![\"\"](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/05\/iStock-2255483268.jpg\")
<\/p>\n<\/p>\n
Singapore prides itself on being unbreakable. Then 2020 happened. A pandemic was bad. A pandemic plus citywide flooding was worse. Emergency services failed to reach people who needed them. The world\u2019s gold-standard smart city had just discovered its expensive seams.<\/p>\n
Felix Tan watched it happen and started rebuilding what his company was.<\/p>\n
\u201cThere were still many areas that the agencies and their data and the services\u2026 are not coordinated,\u201d he says now. Tan is the chief executive officer of\u00a0Nuvola Media<\/a>. This Singapore firm began in signal management and has spent the last few years rebuilding itself into something stranger: a translator between the autonomous AIs that smart cities are now buying by the truckload.<\/p>\n This is the part where chief data officers should take notice.<\/p>\n The smart-city illusion<\/p>\n Tan\u2019s diagnosis of the industry is uncomfortably specific. Every agency runs pilots. Each pilot stands up its own agentic AI: one for traffic, one for water, one for HVAC, and one for the weather bureau. Each is sold as autonomous. Each is, in Tan\u2019s words, working \u201conly within their role of responsibility.\u201d None of them are really talking.<\/p>\n The result across many cities he has audited is the same: \u201cthey actually fail to integrate data across the multiple agencies.\u201d<\/p>\n That is the story your smart-city RFP isn\u2019t telling you. You are not buying a city brain. You are buying a row of clever, well-funded, self-absorbed monologues.<\/p>\n Why Bangkok matters<\/p>\n Tan\u2019s preferred parable is the flood. Bangkok\u2019s, specifically. Ask his audience whether a flood is a water problem, and he lets the silence sit a beat. \u201cIs it just a water problem? No.\u201d<\/p>\n Floods stall traffic. Traffic stalls logistics. Logistics stalls GDP. Floods drown crops, which moves commodity prices, which then moves markets. Floods cost lives. One event, five ministries \u2014 and your agentic AI stack is five solo performers in five different keys. The synoptic layer Nuvola sells is essentially the connective tissue: protocol-agnostic plumbing that lets siloed agents share state without anyone having to rip out their pet system.<\/p>\n CDOs should care because their organizations are running the same play on a smaller scale. Stand up another autonomous agent. Add another model. The fragmentation is not a phase. It is the architecture.<\/p>\n Proof, Keppel Bay edition<\/p>\n The case study on Tan\u2019s pitch deck is hard to argue with. Nuvola\u2019s platform helps operate\u00a0Keppel Bay Tower<\/a>, Singapore\u2019s first zero-energy commercial building \u2014 an 18-story office tower that has become a working museum for smart-building tech.<\/p>\n The headline number is interesting: \u201c8% savings for the HVAC system\u2026 year on year.\u201d The mechanism is delightful. Nuvola wired the building\u2019s meeting-room booking system into its platform. The platform learned that Felix Tan likes his meetings at 20\u00b0C, three degrees below the building baseline. So now it begins, lowering the room temperature twenty minutes before he arrives. He walks in. It is perfect. He walks out. It resets.<\/p>\n Multiply that across 18 floors, an autonomous security robot fleet and a re-purposed CCTV grid, and you start to see his point about choreography over intelligence. The platform is not smarter. It is better at agreeing with itself.<\/p>\n The surveillance question<\/p>\n Then there is CCTV. Tan has thoughts. The kindest: \u201cmost CCTV implementations today\u2026 just a backup plan\u201d \u2014 recordings nobody watches until something explodes. Computer vision changes the economics. The same dumb camera becomes a people counter, a footfall heatmap and a thermal-comfort signal that feeds into HVAC decisions.<\/p>\n It also becomes a face-recognition system, which is the part Wired readers spot immediately. Tan is more careful here than the average smart-city vendor. He respects each jurisdiction\u2019s law \u2014\u00a0Singapore\u2019s PDPA<\/a>, say \u2014 and then goes further. \u201cWe work with the customer to draw the line, where do we stop?\u201d<\/p>\n It is a less satisfying answer than \u201cwe never enable surveillance.\u201d It is also more honest. Anyone selling video analytics into 2026 cities is choosing where the creep factor lives. Pretending otherwise is a dishonest move.<\/p>\n The model<\/p>\n Commercially, Nuvola is the kind of SME that big-vendor smart-city RFPs are designed to box out. Its survival strategy is to sit where megavendors can\u2019t comfortably sit \u2014 in the boring data-exchange layer that everyone else\u2019s product depends on. SaaS for the easy cases. On-prem for the military and government clients who would rather die than touch a cloud. Hybrid for the pragmatic. Regional partnerships with\u00a0SoftBank<\/a> and\u00a0Yokogawa<\/a> in Japan, JLL and CBRE in Thailand. Tan has lost projects on budget. He claims to have never lost one in capability.<\/p>\n The next three years<\/p>\n Ask Tan what the city operations stack looks like in 2029, and he gives you the right answer for a vendor: routine cross-agency coordination, the Bangkok flood scenario actually choreographed. Ask him what stays human, and he gives you the better one. Someone still has to replace the dead LED. Someone has to repair the stalled robot. Someone has to remember why the building exists in the first place.<\/p>\n \u201cWe are not completely Skynet.\u201d<\/p>\n This is the actual takeaway for the CDO crowd: your next AI strategy is probably stacking autonomous agents on top of a data plane that does not yet exist. The unglamorous middle layer \u2014 the one no vendor brochure leads with \u2014 is where smart cities either work or quietly burn another decade of capex.<\/p>\n Mirror your world, Nuvola\u2019s tagline insists. Or just keep paying for five monologues. Pick.<\/p>\nFelix Tan @ Nuvola Media audited many smart cities and found the same failure: there is little to no data integration across departments.<\/p>\n