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
METiS TechBio’s blockbuster IPO signals rising investor interest in AI startups focused on how drugs are delivered inside the body
Updated
May 14, 2026 3:02 PM
HIV-1 virus particles, coloured red. PHOTO: UNSPLASH
Investors are beginning to place bigger bets on AI startups focused on drug delivery infrastructure rather than drug discovery alone. That shift was on display this week after METiS TechBio, a Hong Kong tech-bio startup focused on AI-powered drug delivery systems, debuted on the Hong Kong Stock Exchange.
The listing made METiS TechBio the world’s first publicly traded AI-powered drug delivery startup and the first AI-powered large-molecule biopharmaceutical startup listed in Hong Kong. The startup raised more than HKD 2.1 billion through its IPO, making it the largest healthcare listing in Hong Kong so far in 2026.
Investor demand was unusually strong. The Hong Kong public offering was oversubscribed by more than 6,900 times while the international tranche recorded 82 times oversubscription. More than 280 institutional investors participated in the international placing.
The strong demand reflects a wider shift in AI biotech. Over the past few years, much of the sector’s attention has focused on using AI to discover new drugs or molecules. METiS is taking a different approach. The startup focuses on how medicines are delivered inside the body after they are developed.
That challenge is becoming harder to ignore in biotech. Designing a therapy is only one part of the process. Delivering it precisely to specific organs, tissues or cells remains a major hurdle, especially for newer therapies involving RNA, proteins and large-molecule drugs.
METiS is trying to solve that problem through its proprietary NanoForge platform. The system uses AI to design and test nanodelivery systems that help medicines reach targeted areas inside the body more efficiently. The platform combines AI models, simulation systems and high-throughput screening tools to speed up formulation development and improve delivery precision.
The startup says it has already achieved targeted delivery across eight organs and tissue systems including the liver, lungs, heart, muscles and central nervous system.
One of its lead programs, MTS-004, became China’s first AI-enabled formulation drug to complete a Phase III clinical trial. The drug is being developed for pseudobulbar affect, a neurological condition that affects emotional expression. According to the startup, AI tools helped reduce preclinical formulation development time from up to two years to less than three months.
Investor interest in the IPO also came from some of the world’s largest asset managers and healthcare funds. BlackRock led the cornerstone investments with a USD 50 million subscription. Other participating investors included UBS Asset Management Singapore, Mirae Asset, ORIX Corporation, Deerfield, RTW, Hillhouse Capital and IDG Capital.
METiS is also building what it describes as a “platform collaboration + product partnership” business model. The startup currently works with more than 30 pharmaceutical and biotechnology partners globally, including large pharmaceutical companies and medical research institutions.
The company reported RMB 105 million in revenue in 2025, largely tied to upfront payments connected to its MTS-004 partnership agreements. It also said some platform collaboration contracts could reach milestone values of up to USD 109 million.
Chris Lai said: "The future of biomedicine will no longer be simply about 'taking medicine when one falls ill.' METiS TechBio's ambition is to harness AI to build nano-rockets that can navigate with precision through the inner space of the human body's 30 trillion cells, write the code of nucleic acids and proteins into cells, and reprogram diseased and aging cells into healthy cells. This was our founding aspiration, and it is the mission to which we will dedicate our lives. The IPO marks a new starting point for us to accelerate forward, and we will strive to live up to the support and trust we have received from all sectors."
The IPO also highlights how Hong Kong is increasingly positioning itself as a hub for next-generation biotech and AI healthcare startups. While investor excitement around AI drug discovery has cooled in parts of the market, startups focused on delivery systems and biotech infrastructure are beginning to attract stronger institutional backing.
For METiS, the challenge now will be turning that investor confidence into commercially viable therapies and long-term partnerships. But the listing suggests that AI-driven drug delivery is starting to emerge as a category investors are willing to treat as core biotech infrastructure rather than a niche research experiment.