Putting AI to work. Moving from data to insight.

By | 3 minute read | January 28, 2020

Greener cities. Safer air travel. New treatments for degenerative diseases. Artificial intelligence (AI) just isn’t solving business problems anymore. It’s solving world problems. And you, our Business Partners, have played a critical role in empowering our clients across the globe to tackle big challenges, gain new insights and turn innovative ideas into reality.

Just to mention some examples — Oxford Cancer Biomarkers is now combining leading-edge diagnostics with AI technology to improve outcomes for cancer patients. And Bridger Pipeline LLC is protecting the environment with deep-learning AI. Similarly, for the automobile and industrial sectors, AI is being used for safety, inspection and process improvement.

Those were just a few of the examples I’ve seen, which means that our clients are adopting AI rapidly across multiple industries. And the solutions you are providing are extremely valuable and helpful in differentiating us in the market. However, the complexity of AI models continues to evolve while traditional IT infrastructures that aren’t purpose-built to handle these challenges can’t put AI to work.

Changing the model

You’ve probably heard about the Data-Train-Inference (DTI) model that we introduced last year to help guide clients on their AI journeys. Before diving into specifics, it is important to know that the DTI model is not a linear workflow. It is instead a continuous loop consisting of three stages that interact at all times.

These three stages of the model are:

  • Data: the stage where data is loaded into an AI model for training
  • Training: the magic stage where data becomes AI models
  • Inference: the stage where you derive insights from your model, and without proper inference all previous stages are for naught

For the DTI model to work, it is necessary to have the right infrastructure to support each stage. So how does DTI map to the IBM Power Systems portfolio?

For training, we offer the IBM Power System AC922 server, which you know very well and have implemented for several customers.

And what should you offer to your clients for the data and inference stages?

A new inference server in the AI lineup

Today, I am pleased to share with you the new IBM® Power® System IC922 server, a purpose-built inference server designed to put AI models to work and help clients unlock business insights. The new server supports up to six NVIDIA T4 Tensor Core GPU accelerators, and IBM intends to support up to eight T4 accelerators along with additional accelerator options.1 This level of configuration enables clients to leverage the inference accelerator that best suits their needs.

The Power IC922 uses optimized hardware and AI software — IBM PowerAI Vision — to deliver the necessary components for AI inference whether in a central data center or in a distributed data center closer to the sources of data. And the modular system can scale with business needs.

Storage-rich with 24 SAS/SATA — and future non-volatile memory express (NVMe) support intended1 — the Power IC922 also fits neatly into the data phase to help clients build a solid data foundation. Ultimately, the system is configurable, so whether clients are looking for a data server or an inference server, it offers the needed pieces for their AI solutions.

How does this announcement help you?

We are confident that this new technology and solution will bring to you new opportunities in the IT infrastructure for advanced analytics and AI, where you can leverage your expertise as an IBM Power Systems Business Partner. With the Power IC922, you can expand your footprint and offer your customers a portfolio of secure and robust servers for an end-to-end AI solution that matches with the DTI model and your client’s AI needs.

To help you in this journey and better prepare you for discussions with your customers,  please take a moment to familiarize yourself with the sales enablement material available in Seismic. You can find seller and client presentations along with technical deep dives that will help you understand this new technology.

In addition, I recommend that you visit the IBM Power System IC922 web page for further details. And to start the conversation with your clients, try sharing the announcement blog. Let’s put AI to work and deliver greater value to our clients.

Together we can drive growth and success in 2020.


[1] Statements regarding IBM’s future direction and intent are subject to change or withdrawal without notice and represent goals and objectives only. For more details about the Statements of Direction visit this page.

Explore more IBM Power Systems

Most Popular Articles