One of the defining moments of this year’s Computex came when Supermicro President and CEO Charles Liang took to the stage. His keynote didn’t just echo the AI-first future—it laid down the infrastructure blueprint that will make it real.
Positioning Supermicro at the heart of the AI revolution, Charles highlighted how the company is responding to explosive AI growth with agile manufacturing, liquid cooling leadership and a sweeping portfolio of Blackwell-ready systems already rolling off production lines.
Let’s explore the most critical announcements and what they mean for system integrators, hyperscalers and AI innovators.
Performance, Efficiency and Readiness
Charles opened by addressing the elephant in the room: AI’s exponential compute demand. His message was clear—Supermicro is ready, with a flexible architecture roadmap tailored for Grace Blackwell, Hopper and Gaudi 3, delivering scalable, energy-efficient infrastructure for AI factories across the world.
From rack-scale deployments to developer workstations, Supermicro’s modular designs allow customers to move fast and deploy anywhere.
One standout: Supermicro is already shipping NVIDIA GB200 NVL72 and DGX SuperPOD platforms, with optimised thermal design and ready integration for liquid cooling—meeting both high-density compute and sustainability goals.
The Next Generation of Direct Liquid Cooling
A centrepiece of Charles’ keynote was the introduction of DLC-2, Supermicro’s latest generation of direct liquid cooling. Unlike earlier solutions, DLC-2 cools not just the CPU and GPU, but also the power supply units, memory and voltage regulator modules—capturing more heat and delivering significantly improved thermal efficiency.
At the heart of this system is PGDLC2, a new Supermicro-developed cooling fluid featuring ultra-high electrical impedance—making it far less likely to conduct electricity and offering greater safety in high-density environments.
The result? DLC-2 enables data centres to reduce electricity costs by up to 40%, while cutting total cost of ownership by up to 20%, all while lowering water usage, fan noise and space requirements.
This innovation places Supermicro squarely at the forefront of sustainable, high-efficiency AI infrastructure—and aligns with Boston’s own portfolio of immersion and direct liquid cooling solutions, including Glacier Core servers and Castrol-powered deployments.
From Components to AI Factories: DCBBS
Another major reveal was Supermicro’s Data Center Building Block Solutions (DCBBS)—a complete blueprint for constructing AI data centres from the ground up. DCBBS spans the entire stack, from individual GPUs to full racks, including cooling infrastructure, power delivery and integration services.
Built on a modular philosophy, DCBBS delivers validated, pre-integrated subsystems that streamline deployment and enhance reliability. For hyperscalers and enterprises building AI factories, this represents a new level of flexibility, scalability and speed.
And with Supermicro’s global rack-scale integration capabilities, it’s possible to go from design to deployment faster than ever—making the AI factory a reality now, not next year.
What This Means for Boston Customers
At Boston, we’re proud to be a long-standing partner of Supermicro, delivering pre-validated systems across sectors—from financial services to academia and national AI initiatives. With Charles Liang’s keynote confirming Supermicro’s leadership in AI infrastructure, now is the time to test, scale and deploy.
We’ve already begun enabling customers with access to the latest Grace Blackwell systems via Boston Labs, alongside custom-built immersion-cooled AI solutions and sovereign-scale deployments.
Whether you're looking to explore DLC-2, build an AI factory with DCBBS or test the latest GB200 and DGX SuperPOD platforms, we’re here to help.
Want to test the latest Supermicro AI platforms? Let’s talk.