Optimized intelligence drives AI evolution

Optimized intelligence is redefining how enterprises strategy AI workloads, signaling a shift away from the GPU-first mindset that has dominated latest years. As AI adoption surges, firms are prioritizing structure that aligns with particular use circumstances, delivering smarter efficiency, decrease prices and higher belief throughout hybrid and multicloud environments.

Key conversations throughout the business now heart on aligning {hardware} and software program roadmaps, scaling AI whereas managing whole value of possession and getting ready for next-generation applied sciences comparable to quantum computing. The collaboration between SAS Institute Inc. and Intel Corp. displays these priorities, emphasizing workload-specific acceleration, moral AI practices and human-centered design that ensures confidence in real-world outcomes, based on Muge Tanik (pictured, proper), basic supervisor for AI options and partnerships at Intel Corp.

SAS’ John Carey and Intel’s Muge Tanik speak to theCUBE about optimized intelligence.

“There was a false impression within the market that each one the AI workloads work on GPU solely and likewise gen AI equals GPU, which isn’t appropriate,” Tanik defined. “We take a look at the frameworks, we take a look at the information sort, the accuracy of the information, we take a look at the inference performances. Our platforms have proven very promising and aggressive outcomes on all these platforms. We’re extraordinarily excited and proud.”

Tanik and John Carey (left), vp of world channels at SAS, spoke with theCUBE’s Rebecca Knight at SAS Innovate, throughout an unique broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They mentioned how SAS and Intel are advancing AI by way of optimized intelligence, shifting past GPU-first considering to ship workload-specific efficiency, value effectivity and future-ready innovation. (* Disclosure beneath.)

SAS and Intel advance AI with optimized intelligence

SAS and Intel’s 25-year partnership is constructed round optimized intelligence, with each firms investing deeply in technical alignment to make sure {hardware} and software program co-evolve. By integrating Intel’s silicon improvements, comparable to AMX acceleration, trusted compute and Gaudi GPUs, SAS delivers constant, cost-conscious efficiency on Viya, its AI and analytics platform. This relationship allows shoppers to deploy superior fashions throughout sectors comparable to monetary providers and public sector with out overspending on infrastructure, based on Tanik.

“We’ve got this roadmap alignment, however AI is the important thing focus for us,” she mentioned. “Viya has been displaying nice efficiency on [the] newest Intel platforms. We’ve got seen like [a] 24% efficiency improve of Viya for the velocity take a look at, and we’ve got seen [a] 94% improve of Viya efficiency on throughput assessments.”

On the engineering stage, SAS and Intel keep a suggestions loop that enables SAS to form chip design earlier than finalization. This iterative co-development ensures real-world wants are thought of earlier than merchandise hit the market. SAS’s R&D groups use Intel’s newest {hardware} to unlock new parameters for AI acceleration, permitting sooner deployment of fashions that drive enterprise worth, with out requiring prospects to overtake their current infrastructure except completely vital, Carey defined.

“When these guys ship over {hardware} to our R&D, it’s like Christmas morning,” he mentioned. “These are the items that carry on giving as a result of they permit our staff to get actually excited in regards to the new parameters that they will open up, the brand new acceleration, how they will remedy these issues and take full benefit of all the new innovation that Intel is doing on the chip set.”

This hands-on, customer-first strategy underpins each corporations’ perception in a versatile, consultative AI roadmap. Quite than pushing high-powered GPUs by default, SAS and Intel assess buyer targets, workload sorts and current investments to supply tailor-made options, whether or not which means CPUs, accelerators or hybrid cloud infrastructure.

“One dimension doesn’t match all. All the pieces begins with the workload, and the whole lot begins with what ache factors that buyer is attempting to resolve,” Tanik mentioned. “We provide options to the shoppers and we meet them the place they’re.”

Right here’s the entire video interview, a part of SiliconANGLE’s and theCUBE’s protection of SAS Innovate:

(* Disclosure: TheCUBE is a paid media associate for SAS Innovate. Neither SAS Institute Inc., the sponsor of theCUBE’s occasion protection, nor different sponsors have editorial management over content material on theCUBE or SiliconANGLE.)

Photograph: SiliconANGLE