On 22nd July 2025, Supermicro CEO Charles Liang delivered a keynote at the RAISE Summit, unveiling the company’s bold new strategy for AI infrastructure. Key topics included the modular Data Center Building Block Solutions® (DCBBS) architecture, next-generation direct liquid cooling (DLC-2) systems and Supermicro’s global expansion strategy to meet sovereign and hyperscale AI demand.
DCBBS – The Modular AI Data Centre Blueprint
At the core of Supermicro’s infrastructure push is DCBBS (Data Center Building Block Solutions®). This modular framework enables rapid customisation of AI and cloud data centres through pre-validated, rack-level subsystems that include:
- Compute: CPU and GPU platforms (including NVIDIA HGX, GB200 NVL, AMD Instinct, Intel Gaudi 3)
- Networking: Support for high-speed interconnects like NVLink, Infiniband and Ethernet
- Cooling: Integration of air, liquid and DLC-2 subsystems
- Power: Flexible power delivery models, including high-density redundant PSUs
- Management: Integrated IPMI/BMC, Redfish support and rack-level orchestration
Key benefits:
- Faster deployment: Complete systems can be deployed in weeks, not months
- Reduced integration overhead: All components are pre-qualified and interoperable
- Custom fit for purpose: DCBBS supports enterprise, hyperscale and sovereign AI workloads
Think of DCBBS as LEGO® for AI infrastructure - but industrial-grade, rack-validated and thermally optimised.
DLC-2: Sustainable, High-Density Liquid Cooling
Liang’s keynote introduced DLC-2, Supermicro’s latest direct liquid cooling system. Unlike conventional DLC, DLC-2 goes further by covering:
- CPUs & GPUs
- Voltage Regulator Modules (VRMs)
- System memory
- Power delivery components
Supermicro’s proprietary PGDLC2 cooling fluid offers high electrical impedance, ensuring safety even under dense configurations. Compared with air cooling or legacy DLC, DLC-2 can:
- Reduce energy use by up to 40%
- Lower total cost of ownership by 20%
- Minimise rack footprint, noise and water usage
Supermicro shipped over 4,000 liquid-cooled racks in 2024 alone, with customers reporting major energy and space savings.
DLC-2 and GlacierCore
While Supermicro’s DLC-2 technology featured prominently in the keynote as a critical innovation for rack-level energy efficiency, it’s important to note that liquid cooling isn’t a one-size-fits-all solution. The thermal strategy you choose depends on your data centre design, power density and long-term operational goals.
At Boston, we work closely with customers to identify the right cooling architecture for their needs. Alongside supporting DLC-2 deployments, we also design and deliver immersion-cooled infrastructure through our proprietary GlacierCore range.
GlacierCore systems are ideal for:
- Ultra-high density AI and HPC workloads exceeding traditional rack thermal limits
- Edge or modular deployments where ambient cooling is insufficient
- Organisations aiming for the lowest possible PUE in a compact footprint
Using Castrol ON immersion fluids, GlacierCore enables highly efficient, low-maintenance operation with a reduced environmental footprint. These systems are built to the same high standards as our rack-based platforms and can be integrated into sovereign or enterprise AI strategies, including those leveraging Supermicro compute nodes.
As highlighted in Charles Liang’s keynote, energy efficiency, thermal scalability and global readiness are the foundations of next-generation AI infrastructure. Whether your deployment requires advanced DLC or full immersion, Boston is ready to deliver end-to-end support.
AI-Powered Development Inside Supermicro
In a candid moment, Liang shared that around 40% of Supermicro’s firmware and software development is now assisted by AI tools. These include:
- Board-level diagnostics
- BMC automation
- IPMI/Redfish management integration
- System monitoring dashboards
By using internal AI models to boost engineering productivity, Supermicro is accelerating innovation without adding headcount, a model that aligns with broader industry trends.
Global Expansion and Sovereign Readiness
With rising demand from sovereign AI projects and hyperscale deployments, Supermicro is expanding its global presence with:
- Local manufacturing and logistics hubs across EMEA and APAC
- Regionalised support and compliance for AI infrastructure projects
- Inventory staging close to customer locations for faster delivery
This mirrors Boston’s own investments in regionally deployable infrastructure and fast-start AI Factory models based on Supermicro’s portfolio.
What This Means for Our Customers
As a premier Supermicro partner, Boston is positioned to deliver:
- DCBBS-based solutions tailored for AI workloads – pre-integrated and fast to deploy
- DLC-2 and immersion cooling options, supported by our in-house engineering team
- Hands-on testing in Boston Labs with NVIDIA GB200 NVL, AMD Instinct MI300X, Intel Gaudi 3 and more
- Regionally optimised deployment models across EMEA and beyond
Ready to Build Your AI Factory?
From modular DCBBS systems to DLC-2 cooled nodes and immersion-ready racks, Boston can help you:
- Accelerate your AI infrastructure deployment
- Meet sustainability goals with advanced thermal solutions
- Evaluate hardware before you invest
Get in touch with our experts today to discover how Supermicro’s innovations can power your next-generation AI infrastructure.