Intel Data Center & AI Group

We create world-changing technology that improves the life of every person on the planet.  Intel put the silicon in Silicon Valley. For more than 50 years, Intel and our people have had a profound influence on the world, driving business and society forward by creating radical innovation that revolutionizes the way we live.  

Intel_350x65.png

Intel

Businesses are generating an exponential amount of data as they integrate AI, 5G, and the Internet of Things (lot) into their tech stack.&nbsp;This torrent is crushing current storage capacity and guzzling energy as the power needed to run data centers and servers soars. In turn, companies may face a range of new costs that they have yet to fully grasp.&nbsp; Tirias Research estimates that data center server infrastructure operating costs from generative AI alone <a href="https://www.forbes.com/sites/tiriasresearch/2023/05/12/generative-ai-breaks-the-data-center-data-center-infrastructure-and-operating-costs-projected-to-increase-to-over-76-billion-by-2028">will exceed $76 billion by 2028</a>. The costs to the environment could be just as severe.

By 2025, it is projected the volume of wordwide data will grow to 175 billion terabytes.

Intel_Quicksight_Article1_Preview.png

Intel_Quicksight_Article1_M_Preview.png

AT&T_QS_D_V1_1.5_3.mp4

AT&T_QS_M_V1_1.5_3.mp4

While data generation surges, and enterprise storage needs are expanding exponentially, the current infrastructure to house data is out-of-date and inefficient.&nbsp;
Companies that don’t overhaul their storage solutions to account for this impending influx may find themselves with high power bills, congested data flows and incompatible security techniques.&nbsp;
Without a plan, energy consumption and carbon emissions will rise in lockstep with data production and utilization.
“Inefficient data center infrastructure will result in companies’ emitting increasing amounts of carbon as compute demand increases,” says Patrick Cassleman, Sr. Director of the Sustainability Center of Excellence at Intel.&nbsp;
Up to 73% of data created by enterprises goes unused but is stored anyway, according to <a href="https://www.forrester.com/blogs/hadoop-is-datas-darling-for-a-reason/">Forrester</a>. AI is part of the problem, but it may also be part of the solution.&nbsp;
Processes with built-in AI capabilities are helping to improve data compression and energy use. The 4th Gen Intel® Xeon® Scalable processor, for example, has a range of AI abilities that make data storage more efficient, and less costly to businesses and the environment.

Companies will either have to generate less data or become more efficient regarding how that data is stored to limit the financial burden and environmental impact of emerging technologies.&nbsp;
The 4th Gen Intel® Xeon® Scalable processors can consolidate five servers worth of data into a single server, freeing up data center rack space by 80% and reducing CO2 emissions by 60%.
By upgrading to processors that are both purpose-built to handle workflows from generative AI, and incorporate the capabilities of AI, data center operators get more control over their environment, particularly around power consumption.&nbsp;
With a new feature called Optimized Power Mode, the 4th Gen Intel® Xeon® AI accelerator engine anticipates the power needs of its customers, enabling servers to save power in the utilization range that customers run the most, providing up to 20% of power socket saving with little to no performance impact, according to Cassleman. The Quick Assist feature also speeds up crypto processing for secure storage.
Mostly, though, enterprises will need to become more economical about what data gets stored and for how long. More efficient processors are one very important step in this equation.&nbsp;
"AI can help determine what data is valuable, necessary, and of high enough quality to warrant storage,” Cassleman says. “Superfluous data can simply be discarded, saving both cost and energy."

Data Is Expensive to Store. And Companies Keep Making More of It

The prevalence of a hybrid work model has swiftly increased companies’ adoption of cloud and cloud-based apps. According to <a href="https://learn.pluralsight.com/resource/offers/2023/state-of-cloud" target="_blank">Pluralsight</a>, 70% of organizations have at least half their infrastructure in the cloud. Migrating to the cloud enables companies with many innovative AI capabilities like real-time fraud protection, predictive maintenance, and financial forecasting.&nbsp;However, consolidating data in the cloud also increases the likelihood of a data breach. For organizations, activating sensitive data while remaining secure and compliant is a big challenge in the AI adoption journey.&nbsp;Part of the problem is that the cloud enables workers to access company info on devices that sit outside traditional on-premises security. Legacy IT systems often verify users and devices once without continuous re-validation. This leaves cloud environments vulnerable to infiltration.

AT&T_QS3_D_Preview.png

AT&T_QS3_M_Preview.png

AT&T_QS3_D_V1_1.5.mp4

AT&T_QS3_M_V1_1.5.mp4

Turning to the cloud doesn’t have to be such a concern if companies take the right approach. Zero trust cybersecurity can mitigate these cloud risks by assuming any user, device, or application can be a threat, and requiring continuous authentication before granting access. Projected to be the fastest growing network security market segment, according to <a href="https://www.gartner.com/en/newsroom/press-releases/2022-10-13-gartner-identifies-three-factors-influencing-growth-i" target="_blank">Gartner</a>, zero trust adds new levels of threat intelligence, anomalous behavior detection and multi-factor authentication, doing away with outdated perimeter-based security methods.Zero trust abilities are built into the 4th Gen Intel® Xeon® Scalable processors to help businesses take advantage of the flexibility and scalability of the cloud while reducing the risk of exposing sensitive data.Given the complexity of today’s connected environment—including the sheer number of devices and users, public and private clouds, and as-a-service offerings—it’s important for companies to adopt zero trust principles early in their AI journeys.

Cloud security is imperative, but shouldn't limit data utilization. With appropriate measures, companies can reduce risk while harnessing data's full potential.When secured, data opens up AI’s most innovative capabilities. Businesses can put their data to work to accomplish everything from detecting fraud and developing more responsive supply chains to training breakthrough AI models.When creating a data strategy, business leaders need to foster a culture that sees security as a means of bolstering innovation, rather than treating it as an either/or situation. The 4th Gen Intel® Xeon® Scalable processors empower organizations to have their data, and protect it too, with built-in Intel® Security Engines. These components apply confidential computing—a cloud technology that allows data to be used while reducing the risk of exposure. This opens new possibilities for businesses to harness data that was previously too sensitive or regulated to be put into AI models and other purposes.

As More Organizations Move to the Cloud, They Need to Up Their Cyber Defenses

AI is transforming businesses everywhere with improved process efficiencies, customer experience, and employee productivity. Excited about new possibilities, organizations are eagerly integrating artificial intelligence into their business strategies. According to <a href="https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2023-generative-ais-breakout-year" target="_blank">McKinsey</a>, half of businesses say they have adopted AI.But, despite the excitement, there are still many barriers to adoption. <a href="https://www.ibm.com/watson/resources/ai-adoption" target="_blank">According to IBM</a>, these barriers include lack of AI skills and expertise (34%), cost (29%), lack of tools and platforms (25%), complexity to integrate and scale (24%), and data complexity (24%). Enterprises have tried to address these challenges in a myriad of ways, leading to varying success with AI. Instead, what is needed is a structured approach that begins by understanding the business needs and the AI requirements to support it.

AT&T_QS2_D_V1_1.5_(0-00-03-09).png

AT&T_QS2_M_V1_1.5_(0-00-00-00).png

AT&T_QS2_D_V1_1.5.mp4

AT&T_QS2_M_V1_1.5.mp4

Launching any AI project requires a significant investment of resources. With this commitment of time and money, businesses need to make sure they don’t make it any more complicated. To simplify their AI journey, business leaders should focus on what problems they’re trying to address to better zero in on what AI foundations—be it model development, improved latency or data integrity—they need to focus on.Intel helps organizations minimize complexity by enabling them to integrate AI into the open-standards tools they already use. Intel has a large portfolio of software to help developers accelerate their AI journey, including hundreds of <a href="https://docs.openvino.ai/2021.4/omz_models_group_intel.html" target="_blank">pretrained AI models</a> and <a href="http://www.intel.com/aiplatform" target="_blank">reference kits</a>. An active participant of and contributor to hundreds of open source projects, Intel has optimized software directly into popular software frameworks like <a href="https://www.intel.com/content/www/us/en/developer/tools/oneapi/optimization-for-pytorch.html#gs.4vdx11" target="_blank">Pytorch</a> and <a href="https://www.intel.com/content/www/us/en/developer/tools/oneapi/optimization-for-tensorflow.html#gs.4vgs62" target="_blank">Tensorflow</a>.&nbsp;Intel also empowers developers to streamline their AI efforts with the 4th Gen Intel® Xeon® Scalable processors, which help businesses leverage and scale their AI from day one. These processors have built-in AI accelerators that speed up training and inference. Instead of requiring the purchase of additional hardware for every use case, the processors’ deep learning capabilities enable companies to easily build out the most popular AI use cases, like natural language processing (NLP), chatbots, recommendation systems and image recognition.

To operationalize AI from concept to production, business leaders need the hardware and software portfolio to help them deploy from anywhere, faster. Intel has solutions for every leg of the AI journey, starting with open source software for the AI pipeline, all the way through to data processing, model development and deployment. One Intel tool that increases efficiency is <a href="https://www.intel.com/content/www/us/en/developer/tools/openvino-toolkit/overview.html" target="_blank">OpenVino</a>, which allows developers to write one set of code and deploy it anywhere that’s needed. OpenVino also supports deep learning models including the large language models.Organizations also need to ensure they think through their cyber strategy as they scale their AI, leveraging partners that put security first. Using confidential compute technology, Intel ensures data is encrypted, that access to sensitive data can be better protected, and that multi-party AI models can be built without accessing other party’s data.While AI can help businesses improve revenue, efficiency and productivity, scaling the technology is also a journey. Intel’s <a href="https://www.intel.com/content/www/us/en/developer/topic-technology/artificial-intelligence/platform.html" target="_blank">robust hardware and software solution portfolio</a> helps ensure it is a successful one.