Smart city innovation is surging worldwide, driven by rapid urbanisation, AI and Internet of Things connectivity. Major metropolitan areas across the globe are deploying intelligent infrastructure to transform urban living by optimising traffic flow, reducing carbon footprints and improving public services.
The low-altitude economy, centred on operations by unmanned drones and electric vertical take-off and landing aircraft typically below 1,000-3,000 metres, acts as a vital catalyst for smart city innovation.
In recent years, the Hong Kong SAR Government has been stepping up efforts to drive innovation in smart city management by establishing a high-level Working Group on Developing Low-altitude Economy following the 2024 Policy Address.
Relevant measures include exploring low-altitude flying application scenarios, while pressing ahead with pilot projects and designated venues to explore deploying drones for building maintenance, among others.
Hong Kong’s Alpha AI Technology is supporting these efforts. Its vision is to build the spatial intelligence layer for the next generation of drones and robots, starting with making building and infrastructure inspection safer, faster and more affordable.
Alpha AI started with drone-based external-wall and façade inspection. From there, it built a proprietary spatial AI platform and became the first Hong Kong start-up to secure a Beyond Visual Line of Sight license under the Hong Kong SAR Government’s Low-altitude Economy Regulatory Sandbox in 2025, deploying automated AI drones for infrastructure and spatial analysis.
Rapid growth
Since last year, the Cyberport-incubated start-up has delivered 200+ projects across government, property and construction, quadrupling its revenue growth in two years.
Its biggest challenge was convincing a highly regulated industry to take AI seriously. “Building inspection, surveying and maintenance are standards-driven fields,” observed Co-Founder and CEO Desmond Ho, “so it isn’t enough to build clever AI. The AI has to meet the expectations of professional surveyors, engineers and regulators.”
Those expectations required three things: real domain data, a deep understanding of each sector and constant product refinement to match how the industry works.
Alpha AI had to build all of this from scratch, designing its own algorithms and effectively teaching AI case by case. Moving from one asset type to another – from external building façades to confined spaces, bridges or slopes, for example – the environment, defects and standards changed, and models and workflows needed adjustment.
“The upside is scalability: once we solved it in Hong Kong, the same spatial AI approach transfers readily across asset types and across cities because the underlying 3D understanding carries over, even when local standards differ,” Mr Ho remarked.
The Hong Kong Trade Development Council (HKTDC), particularly its Start-up Express entrepreneurship development initiative, will assist with Alpha AI’s continued growth, as the programme has done for 80 other start-ups since 2018.
“We’re proud to have been selected as a Start-Up Express winner,” said Mr Ho. “The programme is built to help Hong Kong start-ups explore overseas markets, meet partners and investors and raise brand awareness. The marketing package, the start-up zones at HKTDC’s local exhibitions, the overseas flagship events and study missions, and the mentorship all directly support our push into the GBA and ASEAN.”
The drone is just one of the company’s tools, he added. “Our spatial AI turns 2D video footage into a real-world 3D coordinate map. Any defects are pinpointed in space to the exact part of the building, slope or structure.”
An added memory layer means the AI remembers each location and tracks how each defect evolves over time, comparing one inspection to the next to reveal progression. This enables predictive, condition-based maintenance rather than one-off snapshots.
Over the long term, Alpha AI is building this spatial-plus-memory foundation into a world model for physical AI, so the same intelligence extends from drones to robot dogs and humanoid robots, and from one-off inspection to continuous, autonomous operation.