The hidden cost of scaling AI: infrastructure, energy, and the push for liquid cooling.
Updated
January 8, 2026 6:31 PM
The inside of a data centre, with rows of server racks. PHOTO: FREEPIK
As artificial intelligence models grow larger and more demanding, the quiet pressure point isn’t the algorithms themselves—it’s the AI infrastructure that has to run them. Training and deploying modern AI models now requires enormous amounts of computing power, which creates a different kind of challenge: heat, energy use and space inside data centers. This is the context in which Supermicro and NVIDIA’s collaboration on AI infrastructure begins to matter.
Supermicro designs and builds large-scale computing systems for data centers. It has now expanded its support for NVIDIA’s Blackwell generation of AI chips with new liquid-cooled server platforms built around the NVIDIA HGX B300. The announcement isn’t just about faster hardware. It reflects a broader effort to rethink how AI data center infrastructure is built as facilities strain under rising power and cooling demands.
At a basic level, the systems are designed to pack more AI chips into less space while using less energy to keep them running. Instead of relying mainly on air cooling—fans, chillers and large amounts of electricity, these liquid-cooled AI servers circulate liquid directly across critical components. That approach removes heat more efficiently, allowing servers to run denser AI workloads without overheating or wasting energy.
Why does that matter outside a data center? Because AI doesn’t scale in isolation. As models become more complex, the cost of running them rises quickly, not just in hardware budgets, but in electricity use, water consumption and physical footprint. Traditional air-cooling methods are increasingly becoming a bottleneck, limiting how far AI systems can grow before energy and infrastructure costs spiral.
This is where the Supermicro–NVIDIA partnership fits in. NVIDIA supplies the computing engines—the Blackwell-based GPUs designed to handle massive AI workloads. Supermicro focuses on how those chips are deployed in the real world: how many GPUs can fit in a rack, how they are cooled, how quickly systems can be assembled and how reliably they can operate at scale in modern data centers. Together, the goal is to make high-density AI computing more practical, not just more powerful.
The new liquid-cooled designs are aimed at hyperscale data centers and so-called AI factories—facilities built specifically to train and run large AI models continuously. By increasing GPU density per rack and removing most of the heat through liquid cooling, these systems aim to ease a growing tension in the AI boom: the need for more computers without an equally dramatic rise in energy waste.
Just as important is speed. Large organizations don’t want to spend months stitching together custom AI infrastructure. Supermicro’s approach packages compute, networking and cooling into pre-validated data center building blocks that can be deployed faster. In a world where AI capabilities are advancing rapidly, time to deployment can matter as much as raw performance.
Stepping back, this development says less about one product launch and more about a shift in priorities across the AI industry. The next phase of AI growth isn’t only about smarter models—it’s about whether the physical infrastructure powering AI can scale responsibly. Efficiency, power use and sustainability are becoming as critical as speed.
Keep Reading
From plush figures to digital pets, a new class of AI toys is emerging — built not around screens or sensors, but around memory, language and emotional awareness
Updated
February 5, 2026 2:00 PM
Spielwarenmesse toy fair. PHOTO: SPIELWARENMESSE
Spielwarenmesse in Nuremberg is the global meeting point for the toy industry, where brands and designers preview what will shape how children play and learn next. At this year’s fair, one message stood out clearly: toys are no longer built just to entertain, but to listen, respond and grow with children. Tuya Smart, a global AI cloud platform company, used the event to show how AI-powered toys are turning familiar formats into interactive companions that can talk, react emotionally and adapt over time.
The company’s central argument was simple but far-reaching. The next generation of artificial intelligence toys will not be defined by motors, sensors or screens alone, but by how well they understand human behavior. Instead of being single-function objects, smart toys for children are becoming systems that combine language models, emotion recognition and memory to support ongoing interaction.
One of the most talked-about examples was Tuya Smart’s Nebula Plush AI Toy. At first glance, it looks like a soft, expressive plush figure. Inside, it uses emotional recognition to change its LED facial expressions in real time. If a child sounds sad or excited, the toy’s eyes respond visually. It supports natural conversation, reacts to hugs and touch and combines storytelling, news-style updates and interactive games. Over time, it builds memory, allowing it to behave less like a gadget and more like an interactive AI toy that recalls past interactions.
Another example was Walulu, also developed using Tuya’s AI toy platform. Walulu is an AI pet built around personalization. It can detect up to 19 emotional states and speak more than 60 languages. It connects to major large language models such as ChatGPT, Gemini, DeepSeek, Qwen and Doubao. Through simple app-based controls, users choose traits like cheerful, quiet, curious or thoughtful. Those choices shape how Walulu talks and reacts. Instead of repeating scripts, it adjusts its tone and behavior over time. The result is not a novelty item, but an emotionally responsive AI toy that feels consistent in daily use.
Tuya also showed how educational AI toys can extend into learning and exploration. Its AI Learning Camera blends computer vision with interactive content. When it recognizes an object, it links it to cultural and learning material. If a child points it at a foreign word, it offers real-time pronunciation and translation. It can also turn drawings into digital artwork, encouraging active creativity rather than passive screen time. In this sense, AI toys for kids are becoming tools for learning as much as play.
These products point to a larger strategy. Tuya is not just making toys — it is building the AI toy development platform behind them. Through its AI Toy Solution, developers can design a toy’s personality, memory logic and behavior without training models from scratch. The system integrates with leading AI models and supports multi-turn conversation and emotional feedback, turning standard hardware into responsive AI companions.
The platform supports multiple development paths. Brands can use ready-to-market OEM solutions, add AI to existing products or build custom toys around their own characters. Plush toys, robots, educational tools and wearables can all become AI-powered toys without changing their physical design.
Because these products are made for children and families, safety is built in. Tuya’s system includes parental controls, conversation history review and content management. It supports standards such as GDPR and CCPA with encryption and data localization.
From a business standpoint, Tuya’s pitch is speed and scale. The company says its AI toy infrastructure can cut development time by more than half and reduce R&D costs by up to 50 percent. Its AIoT network spans over 200 countries and supports more than 60 languages, making global deployment of AI toys easier.
What emerged at Spielwarenmesse 2026 was not just a lineup of smart gadgets, but a clear shift in the category. AI toys are evolving into emotionally aware systems that talk, listen, remember and adapt. Their value lies not in sounding clever, but in fitting naturally into everyday life.
The fair did not present AI toys as a distant future. It showed them as products already entering the mainstream. The real question now is not whether toys will use AI, but how carefully that intelligence is designed for children.