该报道于2026年3月10日（GMT+8 06:00）发布，指出澳大利亚初创企业Cortical Labs与新加坡DayOne Data Centers Ltd.合作，在墨尔本推出首个生物数据中心，并在新加坡建设第二个。两处设施计划使用CL1生物计算单元，以人脑细胞驱动，而非传统服务器架构中的处理器芯片。文章强调，尽管该公司称其计算能力目前仍然有限，距离与主流AI芯片（如Nvidia Corp.）形成直接竞争仍需数年甚至数十年，但该布局正对应全球AI数据中心快速扩张所带来的可量化压力：电力需求上升、用水压力上升，以及硅芯片供应短缺。
该文提到的技术路径具有明确的能力递进迹象：该公司先前已让其脑细胞系统学习玩朴素游戏Pong，近期又训练其运行更复杂的Doom。Cortical Labs使用由干细胞培养的人脑神经元，并将其置于芯片上，通过芯片向细胞发送并接收电信号，随后软件记录细胞反应并将其解释为计算输出。这一“输入-响应-解码”闭环表明其目标不仅是替代式试验，而是逐步验证生物神经元在复杂任务上的可用性与可扩展性。
Cortical Labs最核心的量化主张是能耗：创始人兼首席执行官Hon Weng Chong称，每个CL1单元耗电量低于手持计算器。墨尔本机房规划部署120个CL1单元；新加坡项目在初期将从新加坡国立大学Yong Loo Lin医学院启动，最终计划分阶段部署最多1,000个单元。按容量目标看，单中心规模在新加坡可达墨尔本的8.33倍（约1000/120），且所用神经元来自人类血细胞重编程，形成从生物材料到可计算单元的供应链雏形。
The story, published on March 10, 2026 at 06:00 GMT+8, reports that Australia-based Cortical Labs and DayOne Data Centers Ltd. have announced two biological data center projects: one in Melbourne and one in Singapore. Instead of conventional AI chips in server racks, the facilities will run on CL1 biological computing units based on human brain cells. The company says the system is still modest in capacity and unlikely to challenge mainstream AI processors such as Nvidia Corp. for years or decades, but the move aligns with a measurable trend: rapid global AI data-center expansion is driving higher electricity demand, increased water use, and silicon shortages.
The article also highlights a measurable technology progression. Cortical Labs had previously trained its brain-cell systems on the simple game Pong, and last month reportedly trained them on the more complex game Doom, suggesting rising task complexity. The neurons, grown from stem cells, are placed on chips that exchange electrical signals with software; inputs are sent, responses are recorded, and software interprets these responses as computing output. This loop suggests the company is developing a repeatable biological compute pipeline rather than a one-off demonstration.
A key numeric claim is energy efficiency. Founder and CEO Hon Weng Chong said each CL1 unit uses less power than a handheld calculator. The Melbourne site is planned for 120 CL1 units, while the Singapore deployment is designed in phases to scale up to as many as 1,000 units, beginning at the National University of Singapore’s Yong Loo Lin School of Medicine. That implies about an 8.33× scale increase over the Melbourne target. The Singapore project also uses neurons converted from human blood cells, indicating a potentially scalable pathway from biological material sourcing to larger compute clusters if performance and reliability continue to improve.