Robotics is reaching live, mixed-case fulfillment

Descartes Systems Group reported Q2 revenue of $201.108 million, up 11.84% year over year, with adjusted EBITDA of $94.4 million and net income of $50 million.

The company said the quarter included completed acquisitions of TIE and Extensiv, adding freight-broker transportation management and 3PL warehouse and fulfillment capabilities to its logistics software portfolio. The reported AI angle is an expanding data and workflow base, not a disclosed standalone model benchmark.

Services revenue reached $188.6 million, or 94% of total revenue, while operating cash flow rose to $81.3 million. For customers, a broader suite could reduce integration handoffs, but migration quality will determine whether consolidation improves execution.

India's industrial and warehousing demand across eight major cities reached nearly 22 million square feet in the first half of 2026, up 12% year over year, while 3PLs represented 30% of leasing.

The report argues that the next advantage will come from connecting inventory, transport, customer and facility data so infrastructure can be used strategically. It contrasts fast-moving consumer and quick-commerce networks with slower B2B logistics adaptation.

As Grade A supply expands, adding sites alone may not improve lead time or working capital. AI-enabled planning and visibility can help decide where stock belongs and how capacity should flex, but only when operating data is shared across the network.

Global Trade Magazine reports that rising labor expense and unpredictable demand are pushing organizations to reconsider dedicated warehouses in favor of shared capacity.

The shared model shifts some fixed labor overhead into flexible operating expense and lets multiple shippers use a facility, while AI and data can forecast volume, staffing and space requirements. The article emphasizes economics and operating flexibility rather than a specific vendor deployment.

For mid-market shippers and 3PLs, the tradeoff is between underused labor in a dedicated site and service risk in a shared operation. The right comparison includes peak throughput, inventory accuracy, accessorial charges and the cost of overflow capacity.

Inbound Logistics' 2026 technology-trends review says AI is becoming part of enterprise execution rather than a standalone feature, as logistics operators seek faster decisions under labor and network pressure.

The review highlights Aera Technology's always-on decision agents, Decklar's combination of physical signals and enterprise data, Lucas Systems' AI orchestration engine, Deposco's causal-AI WMS corrections and multi-agent warehouse platforms such as SnapControl.

The cited examples include claims such as up to 15% logistics-cost reduction from Aera, more than 112 billion picks powered by Lucas Systems and a 90% short-ship reduction for Psycho Bunny using Deposco. Vendor-reported results require local validation, but the common design is insight connected to an execution step.

Supply Chain Management Review previewed 2026 NextGen sessions built around warehouse execution, inventory accuracy, system testing, workforce change and AI decision support.

The listed cases include AI and deterministic testing for WMS and EDI, edge AI cameras at docks and material-handling equipment, an automated fulfillment center that doubled throughput and raised picking capacity by 25%, and autonomous inventory intelligence that audited one million square feet in less than 24 hours while identifying 90% of discrepancies before customers saw them.

The preview presents customer-provider case studies rather than a single comparable benchmark. Its common lesson is that performance improvement requires a test harness, physical evidence, operator roles and a KPI that can be checked after deployment.

AS Watson says it will not use AI to cut its workforce across more than 17,000 stores in 31 markets, instead pairing staff with digital assistants and directing warehouse automation toward ergonomics.

The retailer's digital assistants handle routine lookups and operational tasks, while heavy-lifting robotics operate at a Foshan distribution warehouse. The company reports that women make up 62% of that facility's workforce after the robotics change, compared with a regional manual-handling norm of roughly 80% male participation.

The operating model treats automation as work redesign: remove physical strain and routine searches while preserving customer-facing judgment. AS Watson tracks engagement and customer feedback rather than labor reduction alone, a useful control for warehouse adoption.

TechBullion describes warehouse optionality as pre-negotiated, flexible storage and distribution paths that absorb tariffs, port strikes, demand shocks and geopolitical disruption.

The proposed model combines real-options thinking with connected planning, inventory visibility and decision support so a shipper can activate alternate facilities before an emergency. AI can compare capacity, distance, labor, inventory and service commitments across those paths.

A flexible network trades some fixed-cost efficiency for faster recovery and less dependence on one node. The decision should be measured against stockout exposure, emergency freight, dwell and the cost of keeping alternate capacity available.

Capgemini describes a move from periodic planning cycles toward AI-augmented supply chains that continuously connect demand, supply, operations and financial decisions in automotive.

The model uses live operational data, scenario analysis and AI recommendations to identify changes in demand, capacity, inventory and supplier risk. Its purpose is not only forecast improvement but faster reallocation of working capital and production response.

For logistics networks, always-on planning can reduce the lag between an order signal and an inventory or transport decision. The implementation challenge is governance: planners must distinguish a model recommendation from an approved change to a lane, node or supplier.

Logistics Viewpoints argues that AI investment is becoming a competitive logistics race as operators seek faster decisions, lower cost and greater resilience amid volatility.

The discussion centers on the combination of predictive analytics, automation, digital control towers and connected execution systems rather than on one foundation model. The differentiator is the ability to turn a signal into a coordinated planning or transportation action.

The implication for 3PLs is that a pilot can create little advantage if it does not connect to a real operating constraint. Network response time, carrier capacity, inventory position and customer promise are the practical test points.

Inbound Logistics' 21st annual 3PL survey says providers and shippers reported healthy growth while still facing operating-cost pressure, capacity issues and compliance burdens.

The report says AI and visibility technology have moved to the front of the industry agenda: shippers identify AI deployment as a leading challenge, while 3PL respondents call AI the top disruptive innovation and 50% still cite it as a challenge.

The survey also records growth in fulfillment, value-added services, transloading, e-commerce and distribution-center management. A new customer relationship therefore needs a technology, data and operating-model conversation, not only a rate card.

Yingli Wang and J. Mark Munoz argue that AI in logistics should be treated as a management redesign issue spanning efficiency, resilience, sustainability and accountability.

The analysis covers machine learning, computer vision, autonomous systems, generative AI and agentic AI across forecasting, replenishment, inspection, warehouses, vehicles, procurement and planning. It stresses that fragmented ERP, WMS, procurement and partner data can undermine the decision layer.

The article identifies a prediction-to-action gap: a demand signal creates value only when inventory, procurement, capacity or customer decisions can change. For a 3PL partnership, onboarding therefore needs authority, governance and routines, not only an API connection.

Descartes Systems Group announced an agreement to acquire Extensiv, a cloud warehouse-management and inventory-management software provider serving brands, 3PLs and warehouses.

Extensiv brings warehouse, order and inventory workflows into Descartes' logistics technology portfolio, with integrations across marketplaces, shopping carts, carriers, accounting systems and third-party logistics operations. The combination is positioned to connect warehouse execution with broader transportation and trade data.

For 3PL onboarding, the strategic question is whether a cloud WMS can shorten integration for multi-client operations without weakening inventory control. The acquisition is an ownership and platform move, not proof of a measured productivity gain, so buyers should separate product roadmap from verified site outcomes.

Locus Robotics announced Locus Array, a tower-style mobile robot with an integrated picking arm and AI-powered perception, at MODEX 2026 in Atlanta.

Array combines autonomous movement, robotic pick-and-place, perception and the LocusONE orchestration platform, operating alongside Locus Origin and Vector robots. Early-access deployments are underway in North America, including DHL Supply Chain, with planned expansion to Europe and Asia-Pacific.

The system targets labor constraints, rising costs and variable item handling by bringing robots-to-goods into conventional warehouse racking. It is a company announcement, so operators still need site-specific evidence on item range, replenishment, safety and exception recovery before scaling.

Hirsch Pipe & Supply, which operates 30 locations, reports a 25% improvement in picking efficiency after deploying PathGuide's Latitude Warehouse Management System.

Latitude integrated with the distributor's existing ERP, replaced spreadsheet-driven processes and introduced zone-based picking while making inbound and operational information available across warehouse, purchasing and sales teams.

The case shows how workflow structure and shared data can matter more than a standalone AI label. Better information access and labor allocation can improve receiving and pick flow, but the reported gain still needs replication against a site's order mix.

KNAPP's 2026 logistics outlook identifies AI co-pilots in WMS/WES, swarm intelligence for AMRs, computer vision, forecasting with digital twins and sustainability management as five warehouse trends.

The approach combines machine learning with API-centric data, multi-agent robot coordination, camera-based barcode and condition capture, and simulations that combine stock, orders and external factors. KNAPP also describes computer vision in goods-in and returns management.

The range of use cases puts data quality and orchestration ahead of a single model. Warehouses can improve labor, robotic routing, inspection and returns decisions only when the system can surface uncertainty and retain an operator fallback.

Emiza opened two multi-client fulfillment facilities in Bhiwandi, adding a combined 232,000 square feet and bringing its Mumbai warehouse count to six.

The sites use technology-enabled systems for storage optimization, inbound and outbound operations, inventory visibility and replenishment across downstream networks. Emiza says the facilities support brands that need capacity across channels without repeatedly building their own infrastructure.

The expansion is designed to place inventory closer to demand across Western India, with facilities spanning 75,000 and 157,000 square feet and a stated network footprint of more than 11 cities and 11 states. Emiza expects more than 700 jobs as the operation scales.

VX Logistics describes an end-to-end cold-chain system for fresh produce, including AI vision inspection, automated handling, temperature and humidity management, and rapid warehouse throughput.

Vision assesses color, size and defects, while IoT sensors track temperature and location; the company says its solution supports close to 200 million boxes of berries annually for Driscoll's and covers warehousing, ripening, sorting and packaging for Zespri in China.

The workflow turns product condition into a measurable logistics record rather than leaving quality as a port-to-warehouse black box. The claims are company-reported, but the design ties inspection and sensor data to spoilage risk and saleable inventory.

Inbound Logistics describes warehouse robotics expanding across transport, picking and inbound movement as operators address labor shortages and heavier material-handling work.

The roundup cites Amazon's Proteus mobile robot, which moves carts weighing about 900 pounds and is deployed at 25 U.S. fulfillment centers, plus ShipLab's phased collaborative-robot deployment and an Ambi Robotics-Pickle integration for dock-to-warehouse package movement.

The examples show a staged path from point-to-point transport toward integrated order picking and continuous inbound flow. Each use case still requires safety validation, WMS handoff design and proof that robots reduce travel or handling effort without creating new queues.

GXO and Exotec announced a live Skypod deployment at GXO's Venlo facility serving Guess, using 127 robots, 60,000 rack locations, eight goods-to-person stations and 200 meters of conveyor.

The system handles inbound logistics, value-added services and outbound distribution, with robots navigating the rack structure and bringing inventory to operators. The companies say the installation processes 40,000 to 70,000 pieces per day and can reach 2,200 order lines per hour during peaks.

GXO selected Exotec as a single integrator, making one party accountable for storage, stations, conveyor and handoff performance. The reported capacity must still be checked against SKU mix, labor at stations and exception recovery.

CJ Logistics began operating an approximately 19,000-square-meter cold-chain center in Anseong, South Korea, handling 2,040 refrigerated and frozen products from 112 e-commerce sellers.

Its LoIS eFLEXs fulfillment system links real-time orders, inventory, picking, packing, shipping and delivery; AI recommends order combinations that reduce worker travel and sequences automated equipment and process workloads to prevent bottlenecks.

The center has a daily maximum of 25,000 orders and serves customers including CJ CheilJedang and Starbucks. In a temperature-sensitive operation, the value case combines order throughput with inventory visibility, labor travel, cold exposure and product integrity.

Wurth Australia appointed Dematic to automate fulfillment at its Yatala and Keysborough distribution centers serving more than 22,000 products and 56,000 active customers.

The design combines AutoStore goods-to-person storage with manual picking zones, downstream carton sealing, labeling and dispatch sortation. Split-case and full-case orders will move between automated and manual areas under a common fulfillment approach.

Wurth links the project to accurate, complete orders and national service rather than automation for its own sake. The two-site design creates a useful test of common inventory, routing and exception rules across different facilities.

Supply Chain Management Review reports that tanker attacks near the Strait of Hormuz, procurement restrictions and damage to warehouses around Kyiv are forcing companies to reassess transport and storage risk.

The briefing connects geopolitical events with alternate sourcing, dispersed inventory, responsive forecasting and logistics planning. AI can combine corridor status, supplier exposure, inventory and delivery promises to rank outbound options, but the decisions still require policy and human approval.

For carriers and 3PLs, the question is how quickly an at-risk lane can be replaced without losing customer commitments or inflating premium freight. Resilience has to be visible in cost per shipment, dwell and service recovery.

Kardex is expanding its AutoStore integration business into Canada after reporting dozens of Americas deployments and more than 150 AutoStore deployments globally.

The offering includes FulfillX warehouse execution, SnapVac grid cleaning, BinInductAI for bin induction and an Intuitive Picking Assistant that projects information at the AutoStore port. Those tools target uptime, induction quality and operator guidance around the storage system.

Expansion into Canada broadens access to a dense storage and retrieval pattern for retail, manufacturing, healthcare, e-commerce and 3PL customers. The operational proof is whether induction and port guidance reduce queueing, misplacement and outbound labor minutes.

Hong Kong's 2026-2030 plan prioritizes the city as an international maritime and trade center, an innovation-and-technology center and a high-value supply-chain services hub.

The plan calls for cargo and trade data-platform integration, digital trade documents, cross-border data exchange, and more sea-air intermodal transshipment. It also identifies AI, resilience and data infrastructure as technology priorities and describes an AI-enabled, facial-recognition pilot for contactless vehicle clearance.

For outbound networks, the policy direction points toward faster document, clearance and corridor decisions rather than a single new warehouse asset. Carriers and 3PLs still need to assess interoperability, data governance and the effect on border dwell before assuming service gains.

Buywander raised a $21 million Series A led by Madrona Venture Group and Inspired Capital to expand its auction marketplace for returned and overstocked retail goods.

The model connects retailers including Amazon, Target, Walmart and Home Depot with local buyers who collect purchases from Buywander warehouses; the company says it avoids last-mile shipping and uses timed auctions beginning at $1.

Buywander had eight locations, with Denver and Chicago recently launched and Minneapolis planned next. Each new market requires warehouse operations, retailer relationships and item-level reverse-logistics data to keep recovery value ahead of handling and markdown cost.

Automotive Logistics describes AI as a way for automotive supply chains to recover value and reduce waste while facing complexity, inventory pressure and disruption.

The operating concept combines supply-chain data, predictive analytics and decision support to identify where materials, parts or transport choices can be reused, redirected or changed. The article frames AI as an orchestration aid rather than an autonomous replacement for reverse-network owners.

For returns and reverse logistics, the decision is whether a returned, surplus or damaged item should be repaired, reused, recycled, liquidated or moved to another node. The financial test is recovered value after handling, storage and transport.

ClickPost's 2026 operations guide defines the fulfillment manager as the owner of receiving, storage, picking, packing, carrier coordination, returns and performance measurement.

The guide calls for WMS and TMS control, inventory synchronization, wave and slotting analysis, carrier allocation, exception handling and KPI instrumentation. It identifies return handling as a gap when organizations split ownership across warehouse, logistics and customer teams.

The management pattern is relevant to reverse logistics because a return crosses customer service, carrier, warehouse and resale decisions. A common owner can reduce dwell and missing handoffs, while the article's role guidance is not an independently measured AI deployment.

Market Research Future estimates the logistics market at $10,170 billion in 2024 and projects growth from $10,780.2 billion in 2025 to $19,305.7 billion by 2035.

The report links expansion to automation, artificial intelligence, digital software, analytics, e-commerce and greener operating practices rather than to a single logistics technology. It describes AI as a way to streamline processes, reduce cost and improve decisions across the network.

For 3PLs and warehouse operators, the forecast is a directional planning signal, not proof of site-level ROI. The practical consequence is that capacity, service speed and emissions choices increasingly need to be modeled together.

AI CERTs describes self-healing supply chains as systems that sense risk, adjust decisions and recover from disruption with less manual intervention.

The article combines transport, warehouse, inventory and supplier data in an orchestration loop, and adds sensors for predictive maintenance plus digital-twin simulation. It cites reported potential reductions of up to 40% in disruption recovery time and up to 50% in equipment downtime, with maintenance-cost reductions of 10% to 40%.

These figures are cited market and analyst claims rather than a named customer baseline. For logistics operators, the value case is a controlled exception loop that changes routing, inventory or maintenance only within approved policies.

Talking Logistics argues that visibility alone does not solve a disruption: a team must be able to shift mode, find capacity or reroute freight without rebuilding the network from scratch.

The proposed operating model connects planning, optimization and execution through orchestration, with alternate carriers, modes and playbooks built before conditions change. AI can rank responses from live shipment and capacity data, but the network still needs contracts and decision rights.

The article reframes continuous improvement as response design. A shipment alert has value only when it shortens time-to-action without adding handoffs, premium freight or customer-service surprises.