Microsoft and Intel Expand Azure and Chip Partnership; Microsoft-Designed Chips to Use EMIB Packaging
Microsoft and Intel signed a long-term agreement on November 18 to expand their collaboration on Azure cloud and AI infrastructure. The next generation of Microsoft-designed chips will use EMIB and related packaging technologies from Intel Foundry. Intel will also supply Xeon processors under a multiyear agreement, including the next-generation server CPU “Diamond Rapids.” The two companies will deepen their collaboration in CXL memory expansion, Intel QAT, and confidential computing based on Intel TDX. Azure M-series v4 virtual machines combining CXL memory with Xeon 6 deliver up to 2.3 times higher performance, while Intel QAT accelerates cloud-based SQL backups by up to 2.5 times.
Why read this
AI score 66
The partnership spans advanced packaging, server processors, and system-level optimization, illustrating a concrete path for coordinating cloud providers’ in-house chip development with external supply chains.
Topics
Body · AI translation
On November 18, Microsoft and Intel announced an expanded partnership, signing a long-term agreement to jointly advance Azure cloud and AI infrastructure. Microsoft thereby became the first company to announce that it will use Intel Foundry’s advanced EMIB packaging technology for its in-house chips.
The agreement covers several areas:
Intel confirmed that it will provide Microsoft with a multiyear supply commitment for Xeon processors, including the next-generation server CPU “Diamond Rapids”;
Microsoft will use EMIB technology to package its in-house Azure silicon, including its next-generation Maia AI accelerator and Cobalt CPU.
The two companies will also collaborate closely on CXL memory expansion, Intel QuickAssist acceleration technology, and confidential computing.
According to technical data disclosed by the two companies, the partnership has already begun to deliver results. By combining CXL memory with the flat memory mode of Xeon 6 processors, Azure M-series v4 virtual machines have achieved performance gains of up to 2.3 times, while Intel QAT technology has increased cloud-based SQL backup speeds by as much as 2.5 times.
Microsoft Leads Adoption of EMIB Packaging, Providing Major Endorsement for Intel Foundry
Microsoft is adopting Intel Foundry’s EMIB-T packaging technology.
Building on standard EMIB, the technology introduces through-silicon vias (TSVs) to support vertical power delivery across multiple layers of stacked chips, meeting the packaging requirements of three-dimensional silicon modules.
Microsoft’s in-house silicon portfolio currently includes the Maia 200 AI accelerator for large-scale inference workloads and the Cobalt 200 CPU optimized for agentic workloads. Under the agreement, the next-generation versions of these products will be packaged by Intel Foundry using EMIB and derivative technologies.
Naga Chandrasekaran, executive vice president of Intel Foundry, chief technology and operations officer, and general manager of Intel Foundry, said:
He said that combining EMIB’s heterogeneous integration flexibility with the manufacturing scale required for the AI era will deliver new performance and economic benefits to Azure customers.
For Intel, securing a foundry order from Microsoft, a leading cloud provider, represents significant external validation of its foundry strategy.
Multiyear Xeon Supply Secured, Diamond Rapids Added to Azure Roadmap
In addition to advanced packaging, the agreement also confirms Microsoft’s long-term purchasing commitment for Intel Xeon processors.
Intel will provide Azure with a multiyear supply commitment for Xeon processors, explicitly including the next-generation Xeon 7 server CPU code-named “Diamond Rapids.”
Microsoft said the partnership will extend across the entire Azure roadmap and Intel-based Azure hardware systems. This indicates that Intel processors will continue to hold an important position in Microsoft’s data center infrastructure, rather than being limited to specific product lines.
Saurabh Dighe, corporate vice president of Azure Systems and Architecture at Microsoft, said:
Our long-standing collaboration with Intel brings together advanced packaging, custom compute, memory, and security capabilities to support Azure compute and continued innovation across Azure.
CXL Memory Expansion and Confidential Computing Provide Technological Differentiation
At the system-optimization level, the agreement focuses on three technical areas: CXL memory expansion, Intel QAT hardware acceleration, and confidential computing based on Intel TDX.
Regarding CXL, Azure M-series v4 virtual machines combine CXL memory with the flat memory mode of Xeon 6 processors to deliver performance gains of up to 2.3 times and greater memory flexibility for demanding workloads.
Regarding QAT, Microsoft launched the first cloud virtual machines supporting Intel QAT technology, increasing SQL backup speeds on Azure by as much as 2.5 times while also improving reliability.
Regarding confidential computing, Intel TDX technology encrypts memory and CPU register states when virtual machines migrate between hosts, strengthening security protections for sensitive workloads.
Microsoft said these investments are intended to provide customers with more predictable, scalable, and resilient infrastructure, specifically optimized for the performance, efficiency, scalability, and security requirements of cloud computing and AI workloads. Risk Notice and Disclaimer
Markets involve risks, and investments should be made with caution. This article does not constitute personalized investment advice, nor does it take into account any individual user’s specific investment objectives, financial circumstances, or needs. Users should consider whether any opinions, views, or conclusions in this article are appropriate for their particular circumstances. Users who invest on this basis do so at their own risk.
AI Trends