Cost, Regulation, and Limitations for UAS Nobody Talks About

The aircraft that can fly beyond visual line of sight already exist. What does not exist yet is the permission, the economics, and the battery to fly them everywhere. Those three limits run on different clocks.

The aircraft that can fly beyond visual line of sight (BVLOS) have existed for years. The permission to fly them routinely, at scale, has not kept pace. In the United States, the FAA (Federal Aviation Administration) has been working on a dedicated rule for these flights, and in the meantime operators rely on case-by-case waivers and exemptions. Deadlines set along the way have slipped more than once.

That gap is the subject of this final chapter. Earlier posts followed an aircraft from constraints to design to first flight to production. This one looks at the three limits that no amount of good engineering removes on its own: regulation, cost, and the physics of storing energy. They are rarely discussed together, and they should be, because they do not move at the same speed, and they quietly depend on each other.

A capable aircraft is not a deployable product. Between the two sits a set of constraints that engineers do not control, and they share one awkward property: each one improves on a different timescale. Software and airframe improvements arrive in months. Regulation arrives in years, and not always on schedule. Battery chemistry arrives in decades, and often turns out smaller than the headline suggested. A product reaches the market only when all three have caught up with the same mission, which means the slowest one sets the date.

The second property is less obvious and more important. These limits are coupled. The rules decide how many aircraft one person may supervise, and that number is the biggest lever on cost per flight. The physics of the battery decides how far an aircraft can fly, and distance decides which regulatory category the mission lands in. Fixing one constraint in isolation tends to move the problem rather than solve it.

Engineering runs in months, regulation in years, battery physics in decades. A mission is only possible when all three agree, so the slowest clock, not the best aircraft, decides when it starts.

Picture three clocks on one wall, each ticking at its own speed. The engineering clock is the fast one: a new flight controller, a lighter frame, a better propeller can all land inside a quarter. The regulatory clock is slower and, importantly, not fully predictable. A rule goes from proposal to final text through comment periods, reviews, and political decisions that no flight test can accelerate. The physics clock is the slowest of all: the energy a kilogram of battery can hold improves in small increments after long research cycles, and promising chemistries often spend years between laboratory result and a product anyone can buy at scale.

A useful habit when judging any claim about drone capability is to ask which clock it assumes. A demonstration video assumes the engineering clock. A business plan promising nationwide delivery next year usually assumes the regulatory clock will cooperate. A promise of day-long endurance quietly assumes the physics clock has already advanced. Most disappointment in this industry comes from a claim that was true on one clock and silently assumed progress on another.

Start with the regulatory clock in the United States. The FAA's proposed Part 108 rule would create a standard path for BVLOS operations, but a rule moves through proposal, public comment, government review, and final publication, and deadlines set for that journey have slipped before. The sticking point is not paperwork. One proposed provision would give drones presumptive right of way over manned aircraft that are not broadcasting their position, which is a real safety question about who must yield to whom. This area changes quickly, so for the current status check the Federal Register notice and the FAA's own announcements rather than any article, this one included.

Europe runs the same clock in a different shape. EASA (the European Union Aviation Safety Agency) uses SORA (Specific Operations Risk Assessment), a structured method that scores an operation's risk and assigns it a SAIL (Specific Assurance and Integrity Level), and the guidance behind it is revised from time to time. At the top of the risk range, EASA's Certified category applies an approach it describes as very similar to manned aviation, which is where the Type Certificate and production approval split from post 12 comes in. Europe has a framework in use where the United States is still finishing one, but national authorities apply it, so the clock there is the number of countries, not only the number of months.

Cost is the second clock, and published figures should be read with care because they come from different sources, years, and definitions. One 2023 McKinsey estimate put a single-package drone delivery near $13.50, against $9 to $11 for electric or gas vehicles. Reporting from 2024 cited internal Amazon projections near $63 per package. Zipline's chief executive has said a flight could cost about $1 at scale, which that source presents as a target, not a demonstrated number, and the operator Manna claims roughly $4 per flight today. A spread that wide says something useful: per-flight cost is dominated by assumptions about how often an aircraft flies and how many aircraft one person supervises, far more than by the price of the hardware. That is the same fleet-scale prize this series described in post 11, and it is exactly why the regulatory clock leaks into the cost clock.

The figures above are not directly comparable, and one of them is a target. What they do show is that cost per flight swings by an order of magnitude on assumptions about utilization and supervision, not on the price of the drone itself.

The physics clock is the slowest. With lithium-ion packs at up to about 250 watt-hours per kilogram, and solid-state chemistries promising over 400, it is worth noting that solid-state remains expensive, slower to charge, and unproven for long-term reliability across drone platforms. Liquid fuels hold many times more energy per kilogram than any of these. As post 7 showed, adding battery mass does not buy endurance linearly, because the extra mass has to be lifted too. A better chemistry would help, but it arrives on its own schedule.

The part most commentary misses is the coupling. Regulation looks like an external delay that engineers simply wait out, but it quietly sets the economics. If one remote pilot may supervise one aircraft, delivery costs follow a labor curve. If one operator may supervise a fleet, the curve bends, and the swarm coordination from post 13 starts to matter commercially. The same rule that decides whether the mission is legal also decides whether it is affordable, so a slow rule is not only a delay, it is a cost.

It also deserves saying plainly that the delay is not only bureaucratic friction. The proposed Part 108 right-of-way provision is a genuine safety trade-off: whether a drone or a small manned aircraft should yield, and who bears the burden of being visible. Reasonable people disagree, which is part of why the clock is slow. Teams that treat regulation as an obstacle to route around tend to be surprised by it twice. Teams that read it as a statement of where the safety argument is still unsettled tend to design to the stricter of the two regimes early, so the evidence they need, traceable parts, documented processes, detect-and-avoid, exists before the rule does.

The first self-propelled road vehicles in Britain met this exact pattern. Britain's Locomotives Act of 1865, known as the Red Flag Act, required a crew of three, with one person walking at least 60 yards ahead of the vehicle carrying a red flag, and capped speeds at 4 miles per hour on open roads and 2 in towns. A later act in 1878 dropped the flag but kept the speed limits, and only in 1896 were light vehicles freed from the walker and the caps. For three decades the engine was ready and the permission was not. That walker with a flag is the visual observer of its day, and the story reads like the line between visual and beyond-visual-line-of-sight flight in post 11: the technology was never the constraint, the rule about who must be watching was.

The limits on UAS are less about the aircraft than about three constraints running on different clocks, and they are coupled.

• Engineering moves in months, regulation in years, and battery physics in decades, so the slowest clock sets when a mission becomes possible
• In the United States, BVLOS rulemaking has repeatedly run behind its deadlines, so waivers and exemptions carry much of the load in the meantime
• In Europe, a risk-based framework is in use, national authorities apply it, and the highest-risk flights fall under a Certified category modeled on manned aviation
• Per-flight cost depends far more on utilization and operator-to-aircraft ratio than on hardware price, and regulation sets that ratio
• Waiting on a rule is not wasted time if the team designs early to the stricter regime

The final topical post turns the question around: given all of these limits, when is a UAS simply the wrong tool, and what should be used instead?

Curious to exchange some ideas? Reach out via the contact form or connect on Linkedin!