Haladir Raises $4.3M to Date to Build Operational Superintelligence

Haladir, the operational AI layer for logistics, led by Jibran Hutchins, Quan Huynh, Preston Schmittou, and Joseph Tso, has raised seed round, bringing the total funding to $4.3M, led by BoxGroup and Susa Ventures, with participation from Sunflower Capital, Valkyrie Ventures, XPRESS Ventures, and various angel investors, following earlier backing from Y Combinator, SV Angel, and Josh Browder.

Building Operational Superintelligence

“Superintelligence” in logistics and supply chains refers to a practical idea rather than hype. Large language models (LLMs) are powerful but unreliable as standalone decision-makers, as they primarily generate fast, educated guesses.

Greater value emerges when LLMs augment traditional models such as SMT/SAT, MILP, and other optimization and forecasting tools. In this approach, LLMs translate operational rules into structured constraints, while deterministic solvers compute optimal decisions.

As language models, LLMs perform best within verifiable and constrained systems, where optimization frameworks handle decision-making with precision.

From Data to Decisions

High-impact decisions in logistics require models that are both accurate and flexible, something most organizations lack due to fragmented systems. The first step is unifying siloed data from WMS, TMS, and OMS into a structured, interpretable layer, where inconsistencies are resolved and exposed as a usable operational graph.

Next comes formalization: encoding real operational rules, constraints, and dependencies. Process mining helps uncover how operations actually run, not how they are documented, ensuring models reflect ground reality.

Only then does optimization become meaningful. While some objectives are clear, many involve trade-offs across cost, service levels, and efficiency. Defining the right objective, or balance of objectives, is often the hardest and most critical step.

The Next Steps

Work is underway with 3PLs, distributors, and AI labs to apply operational superintelligence in practice. Initial efforts focus on improving core ML systems, demand forecasting, pick-path optimization, and ETA prediction, where gains in accuracy directly reduce losses.

From there, the same modeling approach extends into decision-making itself. The long-term vision is clear: every operational decision informed by accurate models and automated through systems that deliver the best possible action in real time.