SURFACE · THE ROBOTICS JOURNEY, PART 2. By Richard Hurley, Chevron Technology, Projects & Execution, and Viktor Klein, 8scale. Twenty pages, free, no form.
Download the reportOVERVIEW
A drone in a box earns its place at the moment a site stops treating it as equipment and starts treating it as infrastructure. Part 2 of The Robotics Journey, written by Richard Hurley of Chevron's Technology, Projects & Execution team with our analyst Viktor Klein, is an argument about that threshold and its price. What follows is our reading, including where we expect adoption to stall.
The article opens on a measurement rather than a promise. "Human availability is now the primary constraint on how quickly and how often drones can be used in energy operations." When inspection depends on pilots, travel and on-call crews, frequency is capped and coverage cannot scale across an asset network.
The article documents the condition with sources. PHMSA reports that half of U.S. pipelines are at least 50 years old and that the average U.S. refinery is 74 years old since original commissioning. Less than one percent of field inspection data exists in a form decision makers or analytics systems can use.
Page 8 is the most reusable in the report. Five axes in one frame, each scored 1 to 5 against a written guideline: safety, productivity, maturity, economics and code compliance. Safety follows the hierarchy of controls, productivity is expressed as outage reduction, economics as payback and IRR.
Applied to conventional flare monitoring, the method gives the incumbent its due. Maturity scores 5, because manual observation and fixed instrumentation are fully mature. Code compliance scores 5, because Method 22 observations align directly with the regulatory construct. Safety scores 3, because a person has to be present and attentive near operating areas. Productivity and economics score 2, for a five minute observation that escalates into a two hour watch once smoking is detected, and for the cost of documenting it.
A rating that hands the old method two perfect scores is the only kind worth quoting when it marks the same method down elsewhere.
OVERVIEW
OVERVIEW
Work at heights carries the clearest safety case, with 421 fatal falls recorded in U.S. construction and maintenance in 2023, and the clearest cost case, where the article models piperack inspection falling from an estimated $175 per meter toward $15. Pipelines carry the volume case, with PHMSA requiring right of way surface monitoring at least 26 times a year. Methane carries the compliance case, where drone sensing is defensible as a detection and localisation layer between satellites and ground crews. Upstream operations carry the routine case, where imagery replaces part of the look and report round on well pads.
None of these four reliably carries a dock on its own. The arithmetic of the report rests on stacking them on the same hardware, and that is the finding for a steering committee.
For a mid sized U.S. refinery with two docking stations, the article models initial setup at $350,000 to $625,000 and annual operating cost at $60,000 to $120,000. The conservative case sums roughly $250,000 in gross annual savings across five stacked missions, leaving about $130,000 net and payback in three to four years. It excludes avoided downtime at a modelled $125,000 per hour, scaffolding mobilisations above $20,000, and the value of the inspection data itself. Weighted in, the article models two to three times that value and payback in 12 to 18 months.
Every figure above is an assumption printed as an assumption, with the calculation logic in the sources chapter. That is what makes the page arguable, and a business case nobody can argue with is a business case nobody can approve.
OVERVIEW
OVERVIEW
The regulator is not the wall. The proposed FAA Part 108 would allow low altitude BVLOS without individual waivers, EASA already issues operational authorizations in the Specific category, and both ask for the same four building blocks. The report states the conclusion directly: "The barrier to adoption is no longer technological, it is organizational."
Our reading is narrower. Adoption stalls in procurement. As long as inspection is bought in hours, a technology that removes hours reduces the supplier's revenue while the buyer's savings sit in avoided costs that never appear on an invoice. The report asks asset owners to bid outcomes, meaning uptime and verified inspections. That change is slower than any waiver and it decides the pace of everything else.
One thing worth doing this quarter. Take the five missions above, count how many your largest site would stack on one dock, and name the person accountable for the flight, the data and the work order. If the answer is fewer than three, the case is not ready. If nobody owns the second question, it will not survive the first incident.
THE REPORT
The full report is on this page, twenty pages, free and without a form. It carries the five axis scoring guideline, the four use case families with their asset class numbers, the complete business case with its assumptions, and the sources and methodology chapter behind them. Published with the permission of Chevron Corporation. © 2026 Chevron Corporation.
Richard Hurley is Senior Robotics Engineer in Chevron's Technology, Projects and Execution organisation. Viktor Klein is the named analyst at 8scale, the market intelligence layer between industry, technology and innovation, and the publisher of Scalebook.
Special thanks to Vasu Pillai, Zach Pringnitz and Trey Caylor (Chevron); Jackie Alkobi, Katherine Snyder, Sam Jones and Ariel Avitan (Percepto); John Rasch (Skydio); Eileen Lockhart (Acuren); Philipp Eder (Swisscom Broadcast). Valuable input from Nokia, ETH Zurich, Hextronix, TEAM Industrial, Energy Drone & Robotics Coalition and Sprint Robotics.
For the market behind these use cases, see Scalebook 2025, Inspection & Maintenance Robotics.
OVERVIEW
OVERVIEW
This report may contain forward-looking statements, predictions, or opinions about future technological and market developments. These statements are based on assumptions and market signals available at the time of writing. Actual outcomes may differ due to risks and uncertainties. All figures, market estimates, and forecasts are based on proprietary analysis and publicly available sources at the time of publication.
cite as: Hurley, R. and Klein, V. (2026), The Robotics Journey Part 2: drone in a box, Chevron / 8scale
