quality, and the ability to deliver more customized vehicles without recreating the complexity Ford is trying to remove.

The post Ford’s Reinvention Begins With Fewer Vehicles and Less Complexity appeared first on Logistics Viewpoints.

Most people will look at the latest SpaceX-NVIDIA announcement and see another major AI hardware agreement. Given NVIDIA’s central role in the artificial intelligence market, that reaction is understandable. Every week seems to bring another announcement involving billions of dollars of AI infrastructure investment and another company seeking access to increasingly scarce advanced computing resources.

But I believe the SpaceX announcement deserves a closer look.

During the company’s August earnings call, Elon Musk announced that SpaceX intends to standardize its future artificial intelligence infrastructure on NVIDIA platforms. More importantly, SpaceX and NVIDIA will collaborate on the computing payloads for SpaceX’s planned Starlink AI1 satellites, a program intended to bring significant AI computing capability into orbit.

At first glance, this may sound like a technology procurement decision. In reality, it may represent the beginning of a much larger shift in how computing infrastructure is deployed and operated.

For decades, computing has steadily moved closer to where data is generated. Mainframes gave way to distributed computing. Enterprise data centers expanded into cloud infrastructure. More recently, edge computing emerged to process data closer to factories, warehouses, vehicles, and industrial assets.

The SpaceX-NVIDIA partnership points toward the next logical step: moving certain forms of AI processing closer to the point where space-generated data originates.

Whether that vision ultimately succeeds remains uncertain. However, the announcement highlights an emerging technology direction that supply chain and logistics leaders should begin watching closely.

Why SpaceX Is Uniquely Positioned

What makes this effort particularly interesting is not the AI technology itself. Numerous companies are pursuing advanced AI initiatives. Rather, it is the combination of capabilities that SpaceX brings to the table.

Building an orbital computing platform requires far more than advanced processors. It requires launch capability, spacecraft manufacturing, satellite operations, communications infrastructure, software platforms, and the financial resources necessary to support years of development.

Few organizations possess even a fraction of those capabilities.

SpaceX designs and manufactures its own launch vehicles. It operates reusable rocket systems that have dramatically reduced the cost of access to space. It manufactures satellites at scale. It operates Starlink, one of the largest satellite communications networks ever deployed. It has growing AI ambitions and now intends to build those ambitions around NVIDIA’s computing architecture.

NVIDIA, meanwhile, has evolved from a semiconductor company into the foundational infrastructure provider for the AI economy. Its value extends far beyond GPUs. The company provides the software, networking, development environments, simulation tools, and computing architectures that increasingly serve as the foundation for large-scale AI deployments.

Together, the two companies are attempting to combine launch infrastructure, communications infrastructure, and AI infrastructure into a single integrated platform.

That combination is unusual.

Traditional cloud providers control computing resources but not launch systems. Aerospace companies build spacecraft but generally do not operate hyperscale AI environments. Satellite operators manage communications networks but typically depend on external partners for launch services and computing infrastructure.

SpaceX is attempting to bring all of these elements together under one roof.

Why Put AI in Space?

The obvious question is why anyone would want to place AI computing infrastructure in orbit in the first place.

The answer is not because space is inherently a better location for data centers.

In fact, space creates enormous engineering challenges. Computing equipment must survive radiation, extreme temperatures, launch stresses, and years of operation without direct maintenance. Heat dissipation is difficult. Hardware replacement is expensive. Power generation is constrained. Communications remain dependent on links to terrestrial infrastructure.

These are not trivial problems.

For that reason, orbital computing is unlikely to replace traditional data centers anytime soon. Training large language models and running mainstream enterprise applications will continue to be far more practical on Earth.

The more compelling near-term opportunity involves edge computing.

Modern satellites generate enormous amounts of data. Earth observation systems capture imagery. Weather satellites monitor atmospheric conditions. Communications satellites process vast amounts of network traffic. Scientific satellites continuously collect measurements and observations.

Traditionally, much of that data must be transmitted to Earth before meaningful analysis can occur.

As satellite networks continue to expand, this model becomes increasingly inefficient.

Instead of transmitting every image, every sensor reading, or every observation, future AI-enabled satellites could process information directly in orbit. A satellite might identify a developing wildfire, detect port congestion, recognize vessel movements, assess storm activity, or identify infrastructure damage before transmitting only the relevant insights.

The result is a reduction in bandwidth requirements, lower latency, and faster decision-making.

Rather than acting solely as sensors, satellites become intelligent participants in a larger information network.

What This Could Mean for Supply Chains

While SpaceX has not announced any supply-chain-specific applications, it is worth considering how orbital AI infrastructure could eventually influence logistics operations.

Supply chains increasingly depend on external signals.

Port congestion, weather disruptions, vessel movements, geopolitical events, infrastructure failures, natural disasters, and transportation bottlenecks all influence operational decisions. Yet many of these signals originate outside the enterprise and often require multiple layers of processing before becoming operationally useful.

Today’s visibility platforms have significantly improved access to information, but visibility alone is no longer enough.

Most organizations are now facing the opposite problem. They have more data than they can effectively process.

This is where artificial intelligence becomes important.

The next generation of supply chain platforms will increasingly focus on transforming raw information into machine-readable awareness. Instead of simply reporting events, systems will identify patterns, assess risk, prioritize responses, and coordinate actions.

Imagine an AI-enabled satellite network monitoring activity at major ports around the world.

The system could observe vessel density, weather conditions, terminal activity, and transportation flows. AI operating near the data source could identify emerging congestion patterns and generate structured events before delays become obvious through traditional operational data.

Those events could then flow into transportation management systems, supply chain control towers, and exception management platforms.

Transportation systems could identify affected shipments.

Inventory systems could calculate downstream exposure.

Customer service platforms could anticipate impacts.

Exception management systems could evaluate alternative actions.

The satellite would not be making these decisions directly. Instead, it would become part of a broader ecosystem of intelligent systems that sense, communicate, reason, and coordinate responses.

This vision aligns closely with the broader industry movement toward connected intelligence, where AI systems communicate across functions, maintain context, retrieve relevant information, and support increasingly autonomous decision-making. As discussed in ARC’s recent research on AI-enabled supply chains, the future lies not in isolated AI applications but in interconnected networks of intelligent systems capable of collaborating across the enterprise.

Orbital computing could eventually become another component within that architecture.

The Economics Remain the Critical Question

Despite the excitement surrounding the announcement, significant questions remain.

The largest is economics.

Moving computing infrastructure into orbit only makes sense if it creates enough value to offset the considerable costs involved. Launch costs may be declining, but they have not disappeared. Satellites remain expensive. Computing hardware continues to evolve rapidly. Operational lifecycles are difficult to predict.

Not every workload belongs in space.

In fact, most do not.

The most promising applications are likely those where proximity to space-generated data creates a meaningful advantage or where communications constraints make local processing more efficient than transmitting raw information to Earth.

Earth observation, defense, communications optimization, scientific research, weather forecasting, and autonomous spacecraft operations appear to be among the strongest early candidates.

Whether those use cases ultimately support a large-scale orbital computing market remains an open question.

History suggests caution. Many technically impressive technologies fail because they solve problems that customers are unwilling to pay to address.

At the same time, history also shows that entirely new infrastructure categories often appear unnecessary until they become indispensable.

Cloud computing, mobile internet, GPS, and commercial satellite communications all faced skepticism during their early years.

The same may ultimately prove true for orbital computing.

Looking Beyond the Announcement

The most important takeaway from the SpaceX-NVIDIA partnership is not that SpaceX selected NVIDIA hardware.

It is that SpaceX appears to be pursuing a vision that extends beyond rockets, satellite internet, or even artificial intelligence itself.

The company is attempting to build a new layer of infrastructure that combines launch systems, communications networks, satellite operations, and AI computing into a single integrated platform.

Whether that vision succeeds remains to be seen.

The engineering challenges are substantial. The economics remain uncertain. The market opportunity is still emerging.

Yet the strategic direction is becoming clearer.

As AI continues moving closer to where data is generated and as organizations seek faster ways to transform information into action, the boundary of the data center may begin extending beyond terrestrial networks.

Most AI infrastructure will remain firmly on Earth for the foreseeable future.

But SpaceX and NVIDIA are betting that part of the next generation of computing infrastructure will operate somewhere else entirely.

It will be overhead.

The post SpaceX and NVIDIA Are Preparing to Move AI Infrastructure Into Orbit appeared first on Logistics Viewpoints.

Small modular reactors are moving from a technology discussion toward an industrial execution test.

Governments, utilities, energy-intensive manufacturers, data-center developers, and nuclear vendors are increasingly interested in reactors that can be deployed in smaller increments than conventional nuclear plants. The attraction is understandable. SMRs could provide reliable low-carbon electricity, industrial heat, hydrogen production, remote power, and replacement capacity at retiring coal sites.

But the future of SMRs will not be determined by reactor physics alone.

It will depend on whether the emerging industry can build repeatable designs, secure nuclear fuel, qualify suppliers, manufacture components at scale, license projects efficiently, finance first-of-a-kind plants, and create enough orders to support a durable production system.

That makes the global SMR market fundamentally a supply-chain story.

A Large Pipeline Does Not Yet Equal a Market

The number of proposed SMR designs is impressive.

The OECD Nuclear Energy Agency’s current digital dashboard tracks 129 designs worldwide, although only 79 are included in its detailed public assessment. Some excluded designs remain under development, while others have been paused, cancelled, or lack sufficient financial and organizational support.

This distinction is important.

A market with more than 100 concepts can appear mature when viewed through the number of announced technologies. In reality, only a much smaller group has progressed far enough in licensing, financing, siting, supply-chain preparation, fuel availability, and customer engagement to represent credible near-term deployment candidates.

The NEA assesses SMR progress across these broader readiness dimensions rather than evaluating technical design alone. Its 2025 review found that 51 designs were engaged in pre-licensing or licensing activities across 15 countries, while seven designs were already operating or under construction.

The global ecosystem is therefore broad, but uneven.

Some programs are approaching commercial deployment. Others remain promising engineering concepts. Still others may never secure the capital, customers, regulatory approvals, fuel, or supply-chain capacity required to proceed.

The question is no longer whether engineers can design smaller reactors. It is which designs can become standardized, financeable products supported by repeatable industrial systems.

Modularity Must Become More Than a Design Feature

SMRs are generally described as reactors producing less than approximately 300 megawatts of electricity, although some microreactor concepts are far smaller. Many are intended to use modular manufacturing, factory production, transportable components, and phased deployment.

The economic argument is not simply that a smaller reactor costs less in total.

A smaller project may require less upfront capital, shorten the period between investment and revenue, reduce the consequences of construction delays, and allow capacity to be added incrementally as demand grows.

The deeper promise is industrial repetition.

Instead of treating every nuclear plant as a largely customized megaproject, SMR developers want to produce standardized modules, equipment packages, and construction sequences that can be repeated across multiple sites.

That is the theory. The challenge is that modularity creates economic value only when repetition actually occurs.

A factory cannot achieve efficient production economics if it builds one reactor module, waits several years, and then switches to a different design. Suppliers cannot justify specialized nuclear capacity without credible order volumes. Skilled workers cannot develop learning-curve advantages when projects remain isolated.

The first reactor of a design is therefore unlikely to reveal its mature cost. The critical economic question is whether the developer can move from a first-of-a-kind project to a repeatable fleet.

Too Many Designs Can Fragment the Supply Base

Technical diversity can be valuable. Different customers require different reactor sizes, temperatures, fuels, and operating characteristics.

A remote mine does not have the same requirements as a major utility. A chemical plant seeking process heat may need a different reactor than a data center seeking continuous electricity. Some customers may favor established light-water technology, while others may value the higher temperatures or fuel efficiency offered by advanced designs.

But design diversity creates a supply-chain problem.

If every project requires different forgings, pumps, valves, fuels, control systems, containment structures, and qualification processes, suppliers cannot achieve volume manufacturing. Regulators must evaluate more designs. Operators must develop different training programs. Maintenance organizations must support incompatible equipment families.

The industry could then reproduce one of the central weaknesses of conventional nuclear construction: too much customization and too little repetition.

The NEA has identified standardization and streamlined global supply chains as important opportunities for improving SMR economics.

The likely market outcome is consolidation.

Many designs may continue through research and early licensing, but a smaller number of platforms will probably capture most commercial orders. The winners may not necessarily have the most technically ambitious reactors. They may be the companies that secure customers, obtain regulatory acceptance, establish fuel supply, and create an executable manufacturing strategy.

Fuel May Become the Binding Constraint

Nuclear fuel is not interchangeable across all SMR designs.

Many light-water SMRs can use forms of fuel similar to those used by the existing reactor fleet. Some advanced reactors, however, require high-assay low-enriched uranium, commonly known as HALEU, or other specialized fuel forms.

That creates a potential sequencing problem.

Developers may complete designs and identify customers before sufficient commercial fuel capacity is available. Fuel producers, meanwhile, may hesitate to invest in large facilities without firm reactor orders.

The result is a circular dependency: reactors need fuel supply to become financeable, while fuel suppliers need reactor demand to justify investment.

Governments increasingly recognize this vulnerability. In the United States, recent federal initiatives have focused on fuel fabrication, domestic enrichment, nuclear component manufacturing, and other supply-chain gaps alongside direct reactor support.

Fuel availability will influence which designs reach deployment first. A reactor using an established fuel supply may hold an execution advantage even if another design offers potentially better long-term performance.

The First Projects Will Shape the Entire Sector

First-of-a-kind projects will carry unusually high strategic importance.

They will establish real construction costs, validate schedules, test regulatory processes, qualify suppliers, train workers, and reveal whether modular manufacturing performs as expected.

A successful first project can create confidence for utilities, lenders, regulators, and subsequent customers. A major delay or cost overrun can affect not only one developer but the broader perception of the SMR category.

This is why early projects often require public support.

Private investors are being asked to finance technical, regulatory, construction, market, and supply-chain risks simultaneously. The first plant must absorb expenses that later projects may avoid, including design completion, supplier qualification, licensing work, factory setup, and workforce development.

The U.S. Department of Energy has committed up to $800 million to initial projects involving the Tennessee Valley Authority and Holtec, with the explicit objective of supporting first deployments and associated supply chains. In May 2026, it also announced more than $94 million for eight additional companies addressing licensing, manufacturing, fuel, and site-preparation gaps.

These programs reflect an important reality: early SMR deployment is not merely electricity procurement. It is industrial base development.

Coal Sites and Industrial Campuses Could Reduce Deployment Risk

One of the strongest SMR opportunities may be at existing energy and industrial sites.

Retiring coal plants often have transmission connections, water access, transportation infrastructure, operating workforces, and communities familiar with large energy facilities. Reusing portions of that infrastructure could reduce site-development requirements and preserve local employment.

Industrial campuses may offer another attractive model. Chemical plants, steel producers, refineries, mining operations, and hydrogen producers need dependable energy and may be able to use both electricity and heat.

Data centers have added another source of demand. Their need for large quantities of continuous electricity has increased interest in nuclear generation, particularly where grid capacity is constrained.

These customers could support deployment through long-term power agreements or direct investment. But they will expect predictable costs and schedules.

An industrial customer cannot base expansion plans on a reactor that arrives years late. The nuclear project must fit within the customer’s broader capital program, energy strategy, and risk tolerance.

Global Deployment Will Require Local Supply Chains

SMR developers frequently describe international markets. The same design may be promoted in North America, Europe, Asia, Africa, and the Middle East.

Yet nuclear construction remains deeply local.

Projects must comply with national regulations, labor practices, quality requirements, security rules, environmental reviews, and political expectations. Governments may also require domestic manufacturing or local content as a condition of support.

This creates tension between global standardization and national industrial policy.

The strongest model may be a standardized reactor platform supported by a controlled global network of qualified regional suppliers. Certain high-value or safety-critical components could come from centralized facilities, while civil construction, balance-of-plant equipment, and services are sourced closer to each project.

Achieving that model will require harmonized codes, regulator cooperation, shared qualification standards, and disciplined configuration management.

Without those controls, localization can become redesign, and redesign can eliminate the economics of repetition.

The Market Will Be Won Through Execution

SMRs have credible strategic advantages. They can add capacity incrementally, serve locations unsuitable for very large reactors, support industrial decarbonization, and provide reliable power alongside variable renewable generation.

They also face substantial risks.

First projects may be expensive. Licensing can take longer than anticipated. Fuel supply may constrain advanced designs. Customers may delay commitments. Suppliers may be unwilling to invest without firm orders. Too many competing technologies may fragment the market before scale is achieved.

The industry should therefore be judged by evidence of execution rather than the number of announced designs.

The most important indicators are increasingly clear:

A design approaching regulatory approval

A committed site and customer

Credible financing

Secured fuel

Qualified suppliers

Manufacturing capacity

A realistic construction plan

Follow-on orders using the same design

Those conditions turn a reactor concept into an industrial product.

The global nuclear renaissance will not be delivered by a single technological breakthrough. It will be built through orderbooks, factories, fuel facilities, qualified components, skilled workers, repeatable construction, and regulatory learning.

SMRs may eventually change how nuclear power is deployed.

But first, the industry must prove that it can manufacture and deliver them as a supply chain rather than construct each one as a national experiment.

The post The Global Small Modular Reactor Ecosystem: Why Supply Chains Will Shape the Nuclear Renaissance appeared first on Logistics Viewpoints.