The following partner insight was authored by Jan Krüger, Global SAP Technical Lead at Intel.
Running enterprise systems used to mean following a familiar script. But cloud-native applications and artificial intelligence have changed the rules.
SAP customers now face steeper energy costs, persistent security threats, and AI workloads that demand unprecedented computing power. Traditional approaches to managing these environments can’t keep pace with these mounting pressures.
Cloud adoption has evolved far beyond its initial promise of cost savings through workload migration. Organizations are fundamentally rebuilding applications to extract maximum value from cloud platforms. Cloud computing now powers AI workloads, while AI makes cloud services more intelligent and efficient. Each advance in one technology opens new possibilities in the other, though this creates additional complexity for enterprises to manage.
The Evolution of AI
The relationship between enterprises and AI continues to evolve, particularly in how organizations deploy and utilize these technologies. While AI-as-a-Service has democratized access to these tools, moving away from massive, general-purpose language models toward specialized solutions marks a change in how enterprises implement and optimize AI capabilities.
“As the focus transitions, large language models will continue to grow in intelligence, but their size will decrease to meet specific industry requirements.”
“As the focus transitions, large language models will continue to grow in intelligence, but their size will decrease to meet specific industry requirements. For instance, instead of one massive, generalized model, we will see smaller models tailored to niche applications, such as healthcare, automotive, or retail,” explained Jan Krueger, Senior Technical Sales Specialist at Intel. “These models may only consist of a billion parameters but will be highly optimized for their respective domains or use cases. This will not only ensure efficiency, both in cost and performance but will make AI more accessible for targeted applications.”
SAP customers are seeing firsthand how AI strains their computing resources. Many organizations started by building and training their own AI models — a process that demands intense computing power. Now, as they shift to using these trained models to process new data, they’re looking for ways to run AI workloads more efficiently. The bottom line is simple: AI needs to prove its worth without burdening systems or budgets.
The environmental costs of running AI and cloud systems pose a growing challenge. Training and running large language models requires enormous computing power, and data centers need significant energy and water to stay cool. This resource consumption is particularly intense for large language models during training and operational phases.
With tightening regulations and changing market expectations, environmental impact has become a leading business consideration. By 2034, a company’s environmental practices will not only affect its ESG score — they will determine whether it can compete in a market where sustainability matters. Organizations that don’t embed green principles into their operations risk becoming obsolete as regulatory pressures mount and consumer expectations shift.
Powering Next-Gen Enterprises
Intel is pushing boundaries in six key areas: AI, quantum computing, memory systems, computational storage, neuromorphic computing, and security.
Intel’s new Xeon 6 platform offers two distinct approaches to computing: performance cores that deliver high per-thread processing power and efficiency cores that maintain strong performance while dramatically reducing power use. The platform lets customers choose the optimal mix for their workloads — performance cores for computationally intensive tasks, efficiency cores for better power efficiency, or a combination of both.
The company is advancing both technologies while refining its manufacturing processes. Over the next four years, Intel plans five major process improvements, starting with Intel 3 at 2-3 nanometers and progressing to Intel 18A at 1.8 nanometers by 2025. These new processors will meet the intense demands of AI and modern workloads while consuming less power.
“Quantum computing will likely become a reality, and when it does, it could render much of today’s encryption obsolete.”
Quantum computing is reshaping how we think about data security. Current encryption methods that protect sensitive data today won’t stand up against quantum systems tomorrow. “Quantum computing will likely become a reality, and when it does, it could render much of today’s encryption obsolete. The data encrypted over the past five years might suddenly become vulnerable because quantum algorithms can break those encryption methods,” said Krueger.
For organizations running SAP systems, data security is non-negotiable. Intel’s response has been two-fold: building quantum-resistant encryption into current products and addressing the challenges of distributed storage. The shift to networked data access has created new security risks; every point where data is stored becomes a potential target for attackers.
Intel’s Trust Domain Extensions (TDX) takes a novel approach to memory security. The system keeps cryptographic keys inside the CPU itself, using them to encrypt and decrypt all CPU memory operations in real time. Microsoft Azure and Google Cloud have built TDX into their platforms, protecting data wherever it runs.
SAP HANA puts unique demands on memory systems. Intel’s answer is Compute Express Link (CXL), a technology that changes how systems allocate and use memory. Just as cloud storage frees organizations from physical storage limitations, CXL lets applications request exactly the memory they need when they need it. Picture microservices applications checking memory resources in and out like library books. In essence, CXL transforms memory management into a more agile and efficient process, enabling SAP customers to optimize their memory usage, reduce costs, and maintain smooth system performance.
One of Intel’s most forward-looking investments is in neuromorphic computing, which aims to create computer architectures that function more like biological brains, potentially revolutionizing how AI systems process information. Unlike traditional computing architectures, neuromorphic systems can be more energy-efficient and better suited for certain AI workloads.
Intel’s research in this area could fundamentally change how enterprise systems handle complex AI tasks in the future, particularly for applications requiring real-time processing of sensory data or pattern recognition.
Intel has also founded the Open Platform for Enterprise AI (OPEA) in collaboration with the Linux Foundation. This open-source platform provides businesses with ready-to-use AI tools and use cases that can be seamlessly integrated into their SAP environments, including features like chatbots and coding assistants. OPEA is designed with flexibility in mind, allowing users to swap individual components as needed and run workloads on various hardware platforms, including CPUs, GPUs, or specialized AI accelerators.
Together, Intel’s focus areas reflect a comprehensive approach to helping businesses stay ahead of disruptions rather than constantly playing catch-up. Instead of scrambling to adopt each new technology as it appears, companies using Intel-powered solutions can integrate advances smoothly into their existing systems. This flexibility will determine how well they can innovate while controlling costs and complexity.
SAP customers face several critical technology shifts in the coming years. Getting ahead of quantum computing threats, environmental regulations, and AI requirements will be essential for successful digital transformation.
While raw computing power may become more available over time, using it efficiently will matter more than ever. Organizations will need solutions that deliver maximum performance without driving up energy costs or infrastructure spending.
Here’s what Intel’s technology makes possible:
- Build AI solutions tailored to specific industries, running lean and efficient
- Protect sensitive data with quantum-proof security systems
- Give data-heavy applications the exact memory they need when they need it
- Hit environmental targets while maintaining full performance
- Keep data secure across multiple cloud platforms
- Get the most value from every technology investment
Running enterprise systems means balancing competing demands: security with accessibility, innovation with stability, and performance with sustainability. Intel’s investments in key technologies help organizations manage these trade-offs, with AI, quantum computing, and advanced memory systems initiatives that deliver practical ways to adopt new capabilities while protecting their operations.
Jan Krüger is Global SAP Technical Lead at Intel.