OneRail Leverages Nvidia AI to Optimize Real-Time Last-Mile Delivery

OneRail has introduced an AI-driven delivery platform leveraging Nvidia technology to assist retailers, wholesalers, and distributors in determining the most efficient delivery methods for individual orders.

Named OmniSTAR, the system assesses various delivery channels—including in-house fleets, couriers, parcel services, and other logistics modes—to identify the most cost-effective solution that satisfies service level agreements, according to OneRail.

The platform integrates Nvidia’s cuOpt decision optimization engine and cuDF data processing software with OneRail’s proprietary delivery pricing and performance metrics. Nvidia’s accelerated computing infrastructure powers the routing and delivery-mode calculations.

According to OneRail, the system reduces computation times by up to 10x. Tasks that previously required 20 minutes can now be completed in under two minutes, while weekly calculations are reduced to approximately two days.

OneRail states that this accelerated processing enables optimization within live delivery operations, allowing multiple fulfillment options to be evaluated before an order is assigned.

“If you don’t have the ability to make lightning-fast decisions, you’re giving up margin,” Catania told CNBC in an interview. “Last-mile fulfilment is expensive.”

From prediction to delivery decisions

OneRail’s broader AI systems utilize prediction and optimization across various stages of the delivery process. The company reports that its machine-learning models forecast factors such as service duration, lateness risk, the likelihood of successful first-attempt delivery, and expected price ranges.

These predictions feed into optimization systems that determine order execution strategies. Separately, OneRail explains that OmniSTAR compares different fulfillment modes to select an option based on cost and service requirements.

Research on dynamic vehicle routing distinguishes between predicting changing conditions and recalculating operational decisions as new information emerges. A 2024 review in the European Journal of Operational Research identified travel-time prediction and real-time re-optimization as distinct areas of time-dependent routing research.

Nvidia cuOpt handles route optimisation

Nvidia describes cuOpt as an open-source, GPU-accelerated optimization library designed for vehicle routing and other mathematical optimization challenges.

Nvidia’s documentation indicates that cuOpt considers vehicle costs, capacities, travel times, operating windows, starting locations, and other constraints when calculating routes. Its cost models can incorporate distance, time, monetary cost, or a weighted combination of these metrics.

OmniSTAR applies cuOpt to both routing and delivery-mode selection. OneRail states this enables the system to compare available fulfillment options for an order and identify the lowest-cost option that meets service requirements.

OneRail notes that many retailers still rely on static rules or manual planning for these decisions, and OmniSTAR is designed to evaluate more delivery combinations within shorter operational timeframes.

Nvidia explains that cuOpt does not exhaustively test every possible route. Instead, the solver generates candidate solutions and iteratively improves them using GPU-accelerated heuristics to produce high-quality results within a specified computation time.

The platform also utilizes Nvidia cuDF, a GPU-accelerated library for tabular data processing, including filtering, joining, and aggregating datasets.

OneRail combines these capabilities with its own delivery data and operational models. Its dataset is derived from millions of deliveries across a network comprising more than 12 million drivers and over 1,000 logistics partners.

The data covers pricing and delivery performance across various transportation modes. OneRail states that OmniSTAR uses this information to identify cost-increasing delivery rules and assess how delivery choices impact item-level profitability.

The architecture disclosed for OmniSTAR focuses on GPU-accelerated data processing and mathematical optimization. Nvidia describes cuOpt as the optimization component used for problems such as vehicle routing.

Since cuOpt is stateless, changes in operating conditions require the optimization problem to be modeled and submitted again. Nvidia cites vehicle breakdowns, driver absences, road blockages, traffic, and new high-priority orders as examples of changes that trigger dynamic re-optimization.

OneRail states that OmniSTAR can rerun delivery scenarios as variables such as fuel costs, weather, and shipping conditions change. The company has previously stated that its use of cuOpt allows it to evaluate more routing scenarios and recalculate routes faster than its previous approach.

OmniSTAR moves into live operations

OmniSTAR is currently deployed with select enterprise customers.

At US Foods, OneRail reports that the system identified delivery configurations reducing margins, such as low-margin products being transported long distances using higher-cost equipment. US Foods subsequently used these insights to adjust pricing and restructure certain delivery patterns, according to OneRail.

OneRail also told CNBC that an unnamed large tire distributor using the platform achieved $40 million in run-rate savings over three years. The customer was not identified, and the savings figure was provided by OneRail. The company also told CNBC that it expects OmniSTAR to exceed $6 billion in gross merchandise volume during the fourth quarter of 2026.

CNBC reported that OneRail and Nvidia collaborated on the project for three years prior to its launch. OneRail stated that the collaboration included direct engagement with Nvidia’s cuOpt engineering team on last-mile delivery and large-scale logistics optimization, alongside its participation in the Nvidia Inception programme.

In March this year, FedEx launched FedEx SameDay Local in collaboration with OneRail, connecting customers to a national network of more than 1,000 delivery providers.

See also: A quarter of Nvidia’s business next year comes from labs it is financing

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.