Inside Optimal Dynamics and the Future of Autonomous Freight Planning

Ask any dispatcher in truckload what their day looks like and the answer is usually the same. Whatever plan existed at 8 a.m. has unraveled by lunchtime. Loads shift, tenders fall off, dwell time balloons, weather hits a region, drivers run out of hours in the wrong place, and suddenly the entire network starts collapsing in ways no spreadsheet anticipated. The problem has never been the dispatcher. The problem is that the industry is still planning deterministically in an environment that behaves like a stochastic system. Truckload planning has always looked cleaner on a TMS screen than it does in the real world. Optimal Dynamics is interesting to me because it starts from that reality. Instead of pretending the world is stable, it builds planning around the fact that everything will change.

The more I study truckload economics, the more I am convinced that the biggest inefficiencies come not from lack of data but from planning logic that cannot process uncertainty. Volatility is not a random inconvenience. It is the defining characteristic of the market. ATRI’s 2024 cost report showed total per mile operating costs hitting 2.27 dollars, the highest in the series. Empty miles range from 14 percent to more than 22 percent depending on lane density and region. Tender rejection cycles swing dramatically during seasonality. Dwell time at shippers and receivers can vary by more than 50 percent week to week. Driver availability, weather risk, and regional production patterns all shift daily. Yet the tools carriers rely on assume a straight line from origin to destination and a static world between those two points. When the plan breaks, fleets blame dwell time or market conditions or hiring challenges. But beneath all of that is a more fundamental limitation. The system was never designed for the environment it operates in.

Optimal Dynamics emerged from a body of academic work that attacked that limitation head on. Warren Powell, a Princeton researcher who spent more than thirty years working on high dimensional stochastic optimization, built the foundational models that eventually became the company’s decision engine. His research was focused on problems where the future is uncertain and the number of possible outcomes is too large for humans to reason through intuitively. Energy grids, airline networks, commodity logistics, and disaster response systems all behave this way. Truckload is the same, but the industry never had the computational infrastructure to support probabilistic planning. Powell’s insight was that the only way to improve decisions in a volatile environment is to simulate thousands of futures and choose the actions that hold up across the most scenarios.

Investors recognized the potential scale of this approach quickly. Before 2025, Optimal Dynamics had already raised more than 50 million dollars from institutional funds known for backing category defining platforms. Then came the 40 million dollar Series C led by Koch Disruptive Technologies. That round was not about incremental product development. It was a signal that decision intelligence in freight was moving from a fringe concept to a critical capability. Koch has a history of investing in technology that rewires industrial operations. Their involvement pushed OD into a different tier entirely. With Series C funding, the expectation shifted from helping dispatchers make better choices to building the planning substrate for an industry that handles more than 70 percent of the nation’s freight by volume.

The need for that substrate becomes obvious when you look at the economics of inefficiency. A 100 truck fleet running 100,000 miles per truck per year generates 10 million total miles. At 18 percent empty miles, that is 1.8 million empty miles. At 2.27 dollars per mile, the cost of that waste is roughly 4.1 million dollars annually. Scale that to a 1,000 truck fleet and the number becomes 41 million dollars. And this is just one dimension of inefficiency. Utilization is the next. Most OTR drivers run between 2,300 and 2,600 miles per week not because freight is unavailable but because network timing is poor. When drivers wait for preplans, sit in dwell, or land in the wrong region at the wrong time, their productivity collapses. Fleets that have adopted stochastic planning consistently report 6 to 12 percent utilization gains. At scale, that increment transforms the P&L. It also stabilizes retention. Recruiting costs often exceed 7,000 dollars per driver and turnover in OTR fleets is notoriously high. Better planning is one of the few ways to improve earnings predictably without adding labor cost.

Pricing and bid strategy represent another space where deterministic logic fails the industry. During annual RFP season, carriers submit thousands of lane bids based on historic volume, facility reputation, seasonality, and internal heuristics. But those decisions are made using average case assumptions. If a lane delivers 2 dollars per mile on average, a spreadsheet will show profitability. But if dwell spikes in November, or if a region tightens, or if a particular consignee consistently delays unloads, the average becomes meaningless. OD’s system models the probability distribution of profitability rather than the average. A lane that looks strong in deterministic planning might have a 40 percent risk of turning negative under realistic volatility. A marginal lane might exhibit stable upside because volume swings counterbalance dwell risk. The fleets that understand these distributions bid differently. They build lane portfolios the way investment firms balance exposure and volatility. When markets tighten or loosen, their networks behave more predictably because they were designed that way.

What makes Optimal Dynamics unique is not that it uses AI. The logistics world is full of companies claiming that. What makes OD different is that its architecture is built around one assumption: uncertainty is constant. The model simulates what happens when a driver arrives at a facility early or late, when weather introduces delays, when tender fallout increases on a Wednesday, when a region sees an unexpected inbound surge, when dwell time spikes by twenty percent, when a driver’s hours deplete in an unexpected window, or when a load must be reassigned because downstream constraints shift. Humans can consider two or three of these scenarios at once. OD can consider hundreds. The system is not predicting the future. It is stress testing it.

The direction this technology is heading is even more important. With the Series C behind them, OD is no longer building for one planner sitting behind a screen. It is building for carriers that want to shift from assisted decision making to autonomous planning. The distinction matters. Assisted decision making is when the system recommends a plan and a dispatcher approves it. Autonomous planning is when the system executes decisions in real time because the model has already evaluated hundreds of outcomes and selected the most resilient. The industry is moving toward that future whether fleets are ready or not. Rising costs, shrinking margins, volatile volumes, higher insurance premiums, and increasingly complex customer expectations make manual planning untenable. More freight is flowing through drop trailer networks. More loads require real time adaptation. More shippers are implementing fines for late arrivals. More drivers want predictable miles and home time. All of these pressures favor a planning engine that operates continuously, not in morning batches.

There is also a broader strategic implication in OD’s model. Trucking has historically treated trucks and trailers as separate planning problems. That separation is artificial. Trailer imbalances create millions of dollars in capital waste. ATRI reports that non fuel operating costs reached 1.78 dollars per mile in 2024, which means low trailer turns inflate the cost base of every asset a fleet owns. Power only carriers lose significant margin because they lack visibility into where trailers are sitting, how long they have been idle, and how many cycles they support per month. If OD extends its simulation engine into trailer pool optimization, it can attack one of the biggest hidden costs in the industry. A trailer that turns six times a month instead of four materially reduces capital requirements for a given throughput. Multiply that across thousands of trailers and the financial impact becomes enormous.

What I find most compelling about the company is not the technology itself but what it represents. For decades, trucking has been an industry where volatility wins. The fleets that survive are the ones that endure the chaos better than others. But endurance is not a strategy. Modeling uncertainty is. The companies that can reduce the variance in their networks will outperform those that rely on intuition and static planning. The more I study OD, the more I see a shift in what the industry will value in the next decade. Historically, advantage came from scale, relationships, buying power, or fleet size. In the future, advantage will come from mastery of uncertainty. The fleets that can plan across thousands of potential outcomes will absorb volatility without losing margin. They will accept freight others avoid. They will make network commitments with confidence. They will price lanes based on probability rather than instinct. They will reduce empty miles and improve utilization not because they work harder but because their planning logic is built for the real world.

There is no guarantee that OD becomes the dominant platform, but the direction is clear. The industry is moving toward autonomous planning because the economics demand it. Manual planning cannot keep up with the complexity of modern freight networks. Stochastic optimization is no longer academic theory. It is an operational requirement. With its funding, research lineage, and early market traction, Optimal Dynamics is positioned to become the system that carriers depend on when the old logic breaks. If it succeeds, it will not just automate decision making in trucking. It will change the definition of what good planning looks like.