Businesses are on the cusp of a major paradigm shift. AI is here to stay and businesses that invest now will be in the driver’s seat when it comes to market leadership tomorrow.
A new blog from IDC Group Vice President for Infrastructure, Ashish Nadkarni. Sponsored by HPE and Intel.
The adoption of new AI applications and AI functionality within existing applications is highly disruptive but nevertheless necessary for enterprises as they shift into the next phase of DX. While DX continues to initiate organizational, process, and technological changes, AI will turbocharge them. IDC estimates that in 2024, 33% of G2000 companies will be exploiting innovative business models to double their monetization potential of generative AI.
AI and analytics investments also further change an organization’s relationship with data. This change is not just about the volume, variety, velocity, and veracity of data, something that many organizations know how to manage. It's about adding additional dimensions to distributed analytics, data logistics, compliance, governance, and data security. The resulting insights can be so deep that the business must almost ensure such insights can be acted upon immediately to differentiate itself from its competitors.
AI workloads include software applications and their corresponding data sets. They fall broadly into three categories: Training, Tuning, and Inferencing.
- Training is usually a batch and iterative process and involves building new AI models and/or modifying existing ones. It requires massive amounts of compute for a finite amount of time. Time to results can be hours, days, or weeks.
- Tuning is the process of adjusting or optimizing pre-built models with the company’s own data. This is an important set of activities for enterprises looking to adapt LLMs to their business. Like training, time to results can be hours, days, or weeks but is generally shorter than the task of building the model from scratch.
- Inferencing is more of a continual process and involves running trained and optimized models with corporate data and real-time or near-real-time execution.
Use cases include generative AI and computer vision in which businesses train, retrain, tune, or optimize large language models (LLMs), scientific discovery, and financial market modeling.
Planning for AI and digital infrastructure investments
As with any business initiatives, deploying AI and analytics at scale requires well-planned investments in fit-for-purpose digital infrastructure (DI) that complements workload-optimized computing platforms at the core and edge with the appropriate application environments. Training workloads are largely limited to large clusters of highly performant servers or supercomputers deployed in the datacenter or in a public cloud. Inferencing workloads can be deployed on standalone servers in a variety of locations. IDC anticipates that while initial investments are skewed towards training and optimization tasks, it is the outcomes of inferencing tasks that hold the key to business results in the long term. Accordingly, businesses must plan their AI and digital infrastructure investments equitably across training and inferencing workloads.
IT leaders can deliver consistent and reliable outcomes to their business by partnering with proven technology providers that have a successful track record of delivering full stack AI solutions. Some deliver optimized training solutions but fall short when it comes to platforms optimized for inferencing tasks. Here, the provider's investments in creating an ecosystem of certified software stacks are equally important to its investments in creating a portfolio of secure, reliable, optimized, and sustainable computing platforms.
Message from the Sponsors
|
HPE ProLiant Gen11 powered by Intel Xeon processors is secure, efficient, optimized, and it’s engineered for your hybrid world. At HPE, we understand for your IT initiatives to be successful, you need accelerated performance that can meet the needs of your existing and emerging AI workloads and deliver critical insights that drive tangible benefits. From edge to cloud, the right choice of compute—one that delivers a cloud operating experience built from the ground up with a fundamental foundation security approach—can set you apart from the competition. Read more at: https://www.hpe.com/us/en/hpe-proliant-servers.html The world needs computing power that is up to the challenges of emerging AI capabilities and other intensive workloads, and organizations are already leaning into the capabilities of Intel Xeon Scalable processors. Whether it is empowering solid foundations for AI innovation and HPC, supporting critical workloads at the edge, building a secure cloud, or helping professionals stay productive, there is an Intel Xeon processor designed to meet your organization's computing needs. Learn more: https://www.intel.com/content/www/us/en/products/details/processors/xeon.html
Ashish is Group Vice President within IDC's Worldwide Infrastructure Practice. He leads a team of analysts who engage in delivering qualitative and quantitative research on computing, storage, and data management infrastructure platforms and technologies, via syndicated research programs (subscription services), data products (IDC Trackers) and custom engagements. Ashish’s vision for his team is to take a holistic, forwarding-looking and long-term view on emerging as well as established infrastructure-related areas in the datacenter, in the cloud and at the edge. It is complemented by research on current and next-gen applications and workloads, vertical and industry-specific use cases, emerging storage and server form factors and deployment models, and up-and-coming IT vendors. Ashish also takes a keen interest in tracking the ongoing influence of open and open-source communities like OpenStack and the Open Compute Project on infrastructure. Ashish holds an M.B.A. in Entrepreneurship from Babson College, Massachusetts, and an MSc. in Physics from the University of Pune, India. Compute Experts twitter.com/hpe_compute
|
ComputeExperts
Our team of Hewlett Packard Enterprise server experts helps you to dive deep into relevant infrastructure topics.
