Few tech companies have shaped computing like Intel has since its “Intel inside” campaign decades ago. Yet today, as SAP customers navigate a landscape of hyperscalers, AI accelerators, and cloud migration, Intel’s foundational role has paradoxically become more critical and less visible.
According to Jan Krueger, Account Director & CTO for Team SAP within Intel’s Sales and Marketing Group, the behind-the-scenes dynamic defines Intel’s approach to the SAP ecosystem. Drawing on decades of experience integrating Intel® Xeon® processors into SAP environments and spearheading the company’s technical strategy for SAP workloads, Mr. Krueger embraces Intel’s unique position: “Even though we are not often seen as front and center, we are running SAP workloads across the board.”
This behind-the-scenes partnership spans over 25 years and operates along what he calls “two swim lanes”: ensuring SAP’s diverse portfolio performs optimally while simultaneously empowering customers in heterogeneous environments.
In these lanes, Intel enables everything from SAP ERP, starting with SAP ECC and SAP Suite to the latest S/4HANA systems to cutting-edge AI implementations. Whether an established enterprise or a nimble start-up, companies of all sizes “can benefit from this partnership because it drives their success at the end of the day,” noted Mr. Krueger.
Beyond the SAP ecosystem lies the reality of enterprise computing: most organizations maintain diverse technology stacks, with processing needs extending from hyperscaler data centers to factory-floor edge devices. Here, Intel’s consistent architecture facilitates workload portability and horizontal scalability that would otherwise be next-to-impossible.
In a recent conversation with ASUG, Mr. Krueger provided insights into Intel’s partnership with SAP and detailed Intel’s approach to AI implementation, hybrid environment support, and the capabilities of their new Intel® Xeon® 6 processor portfolio.
This interview has been edited and condensed.
Q: How does Intel’s long-standing relationship with SAP help it stay agile with emerging technologies like large language models, and what specific role does Intel play in making these innovations technically viable and practically profitable?
SAP benefits from this partnership by getting access to Intel technology much earlier than the rest of the market. If they had to wait until Intel’s product launch to make decisions on support, and feature adoption, they would lose a lot of time to bring those benefits to customers.
We share information early with SAP’s engineering teams. We have co-development initiatives, task-finding activities, and research activities where SAP works with us to determine which technologies to use, where they help most, and how to productize them. This ensures new technologies work jointly on platforms that provide resiliency, security, and stability.
With AI, we’re working on the next iteration beyond the hype — turning models into use that generates benefits and revenue. We’re at an inflection point where SAP is starting to turn AI applications into profitability streams.
We help SAP with scalable infrastructure to run different LLMs in ways that become profitable solutions. SAP’s strategy is partnering, not developing models themselves, but fine-tuning and optimizing them for their market segment.
Q: SAP recently announced Business Data Cloud, which integrates Databricks and other technologies. How is Intel’s partnership with SAP evolving to support these newer cloud-based offerings, and what unique value does your technology bring to this ecosystem?
We tend to focus on the established part of the business at SAP. Over the last couple of years, SAP has transformed into everything as a service, RISE, and now with AI and solutions like Business Data Cloud (BDC).
We maintain focus on the established business because there are still a lot of customers asking for our solutions. Most of what we do applies to SAP RISE as well because the software’s optimizations and benefits are the same as on self-hosted or co-located SAP landscapes outside SAP RISE. The same software is now running in the cloud on hyperscalers’ infrastructure, but the benefits remain.
Now, SAP is benefiting directly from the Intel advantage because it is running the software and maintaining the infrastructure as part of the SAP RISE offering. That’s why we also partner with Microsoft, Google, and AWS to ensure that what we’ve achieved transitions into this new era, where SAP is the managed service provider through SAP RISE.
For solutions like BDC, customers benefit from Intel partnering with cloud-native companies. Intel has been working with Databricks over the past couple of years, similar to what we’ve done with SAP, making sure the solutions run efficiently, perform optimally, and remain resilient on Intel-based cloud instances. BDC inherits benefits that Intel and Databricks have put together, as well as benefits driven by Intel and SAP.
With BDC, individual partnerships have turned into multi-company partnerships — Intel, SAP, Databricks, Microsoft Azure, Amazon Web Services, and Google Cloud Platform all host SAP services. In the AI space, Business AI is turning into agentic AI, but many of these are inference-based workloads. SAP is taking data from BDC and SAP RISE and utilizing this business process expertise to turn them into customer value.
Many of these workloads run on Intel platforms, though some use other vendors for AI acceleration. Intel is catching up in the AI accelerator market with our recently released Intel Xeon 6 processor family and Intel® Gaudi® 3 AI Accelerators. We are optimistic that with the introduction of Intel Gaudi 3-based instances with our hyperscaler partners, SAP will be able to use them for their solutions as well.
Q: The market for AI hardware is evolving rapidly, with dedicated accelerators becoming increasingly important. Could you elaborate on Intel’s approach to AI acceleration, particularly how it differs from competitors and addresses the specific needs of SAP workloads?
Regarding AI accelerators, the market perceives an AI accelerator as a dedicated hardware component, like an Intel Gaudi or Nvidia GPUs. But for Intel, accelerators have also always been present on our CPUs. We have had many different accelerators on our CPU for over 20 years, such as cryptography accelerators that encrypt data much more efficiently than software.
For a couple of years, we have had Intel Xeon accelerators for AI-related functionality, like our Intel® Advanced Matrix Extension (AMX), which accelerates the foundational vector operations of large language models utilized for embeddings or inferencing services.
Intel has a long history developing accelerators and has entered the market with Intel Gaudi AI accelerators, currently in 3rd generation – Intel Gaudi 3. We expect it to provide access to Intel Gaudi 3-based instances via the hyperscalers and OEMs throughout this year.
Like any other technology, there is no one-size-fits-all. SAP provides multiple services and solutions for AI, and they are generally implementing a Retrieval- Augmented Generation (RAG) that utilizes various components, including grounding or embeddings for specific reference data, so that the responses are based on factual content. Often, lightweight LLMs are suitable for embeddings or grounding services, but these work ideally on Intel Xeon processors. Intel Gaudi 3 accelerators support the needs of larger models — above 20B parameters and high-performance use cases.
We are seeing the market deploying LLMs and RAGs onto Intel Xeon-based platforms, since it optimizes the overall costs, accelerates the time to market, and finally enables higher profitability. Any AI accelerator comes at an additional cost, therefore a company’s choice between CPU and AI Accelerator usage can impact enterprise success.
SAP benefits from using Intel Xeon-based accelerators for various parts of the portfolio, and we are continuing to support SAP’s AI strategy with our product features. Our focus is enabling AI use cases so that companies like SAP can turn them into profitable solutions. If it doesn’t become profitable, it’s not sustainable.
Q: Beyond supporting SAP’s AI initiatives, many organizations want to develop their own AI solutions. How is Intel helping these customers build effective AI implementations while addressing the critical concerns around data privacy, security, and regulatory compliance?
Helping people reach a commercially viable solution involves accelerators in different price categories that you can choose based on your use case.
For example, last December, SAP HANA released an embedding service that allows customers to upload documents that are transformed into vectors. When you run an LLM-based workload, your question is matched to vectors in the HANA Cloud Vector database module for accurate answers.
This embedding service doesn’t have time criticality — if processing 1000 documents takes a few minutes, that’s fine. This could run on an AI accelerator in seconds at a high price, but running on CPU takes a bit longer at a much lower price point.
Outside the SAP context, Intel integrated an opensource project under The Linux Foundation called Open Platform for Enterprise AI (OPEA). This is an open, standardized platform where many vendors participate to provide a freely available platform for any Gen AI use case.
The OPEA platform has already 10+ use cases and 20+ Sample Applications implemented, such as ChatQnA, CodeGen, CodeTrans, DocSum or translation services. Customers can run these container-based Kubernetes solutions anywhere on-prem with our partners like Dell or consume it via the hyperscalers’ marketplaces at AWS or self-hosted in cloud instances.
OPEA has a modular, API-driven architecture that allows you to replace components within the infrastructure. If you don’t want to use the Redis database for storing vectors, you can replace it with HANA Cloud Edition or MongoDB. Similarly, with LLMs, you can swap in models that better suit your language needs or point to other API-based LLM services like OpenAI. Conversely, it also allows you to move from API-based OpenAI usage to open-source, self-hosted LLM usage.
Cost-related optimizations, like running LLMs on CPU or choosing different accelerators, only requires changing a configuration file. The complex integration is already done by Intel and other contributors of OPEA.
For federated/confidential AI, security has already been implemented into the platform. Intel uses two technologies: Intel® Software Guard Extensions (Intel® SGX), an enclave technology Intel has offered for many years, allows running AI processing with confidential information inside a “safe” where nothing goes in or out, and secondly Intel® Trust Domain Extension (Intel® TDX) that runs the entire VM / Container in a protected, encrypted “safe.” Intel SGX offers a very small Trusted Compute Base (TCB), implemented through the application, while Intel TDX TCB is extended and allows an entire VM / Container to run confidentially. Both have pros and cons for customers, but both ensure data used in AI services will remain secure during its processing, not just at rest or transport.
Both implemented via OPEA provide ease of use for customers wanting confidentiality or privacy preservation today. They can use the same AI framework, additional components, and get all the privacy without implementing the confidentiality features themselves. By choosing either Intel SGX or Intel TDX in the OPEA configurations, and using the latest Intel Xeon 6 processor family, customers are ready to run confidential AI.
Q: Many SAP customers operate in hybrid environments, with workloads spanning cloud and on-prem systems. How does Intel’s approach specifically support these complex hybrid architectures, and what advantages does this provide for enterprises managing diverse computing environments?
When it comes to hybrid environments, Intel and its OEM partners continue offering SAP-certified Intel Xeon-based platforms that scale from entry level 2-socket to very large 16-socket platform with up to 32TB memory capacity available to a single scale-up SAP HANA installation.
When it comes to solutions like SAP’s Edge Integration Cell, we are partnering with the ecosystem to provide SAP certified platforms as well. Intel has a wide range of products that can scale up or out and support customers’ requirements.
Customers can use Edge Integration Cell to bridge RISE or SAP BTP solutions with what they have on-premise. For those who run hybrid—an S/4 system on-premise and another in the cloud—it all runs on Intel. End customers have the assurance that whether they run the S/4 system on a Dell, HPE, or Lenovo server or on an AWS, Azure, or Google Cloud instance, the Intel Xeon processor capabilities and features will remain the same.
Intel is behind the scenes with all these vendors to ensure SAP software runs with its expected performance, security and resiliency across all platforms.
Additionally, the Intel Xeon 6 processor with E-Cores provides a new power-efficient platform for SAP landscapes that customers can utilize to reduce their power consumption and support their sustainability goals.
We are also partnering with companies like Lemongrass, which are involved in the platform decision-making process. They analyze requirements and recommend infrastructure based on customer needs, perhaps a top-tier option with the best performance but the highest cost, a middle option, or a lower-cost option with decent performance but fewer regional availability options. The value for SAP customers lies in the flexibility and choice Intel Xeon-based platforms provide globally.
With SAP RISE, SAP provides the landscape as part of its Managed Service Provider (MSP) offering so that customers can jump-start the migration process. Intel-based instances are also used for SAP RISE and provide the same capabilities, as if customers had opted for a self-managed approach.
For customers that need embedded devices that operate in harsh and regulated environments, Intel partners with the ecosystem to provide those solutions as well. SAP customers can run many workloads right into the machines with high performance and more capabilities than ever before.
With a customer developing mining equipment, using Intel-based embedded devices allows them to process more information in the vehicle and bring them back to SAP solutions via gateways like Edge Integration Cell. Intel-based platforms are the backbone for this type of business success in partnership with SAP and our shared ecosystem.
Q: You’ve recently launched the Xeon 6 portfolio with dozens of new products. How does this platform deliver on the promise of a truly scalable, multi-purpose infrastructure that can support everything from cloud deployments to edge computing for SAP customers?
What you’ve outlined is exactly the essence of our value for the industry, though it’s not as visible as we’d like. In the area of cloud computing and SAP RISE, the infrastructure discussion has historically moved into the background, even though it’s essential to business success. Without fully exploring the options available and doing a thorough assessment, companies risk choosing the wrong platform for their SAP landscape which can affect their ability to innovate, adopt, and grow.
Intel just launched the Xeon 6 processor portfolio, consisting of Performance Core (P-cores) & Efficiency Core (E-cores) products. The Intel Xeon 6 processor with E-Cores provides a power-efficient and scalable platform, while the Intel Xeon 6 with P-Cores provides a very high per-Core performance: the first is more suitable for Business Warehouse scenarios, and the second for high transactional workload scenarios.
This huge product portfolio spans the ecosystem— on-premise, embedded systems, telco systems, and certainly cloud—all based on the same Intel Xeon 6 processor technology built on our Intel 3 process node.
Intel Xeon 6 processors with P-Cores for SAP workloads will support the memory footprint of 4TB per socket, the largest we’ve ever had. Customers can bring up to an 8TB SAP HANA system onto a 2-socket Xeon 6, whereas our previous generation required a 4-socket system for the same capacity. This provides enormous consolidation opportunities with up to 26:1 ratio, meaning 26 servers from the past can be replaced with one Xeon 6-based server, considering the performance gains only. Practically speaking, such high consolidation is not always possible, but it shows the huge improvements of Intel Xeon 6 over its predecessors.
Customers can deploy onto the Intel Xeon 6 with E-Core as well, since it offers SAP support for 2TB per socket, allowing 4TB HANA Deployments on 2-socket systems. The advantage of Intel Xeon 6 with E-Cores is up-to-30%-lower power consumption / socket with top performance for SAP HANA.* This represents significant savings for organizations with large SAP landscapes over a usual lifetime of five years.
Intel Xeon 6 with E-Cores and Xeon 6 with P-Cores help to consolidate systems and reduce power consumption to ultimately improve cost performance and overall TCO.
All of this is SAP-certified: fully tested, validated, and ready to be deployed though the SAP-certified partner ecosystem.
Q: Looking at both technical capabilities and market potential, what excites you most about the new Xeon 6 platform? And what does this generation of technology suggest about the future direction of enterprise computing infrastructure for SAP environments?
I’m very excited about Intel Xeon 6 processors, because they represent a breakthrough for Intel in recent years. Having a product that scales horizontally across all segments is a huge advantage.
For SAP customers and for SAP itself, the key benefit is flexibility. Customers seeking TCO-optimized, energy-optimized SAP landscapes now have the option of using Intel Xeon 6 with E-cores. Those requiring high-performance workhorse systems that scale up to 32 terabytes and beyond are well-suited by Intel Xeon 6 with P-Cores. Especially in multifaceted compute environments, a mixture of Intel Xeon 6 with E-Cores and P-Cores will provide the most benefits.
This flexibility is unique in the market. No other vendor in the SAP landscape has this breadth of portfolio, horizontally and vertically. No one else has this focus on providing computing products specifically aimed at sustainability.
Intel Xeon 6 specifically targets sustainability goals without sacrificing performance and SAP’s certification shows that we delivered by designing CPUs with top-notch performance and energy savings.
This strategy will continue with our next-generation Intel Xeon products arriving in phases through the coming years. We’ll deliver different types of products through the same product stack, ensuring interoperability. Whether you’re coming from performance cores, efficiency cores, or neither, all your software can run on these platforms.
This enables end customers and SAP to consume these values within their offerings. That’s the true power of Intel Inside, especially now with Intel Xeon 6.
To learn more, visit the SAP community blog written by Stefan Baeuerle.
*Compared to 4th Gen Intel Xeon Scalable Processors. Based on architectural projections as of August 21, 2023 relative to prior generation. Your results may vary.
About Intel
Intel (Nasdaq: INTC) is an industry leader, creating world-changing technology that enables global progress and enriches lives. Inspired by Moore’s Law, we continuously work to advance the design and manufacturing of semiconductors to help address our customers’ greatest challenges. By embedding intelligence in the cloud, network, edge and every kind of computing device, we unleash the potential of data to transform business and society for the better. To learn more about Intel’s innovations, go to newsroom.intel.com and intel.com.
About ASUG
ASUG is the world’s largest SAP user group. Originally founded by a group of visionary SAP customers in 1991, its mission is to help people and organizations get the most value from their investment in SAP technology. ASUG currently serves thousands of businesses via companywide memberships, connecting more than 130,000 professionals with networking and educational resources to help them master new challenges. Through in-person and virtual events, on-demand digital resources, and ongoing advocacy for its membership, ASUG helps SAP customers make more possible.