The AI landscape doesn't move in one direction — it lurches. Some techniques leap from experiment to table stakes in a single quarter; others stall against regulatory walls, technical ceilings, or organisational inertia that no amount of hype can dislodge. Knowing which is which is the hard part. The State of Play cuts through the noise with a rigorously maintained index of AI techniques across every major business domain — classified by maturity, evidenced by real-world adoption, and updated daily so you always know where you stand relative to the field. Stop guessing. Start knowing.
A daily newsletter distilling the past two weeks of movement in a domain or two — delivered to your inbox while the index updates in the background.
Each dot marks the weighted maturity of practices within a domain — hover for a brief summary, click for more detail
AI systems that navigate and operate in the physical world across road, air, marine, rail, and space. Heavily leading-edge: ADAS and warehouse robotics are good practice, but full autonomy in open environments remains elusive. Evenly split between advancing and stalled — regulatory frameworks and edge-case safety are the binding constraints. One of the largest domains with 22 practices spanning a wide maturity range.
The headline: Autonomous vehicles are now in real commercial service across roads, rails, and oceans — but operating at scale and operating profitably remain two different problems, and no one has solved both.
Self-driving technology works. That is no longer the question. Waymo runs 500,000+ paid taxi rides a week across 11 American cities. Baidu does 250,000+ driverless rides weekly across 26 Chinese cities. Japan launched the world's first fully autonomous container ship on a scheduled commercial route. Nearly 2,000 kilometers of driverless metro track operate in China alone. The question now is whether any of this makes money — and who bears the liability when it goes wrong. Most companies evaluating autonomous systems will find proven technology available in defined settings (mining, metro rail, highway trucking corridors, drone inspection), but the economics and legal frameworks outside those niches remain unsettled. Organizations already using autonomous mining trucks or drone inspection are getting documented returns. Everyone else faces a narrowing window to build internal capability before vendor lock-in and regulatory complexity make late entry expensive and disruptive.
Japan's autonomous container ship entered scheduled commercial service. GENBU, a 134-meter vessel, now runs a fixed Kobe-Tokyo coastal route with full regulatory certification — the first Level 4 autonomous ship operating commercially anywhere in the world. Samsung Heavy Industries simultaneously partnered with Orca AI to install autonomous navigation as standard on newbuild vessels, signaling the shift from aftermarket retrofit to factory-integrated autonomy. If your organization ships goods by sea, autonomous navigation is moving from optional upgrade to default specification on new vessels.
Autonomous rail crossed a procurement threshold. Copenhagen confirmed a EUR 3 billion contract for 226 driverless trains; Dublin specified fully autonomous operations in its 25-year metro tender; India tendered driverless trains with AI-based predictive maintenance for Lucknow. Driverless metro systems are now the default specification for new urban rail, not an experimental choice. Rail operators still running manual systems on new lines face growing cost disadvantage.
L4 robotaxi competition widened — and so did the safety record. Zoox partnered with Uber for Las Vegas and Los Angeles deployment; Nuro secured a California driverless permit; Verne launched the EU's first commercial robotaxi in Zagreb at EUR 1.99 per ride. But Avride logged 16 crashes in four months in Dallas (NHTSA investigation opened), and Tesla disclosed its Cybercab crashes at four times the human rate. More operators entering the market means the gap between best-in-class and new entrants is becoming a safety and reputational risk for the entire sector.
Baidu's fleet-wide cloud outage exposed centralized architecture risks. Over 100 robotaxis froze simultaneously on Wuhan highways, stranding passengers for two hours after a cloud dispatch failure. Chinese regulators froze new robotaxi permits indefinitely. Any organization depending on cloud-connected autonomous systems — vehicles, drones, port equipment — should be asking vendors about single-point-of-failure resilience.
Mining autonomy expanded beyond metals. Heidelberg Materials (cement and aggregates) is scaling from 1 site to 6 with 100+ autonomous vehicles targeted by 2028. Third-party software platforms proved they can run mixed fleets from different manufacturers without equipment replacement, challenging the OEM lock-in model that has defined the sector for two decades.
California's new autonomous truck permits take effect. The state lifted its ban on driverless heavy trucks in April 2026, with a three-tier permitting process requiring 500,000 miles of testing per phase. Freight operators with California exposure should begin evaluating autonomous corridor strategies now — the Sun Belt trucking routes are filling up with competitors.
The IMO MASS Code goes live (non-mandatory) in 2026, mandatory by 2032. The international regulatory framework for autonomous ships is finalized. Shipping and port organizations have a six-year window to prepare for mandatory compliance. Those building new vessels today should specify MASS-ready systems architecture.
Battery certification will gate urban air mobility timelines. Joby and Archer both target 2026 commercial launch, but the gap between required and available battery energy density is a five-year problem. Do not budget for eVTOL (electric vertical takeoff and landing aircraft) operations before 2028 at the earliest unless you are in a pilot program.
Technology works but economics don't — yet. Waymo's near-term finances are flagged as weak by independent analysts despite 500,000+ weekly rides. Kodiak raised $100 million at a steep discount to market price. The autonomous systems that demonstrably pay for themselves today are mining haulage, drone inspection, and metro rail — all operating in controlled, high-value environments. Open-road autonomy has not proven fleet-level profitability anywhere.
Supply chains are a strategic vulnerability. China controls 90 percent of rare-earth processing and 99 percent of the lithium-ion battery cells that autonomous systems depend on. The U.S. ban on DJI removes the 80-percent market leader in commercial drones with no domestic price-competitive alternative. These are not short-term disruptions — they constrain what you can deploy, from whom, and at what cost.
Liability assignment is unresolved across every sub-sector. When an autonomous truck causes an accident during a handoff, who pays — the manufacturer, the fleet operator, or the software provider? No jurisdiction has a complete answer. The $243 million Tesla Autopilot verdict (design defect and failure-to-warn) is the closest legal precedent, and it applies only to Level 2 systems. Level 3 and 4 liability frameworks remain legally undefined.
Go deeper: the full Autonomous Systems & Vehicles briefing — the longer analytical write-up, plus every practice we track in this domain with its maturity rating, the tools to consider, and the evidence behind our assessment.