📌 TOPINDIATOURS Breaking ai: Microsoft’s new datacenter mimics ‘one massive AI sup

Microsoft has announced a new wave of datacenters built specifically to power artificial intelligence (AI) workloads, including what it describes as its “largest and most sophisticated AI factory yet” in Wisconsin.

The facility, named Fairwater, is the first of several identical datacenters under construction in the U.S.

In addition to Wisconsin, Microsoft revealed plans for a hyperscale AI datacenter in Narvik, Norway, in partnership with nScale and Aker JV. In the U.K., it will work with nScale to build what it calls the country’s largest supercomputer.

According to the company, these facilities represent “tens of billions of dollars of investments and hundreds of thousands of cutting-edge AI chips” across its global network of more than 400 datacenters.

The Wisconsin site spans 315 acres in Mt. Pleasant, with three buildings totaling 1.2 million square feet. Microsoft says the project required 46.6 miles of deep foundation piles, 26.5 million pounds of steel, and 120 miles of underground cable.

Unlike conventional datacenters designed for independent workloads such as email or web hosting, Fairwater has been built as “one massive AI supercomputer” powered by hundreds of thousands of NVIDIA GPUs.

As per a blog post by the company, Fairwater will deliver “10X the performance of the world’s fastest supercomputer today.”

The claim, while bold, highlights Microsoft’s ambition to compete in the high-performance computing race that underpins AI development.

Building AI’s new factories

The facility’s architecture uses NVIDIA GB200 servers, interconnected in large clusters to enable parallel AI training.

Each rack contains 72 GPUs linked with NVLink, providing high-bandwidth communication and pooled memory across chips.

Microsoft says this configuration allows the cluster to process up to 865,000 tokens per second. Future sites in Norway and the U.K. are expected to use NVIDIA’s upcoming GB300 chips.

Microsoft argues its design ensures AI models can be trained at an unprecedented scale.

“By co-engineering the full stack with the best from our industry partners… Microsoft has built the most powerful, tightly coupled AI supercomputer in the world,” the company claimed in its blog post.

The Wisconsin datacenter also incorporates a two-story layout that reduces physical distance between racks to minimize network latency.

At scale, the company says, the system behaves as a single global supercomputer rather than isolated machines.

Cooling, storage, and sustainability

Given the enormous computing density, Fairwater uses a closed-loop liquid cooling system instead of traditional air cooling.

Microsoft said the system “ensures zero water waste,” with liquid continually recirculated after an initial fill.

The facility is supported by one of the world’s largest water-cooled chiller plants.

On the storage side, Microsoft has reengineered Azure Blob Storage to sustain over 2 million transactions per second per account.

The company said this eliminates the need for manual sharding and supports workloads at an exabyte scale.

Satya Nadella, Microsoft’s CEO, underscored the project’s significance in a post on X. “If intelligence is the log of compute… it starts witha lot of compute! And that’s why we’re scaling our GPU fleet faster than anyone else,” he wrote.

He added that Fairwater is “a seamless cluster of hundreds of thousands of NVIDIA GB2 0s, connected by enough fiber to circle the Earth 4.5 times.”

Nadella also emphasized sustainability efforts, noting that the facility uses renewable energy and partners with local communities.

“With Fairwater, we’re charting a new path… designing closed-loop energy systems to meet real-world computing needs,” he said.

Microsoft says similar datacenters are under construction across its 70-plus global regions, with Wisconsin serving as a model for future AI infrastructure.

🔗 Sumber: interestingengineering.com

📌 TOPINDIATOURS Hot ai: Researchers from PSU and Duke introduce “Multi-Agent Syste

Share My Research is Synced’s column that welcomes scholars to share their own research breakthroughs with over 2M global AI enthusiasts. Beyond technological advances, Share My Research also calls for interesting stories behind the research and exciting research ideas. 

Meet the author
Institutions: Penn State University, Duke University, Google DeepMind, University of Washington, Meta, Nanyang Technological University, and Oregon State University. The co-first authors are Shaokun Zhang of Penn State University and Ming Yin of Duke University.

In recent years, LLM Multi-Agent systems have garnered widespread attention for their collaborative approach to solving complex problems. However, it’s a common scenario for these systems to fail at a task despite a flurry of activity. This leaves developers with a critical question: which agent, at what point, was responsible for the failure? Sifting through vast interaction logs to pinpoint the root cause feels like finding a needle in a haystack—a time-consuming and labor-intensive effort.
 
This is a familiar frustration for developers. In increasingly complex Multi-Agent systems, failures are not only common but also incredibly difficult to diagnose due to the autonomous nature of agent collaboration and long information chains. Without a way to quickly identify the source of a failure, system iteration and optimization grind to a halt.
 
To address this challenge, researchers from Penn State University and Duke University, in collaboration with institutions including Google DeepMind, have introduced the novel research problem of “Automated Failure Attribution.” They have constructed the first benchmark dataset for this task, Who&When, and have developed and evaluated several automated attribution methods. This work not only highlights the complexity of the task but also paves a new path toward enhancing the reliability of LLM Multi-Agent systems.
The paper has been accepted as a Spotlight presentation at the top-tier machine learning conference, ICML 2025, and the code and dataset are now fully open-source.

Paper：https://arxiv.org/pdf/2505.00212
Code：https://github.com/mingyin1/Agents_Failure_Attribution
Dataset：https://huggingface.co/datasets/Kevin355/Who_and_When
 
 
Research Background and Challenges
LLM-driven Multi-Agent systems have demonstrated immense potential across many domains. However, these systems are fragile; errors by a single agent, misunderstandings between agents, or mistakes in information transmission can lead to the failure of the entire task.

Currently, when a system fails, developers are often left with manual and inefficient methods for debugging:
Manual Log Archaeology : Developers must manually review lengthy interaction logs to find the source of the problem.
Reliance on Expertise : The debugging process is highly dependent on the developer’s deep understanding of the system and the task at hand.
 
This “needle in a haystack” approach to debugging is not only inefficient but also severely hinders rapid system iteration and the improvement of system reliability. There is an urgent need for an automated, systematic method to pinpoint the cause of failures, effectively bridging the gap between “evaluation results” and “system improvement.”

Core Contributions
This paper makes several groundbreaking contributions to address the challenges above:
1. Defining a New Problem: The paper is the first to formalize “automated failure attribution” as a specific research task. This task is defined by identifying the

2. failure-responsible agent and the decisive error step that led to the task’s failure.

Constructing the First Benchmark Dataset: Who&When : This dataset includes a wide range of failure logs collected from 127 LLM Multi-Agent systems, which were either algorithmically generated or hand-crafted by experts to ensure realism and diversity. Each failure log is accompanied by fine-grained human annotations for:
Who: The agent responsible for the failure.
When: The specific interaction step where the decisive error occurred.
Why: A natural language explanation of the cause of the failure.

3. Exploring Initial “Automated Attribution” Methods : Using the Who&When dataset, the paper designs and assesses three distinct methods for automated failure attribution:
All-at-Once: This method provides the LLM with the user query and the complete failure log, asking it to identify the responsible agent and the decisive error step in a single pass. While cost-effective, it may struggle to pinpoint precise errors in long contexts.
Step-by-Step: This approach mimics manual debugging by having the LLM review the interaction log sequentially, making a judgment at each step until the error is found. It is more precise at locating the error step but incurs higher costs and risks accumulating errors.
Binary Search: A compromise between the first two methods, this strategy repeatedly divides the log in half, using the LLM to determine which segment contains the error. It then recursively searches the identified segment, offering a balance of cost and performance.
 
Experimental Results and Key Findings

Experiments were conducted in two settings: one where the LLM knows the ground truth answer to the problem the Multi-Agent system is trying to solve (With Ground Truth) and one where it does not (Without Ground Truth). The primary model used was GPT-4o, though other models were also tested. The systematic evaluation of these methods on the Who&When dataset yielded several important insights:

• A Long Way to Go: Current methods are far from perfect. Even the best-performing single method achieved an accuracy of only about 53.5% in identifying the responsible agent and a mere 14.2% in pinpointing the exact error step. Some methods performed even worse than random guessing, underscoring the difficulty of the task.
• No “All-in-One” Solution: Different methods excel at different aspects of the problem. The All-at-Once method is better at identifying “Who,” while the Step-by-Step method is more effective at determining “When.” The Binary Search method provides a middle-ground performance.

• Hybrid Approaches Show Promise but at a High Cost: The researchers found that combining different methods, such as using the All-at-Once approach to identify a potential agent and then applying the Step-by-Step method to find the error, can improve overall performance. However, this comes with a significant increase in computational cost.

• State-of-the-Art Models Struggle: Surprisingly, even the most advanced reasoning m…

  Konten dipersingkat otomatis.

  🔗 Sumber: syncedreview.com

  ---