I’ve observed numerous such plans being crafted over the past year as businesses begin to embrace one of the most transformative technological shifts ever, indeed, the only one of its kind. True, we had the advent of the Internet and cloud computing, along with Moore’s law shrinking hardware, but we never had technologies truly capable of passing the Turing test as they can today.

When examining all the components involved in any AI initiative, two primary foundational elements emerge – one is data, the digital flow necessary to interact with a neural network. The second is energy – the power required to operate these systems.

After all, our central nervous systems rely on a specific type of power as well. The data centers of the AI era will demand substantial amounts of energy, so you might consider energy as the physical counterpart complementing data, which represents the actual content within the system.

With this duality in mind, let us delve into how companies are advancing these efforts.

An article recently published on Supermicro discusses “a comprehensive approach to AI focusing not solely on introducing cutting-edge hardware but on the entire AI stack, from computation to network architecture and energy efficiency.”

Supermicro is notably a contractor for the xAI Colossus project, one of the largest data centers globally.

I covered its massive construction early on, as Musk continued to double down on the number of GPUs included in the project.

In any case, the liquid cooling systems developed at Supermicro serve as an example of applying energy concepts to the AI domain.

Read the article to see how Supermicro has innovated its design.

“These may look like standard servers, but the infrastructure needed to cool and power them at this scale does not exist,” says Johnson Eung, a senior growth products manager in AI supercomputing for Supermicro. “So we’ve worked with customers to create this from scratch, providing clear guidelines to ensure it’s done responsibly, safely, and sustainably.”

The Department of Energy is also weighing in on these processes. Highlighting the scale of AI growth, agency leaders are offering recommendations and assessing how all this operates in enterprise settings.

Here’s a list of stakeholders outlined by the department for collaboration:

· Hyperscalers: Amazon, Google, Meta, Microsoft, OpenAI

· Data center developers/innovators: Blackstone/QTS Data Centers, Digital Realty, Verrus

· Technology providers: Fervo, General Electric, Hitachi, Intel, HPE, Long Duration Energy Storage Council, Nvidia

· Electricity companies: Associated Electric Cooperative, Constellation, Duke Energy, Evergy, NPPD, NextEra, PPL, Portland General, PSEG, Southern Company/Georgia Power, Vistra

· Independent system operators and regional transmission operators: CAISO, MISO, PJM, SPP

· Environmental NGOs: NRDC

· Researchers: Association for Computing Machinery, Brattle, Caltech, Carnegie Mellon, Department of Energy, EPRI, Johns Hopkins, IEEE, LBNL, MIT Lincoln Lab, NYU, UC-Santa Barbara, University of Chicago

You can read the remainder of the report here.

Interviews with Founders and Leaders

I gained further insights into this type of architecture during two IIA interviews in April.

I spoke with Sridhar Ramaswami of Snowflake about maintaining a marketplace for data, with collaborative approaches and cross-cloud systems.

He discussed AI as a new medium of interaction and shared product philosophies for chatbots and workflows.

“AI must be user-friendly, efficient, and trustworthy,” Ramaswami stated.

He also touched upon customer service models.

“When it comes to investments, we don’t require our customers to commit to purchasing a specific amount of AI; it’s very much a pay-as-you-go model,” Ramaswami noted. “And the final point on trustworthiness is something we emphasize continuously, which is that every software product inherently understands what is right and wrong.”

Another interview was with Chase Lochmiller and Nadav Eiron of Cruso.

This is where we explored energy in the equation.

Discussing “vertically integrated AI infrastructure” and the fundamental role of AI in business, Lochmiller revisited the concept of data usage supported by energy:

“AI infrastructure at scale leads to the convergence of energy and computing,” he remarked. “Running AI at a significant scale simply demands vast amounts of energy… It’s crucial to contextualize scale because sometimes, if I just throw out huge numbers, they lose meaning.”

He explained that the company is considering a project in Abilene intended to consume 1.2 gigawatts of power and mentioned how the firm aims to support this system.

“We’ve adopted an energy-first approach, meaning we bring the demand for computing to areas where we can access low-cost, clean, abundant energy, energy we can effectively produce at scale in a cost-effective manner,” he said. “We can deliver computing infrastructure that is both cleaner and cheaper than existing infrastructure in major markets like Virginia. We achieve this by making clean energy solutions the more affordable way to generate energy for the first time in history.”

Thus, powering the appropriate data sets results in the kind of equation you aim to promote when constructing AI data centers of the future.

I’ll continue sharing these perspectives as we brainstorm the optimization of data centers in 2025 and beyond.

The above is the detailed content of Data And Energy Are Building Blocks For AI. For more information, please follow other related articles on the PHP Chinese website!