Scaling Enterprise AI – Organically

The adoption of Artificial Intelligence (AI) within enterprises is no longer a question of "if," but "how."

For organizations looking to integrate AI seamlessly, an organic approach to scaling AI is often the most effective. This method focuses on gradual, sustainable growth that aligns with the organization’s existing processes, culture, and goals.

When I think back to my climb up Mount Everest, what I remember most is not the frigid temperatures, slippery slopes or dizzying altitude of being 17,000 feet above sea level. What I remember most is my Sherpa guide, my tent mates and my friends and family who cheered me on. Scaling a mountain and scaling enterprise AI share something important: Success requires strong partners.

Unlike many industries where companies rally to be “number one” or “the leader” or “the best,” technology innovation demands collaboration. I think tech companies inherently understand that every company in the industry has expert knowledge and varying levels of user acceptance. Combining them, however, accelerates ideas that work. The recent announcement by IBM and Intel about deploying Intel® Gaudi® 3 AI accelerators (Gaudi was a Spanish architect born in the mid-1800s and known for elaborate, highly stylized designs) as a service on IBM Cloud’s virtual servers for VPC is a great example. Expected to launch in early 2025 , IBM and Intel Corporation’s announcement is a very telling story about how companies will successfully scale enterprise generative AI (GenAI).

How is that going to work, you ask? Expect to scale AI organically.

Organically Scaling Enterprise AI? Big Yes.

There are varying statistics around AI’s successful deployment in large enterprise companies, and none are too impressive thus far. In 2020, IBM commissioned Forrester to conduct a study on the obstacles to scaling AI. Findings revealed that 90% of firms faced serious challenges leaping beyond the pilot stage and actually scaling AI across the enterprise. Data, it found, was one big reason why. Without high data quality and data integration, AI becomes impossible. Lack of skills along with a lack of tools to develop advanced analytics and machine learning models were also mentioned.

From my experience, the 90% statistic is most likely less today. After all, AI is rapidly moving forward. Yet, we’re also still climbing the mountain. Embedding AI into the technology we’re already using and not thinking of AI as a new, separate endeavor will make a difference. This organic approach is one reason IBM and Intel’s collaboration will impact AI adoption significantly: IBM Cloud is a mainstay. Gartner ranks it as one of the top cloud companies. This octet controls 97% of the market.

Companies using IBM Cloud include highly regulated industries like financial services, healthcare and government. The IBM Cloud focuses on efficiency and has one of the broadest offerings. With Intel’s Gaudi 3 accelerators and it’s open software on IBM Cloud, businesses are adding value without adding significant costs.

Here are four specific points I find interesting about Intel and IBM’s collaboration.

1. Enhanced GenAI. With the option of putting Gaudi 3 AI accelerators on the IBM Cloud, we’ll see an uptick in the cost-to-performance ratio for AI workflows as AI innovation becomes more efficient and accessible for enterprises.  Gaudi 3’s increased and faster HBM memory per card allows end users to load larger or/and multiple models on a single card.  Handling massive workloads in the cloud and in data centers, which goes hand in hand with GenAI, is a true breakthrough. We’ll see faster computing power, less time spent training models and reduced total cost of ownership.

2. Scalability. Hardware and software integration reaps benefits from two worlds: Intel’s powerful software and IBM’s improved AI scalability. I’ve been on many projects where a platform is built without considering the power needed to maximize the infrastructure. Having flexibility while scaling AI takes the guesswork out of the power equation. Also, Gaudi 3 has 33% more I/O connectivity per accelerator compared to competitiors for smarter scaling.

3. Security Resiliency. In many of my conversations with CTOs, the biggest fear of AI is lack of security. So what do people do? They turn to shadow AI, meaning they use AI on the side. This heightens the risk. IBM has the security services to monitor data and minimize risk through onboarding of Intel confidential computing technologies, watsonx.governance, and compliance certifications.

4. Industry Impact. Intel and IBM are aiming to enhance capabilities, which improves efficiency and performance across many industries like healthcare, financial services, manufacturing and technology. Businesses can extend Intel’s solutions and IBM’s consulting prowess to improve various parts within the enterprise, from operations to finance to human resources.

So much of AI right now is nuanced. I see the next big AI breakthrough being a faster, more seamless path from pilot stage to enterprise-wide adoption. AI is not a cool trend, but a competitive advantage embedded into solutions people are already using. If you use IBM Cloud or IBM watsonx, there’s no need to go somewhere else. Add AI into existing workflows. Like smartphones, everyone has one. It’s just part of our lives.

The challenge today is AI’s intimidating cost factor, but, like smartphones, those costs will come down. Technically, you can fit a Gaudi card in your pocket, but typically they live in an 8x card configuration in a 4U server.  Earlier versions used to be larger than a laptop. Now, I can put Gaudi 3 in my pocket.

More on Intel and IBM: See it Live

Tune in for my CXO Spice show where I interview Rohit Badlaney, general manager, cloud products and industry platform for IBM, live from Intel’s 500-acre campus in Hillsboro, Oregon. We’ll explore the top benefits enterprise companies can expect from using Gaudi 3 within IBM’s watsonx, the company’s GenAI and data platform, plus go deep on scalability, cost efficiency, performance and security. You might even get an exclusive look at what’s next for Intel and IBM as a power couple.

In the quest to conquer the mountain of AI innovation, IBM and Intel unite as experienced climbers bound by a shared expertise. Together, they are ensuring GenAI is within reach for everyone, no matter where they begin their ascent.

GenAI is changing the way we work and live. When it comes to AI solutions, not everyone can drive around in a Lamborghini – but everyone can have a basic car. IBM and Intel’s collaboration is making GenAI possible for everyone, everywhere.