Spatial Web: Transforming Device Interactions

When it was invented in 1991, the World Wide Web connected together an Internet that was overrun with many thousands of individual, fragmented digital documents. HTML, hypertext markup language, represented a daring leap. It combined the age-old idea of hypertext with the Internet’s global reach. Tim Berners-Lee’s new language offered up a lingua franca for interconnected information.

Today, following the social media revolution, a new phase of the Internet is emerging. The Spatial Web promises to connect together a physical world-full of devices, phones, wearables, robots, drones, and even AI agents. In May, the IEEE Standards Association [who shares a common parent organization with IEEE Spectrum] ratified a set of standards (IEEE 2874-2025) that defines the Spatial Web.

The original World-Wide Web introduced the idea of URLs that point to HTML files, which are accessed remotely via the HTTP standard.

Now the Spatial Web puts forward a new set of defining principles. HSML (hyperspace modelling language) behaves like nouns and verbs on the Spatial Web, describing what an entity is and what it does. HSTP (hyperspace transaction protocol) behaves like the Spatial Web’s grammar, defining how each entity functions and how it can interact with others. And the UDG (universal domain graph) acts as the directory that keeps track of every entity along with its activities and relationships. [See chart.]

We’ll come to some of the practicalities of the Spatial Web a little later. (For instance, where these various files might be stored, and how each entity can stay up to date with everything else in its network.) But for now, let’s first consider a few specific ways devices and AI agents can use the Spatial Web—via HSML, HSTP, and UDG standards—to more seamlessly interact with people, objects, and physical spaces.

EcoNet Gives Thermostats and Batteries the Power to Make Deals

Verses, the Los Angeles–based AI company where we work, recently collaborated with researchers at University College London on a project called EcoNet, a test home where two AI agents—one controlling a thermostat, the other a wall-mounted energy storage battery system—worked together to keep the space comfortable while saving money and cutting emissions.

Every ten minutes, the AI agents evaluated 729 possible strategies to balance comfort, cost, and carbon footprint. The thermostat prioritized occupant safety and warmth. The battery agent focused on charging during off-peak hours and using stored energy during expensive periods. It used HSML to describe a set of competing goals. One goal involved keeping the living room between 22 and 25 degrees Celsius. A second goal involved avoiding discharging the house’s energy storage below 50 percent during peak hours. Here’s how that looks in HSML code:

  A new Spatial Web standard defines Hyperspace Modeling Language (HSML), above, which helps define how devices can interact with the physical world.Spatial Web Foundation

The Spatial Web’s shared digital network protocol, the UDG, helped the smart devices work together in real time. Then, its secure communication method (via the HSTP standard) enabled decisions that followed the system’s rules and commands. The system adjusted automatically to changing weather conditions and energy prices—and cut both energy costs and carbon emissions by 15 to 20 percent. Verses demonstrated EcoNet in March at the 2025 AI UK exhibition at the Turing Institute in London.

At scale, an EcoNet-like architecture might enable entire neighborhoods to act a little like intelligent organisms, optimizing collective energy use and accelerating the shift to a more resilient, renewable grid.

Coordinated Mobility Standards Show Autonomous Vehicles the Way

When an ambulance rushes to an emergency, the ambulance driver still depends on surrounding traffic to notice and react to the siren. But autonomous vehicles may not know which direction the ambulance is coming from or how to properly respond to the ambulance in time, because autonomous vehicles operate without shared context.

The Spatial Web can address this shortcoming via HSML. A shared HSML document describes the state and relationships of things in a given neighborhood or at a given intersection. Properties being recorded in the HSML document might include the color, location, and behavior of a given traffic light.

With this shared context, an ambulance can issue a Spatial Web query like “find all autonomous vehicles and traffic infrastructure within 200 meters of my route.” Using the HSTP, it can request green lights, reroute cars, and alert pedestrians through connected devices.

How Drones Can Use HSML to Read the Same Map

Altitude limits, flight windows, and no-fly zones for drones today are difficult to enforce, in part because most drones follow static rules coded at the factory. They cannot respond to changing conditions or dynamic policies.

The Spatial Web provides drones with the necessary context to navigate responsibly. Regulators can use HSML to define constraints like “no flights above 120 meters after sunset and within 500 meters of a hospital.” Those constraints would then be published to the UDG, where drones operating within the relevant airspace can apply these constraints in real time.

Before take-off, a drone might issue a Spatial Web query such as “What restrictions apply to my delivery route?” HSTP allows it to confirm its airspace authorization, share its intended path, and adjust mid-flight if conditions or regulations change.

The same Spatial Web infrastructure can also be used in emergencies. After a natural disaster, drones could be temporarily authorized to enter restricted zones to assist with search and rescue or deliver supplies—all within a secure, trackable framework.

Lunar Rovers Will Bring the Spatial Web to the Moon

Coordinating autonomous systems in the air is difficult. In space, it is even more difficult. NASA’s Jet Propulsion Laboratory frequently collaborates with multiple agencies, universities, and contractors, with each using different simulation environments and proprietary platforms. Testing how multiple teams and rovers will one day cooperate on the Moon requires a shared language and a common model of the rovers and environment. The Spatial Web makes this possible.

In one demo, rover teams from The Jet Propulsion Laboratory in Pasadena, Calif. and California State University, Northridge each operated their own digital twin and simulation environments using HSML to coordinate a simulated lunar rescue. When one virtual rover got stuck in a crater, HSML allowed the stuck rover to send out real-time geometry, sensor observations, and activity data to the other rovers nearby. The virtual rovers also shared internal models from different physics modeling engines, including parameters like position, velocity, acceleration, and mass. The rover simulation, in other words, demonstrated how HSML-powered digital twins can assist in autonomous collaboration over challenging environments—even on the (virtual) Moon.

Digital Orchards Use the Spatial Web for Zero-Waste Supply Chains

Roughly one-third of global produce spoils before it ever reaches a plate, driving up emissions, reducing profits, and contributing to global hunger.

However, using Spatial Web standards, for instance, a peach orchard could use HSML to describe the ripeness, temperature, and shelf life of each crate. These descriptions are published to the local UDG, where retailers can query live inventory across regions. Using the Spatial Web, a buyer might query their local network, “What peaches are ready to harvest within 500 kilometers and meet my freshness criteria?”

HSTP can simplify the negotiation, delivery, and policy verification of such a query. If a buyer rejects a shipment, the grower can redirect it to a new buyer, such as a juicer or a nearby store, before the fruit goes to waste.

Instead of rigid logistics and guesswork, Spatial Web supply chains have the potential to become more adaptive, intelligent, and responsive to both external demand and internal conditions. The result will be less spoilage, better margins, faster payments, and fresher food.

The Road from Protocol to Practice

The Spatial Web Standard is still in a very early phase. HTML was published in 1991, but the first browser didn’t arrive until 1993. Additional Web standards on top of that, like cascading style sheets (CSS), didn’t come in until 1996. IEEE 2874 is similarly rolling out in stages. Ultimately the foundation we are laying in place paves the way for a Spatial Web that spans not so much pages and data files, but rather people, places, and things.

Standards succeed only when they disappear into the background. No one thinks about TCP/IP standards when reading email, although email relies on these standards in every message that is sent or received. Similarly, no layperson will need to understand how standards like HSML, HSTP, or UDG work. These components of the Spatial Web will all simply, like other protocols and standards before it, just do the hard communication and computation work behind the scenes.

Where, then, do HSML, HSTP, and UDG assets ultimately reside? Do they all sit on some cloud server somewhere? Or perhaps are these various digital files all scattered across individual devices and Internet of Things nodes?

Unfortunately, there is no single answer to these pertinent questions. On the other hand, the World-Wide Web didn’t launch fully formed either. Its earliest days often tested out trial implementations of new standards and technologies—because nothing like the truly widespread, instantaneous, global scale of the Web had ever been rolled out before.

For the Spatial Web, simple agents like IoT devices, for instance, could host HSML files and other Spatial Web assets on-device. In more complex settings, like smart cities or industrial systems, cloud servers or shared storage systems would provide a more remote and cloud-based kind of HSTP, HSML, and UDG deployment.

But no matter the Spatial Web implementation, whether fully remote or fully localized, cybersecurity will remain a key priority. HSML, HSTP, and UDG standards embed identity, access, and policy enforcement, via decentralized identifiers. Furthermore, the HSTP standard ensures that all transactions can be signed and auditable.

Ultimately, too, another aspect of any Spatial Web deployment will be the registries that must scale to manage billions of entities and agents. That is a larger, later-stage question to be tackled, no doubt, in future implementations of the Spatial Web. Nevertheless, even in the Spatial Web’s earliest incarnations today, we have already abstracted these complex concerns behind a secure, standards-based interface.

The standards that defined the World-Wide Web connected information. The Spatial Web will begin to interconnect the physical world and the many devices and AI agents operating in it. And with the new Spatial Web standards—and trial runs in homes, streets, skies, and on the (virtual) Moon—an increasingly interconnected Spatial Web future is no longer theoretical. A standardized Spatial Web is today as actual, and as actualizable, as HTML.