Frontier AI development relies on powerful AI supercomputers, yet analysis of these systems is limited. We create a dataset of 500 AI supercomputers from 2019 to 2025 and analyze key trends in performance, power needs, hardware cost, ownership, and global distribution. We find that the computational performance of AI supercomputers has doubled every nine months, while hardware acquisition cost and power needs both doubled every year. The leading system in March 2025, xAI's Colossus, used 200,000 AI chips, had a hardware cost of \$7B, and required 300 MW of power, as much as 250,000 households. As AI supercomputers evolved from tools for science to industrial machines, companies rapidly expanded their share of total AI supercomputer performance, while the share of governments and academia diminished. Globally, the United States accounts for about 75% of total performance in our dataset, with China in second place at 15%. If the observed trends continue, the leading AI supercomputer in 2030 will achieve $2\times10^{22}$ 16-bit FLOP/s, use two million AI chips, have a hardware cost of \$200 billion, and require 9 GW of power. Our analysis provides visibility into the AI supercomputer landscape, allowing policymakers to assess key AI trends like resource needs, ownership, and national competitiveness.

The whiskers extend to 4x the IQR ( Supplementary Figure 1), which is a conservative threshold that ensures only the most extreme variations in PL are classified as outliers (denoted by diamond symbols). Outliers in PL property distributions may indicate experimental errors, sample inconsistencies, or data proces sing anomalies, thus, they are removed from the ML analysis. Data Visualization: PCA orthogonally transforms the original variables into a new set of linearly uncorrelated variables termed principal components (PCs). The first PC captures the maximum variance present in the data, and each subsequent component has the highest variance p ossible under the constraint of being orthogonal to the preceding ones. The methodology involves standardizing the dataset, calculating the covariance matrix, and then extracting the eigenvalues and eigenvectors of this matrix, which, in tur n, dictate the magnitude and direction of the new space, respectively. By projecting the original data along these new axes, PCA provides a means to reduce the dimensionality of the dataset. Supplementary Figure 1A illustrates the distribution of the samples in the space defined by the PCs, with each point representing a single sample's location within this novel coordinate system. Here, the colors indicate the value of each PL property, offering a visual insight into how these factors correlate with the PCs.

As demand for Large Language Models (LLMs) and AI agents rapidly grows, optimizing systems for efficient LLM inference becomes critical. While significant efforts have focused on system-level engineering, little is explored from a mathematical modeling and queuing perspective. In this paper, we aim to develop the queuing fundamentals for large language model (LLM) inference, bridging the gap between the queueing theory and LLM system communities. In particular, we study the throughput aspect in LLM inference systems. We prove that a large class of 'work-conserving' scheduling algorithms can achieve maximum throughput for individual inference LLM engine, highlighting 'work-conserving' as a key design principle in practice. In a network of LLM agents, work-conserving scheduling alone is insufficient, particularly when facing specific workload structures and multi-class workflows that require more sophisticated scheduling strategies. Evaluations of real-world systems show that Orca and Sarathi-serve are throughput-optimal, reassuring practitioners, while FasterTransformer and vanilla vLLM are not maximally stable and should be used with caution. Our results highlight the substantial benefits that the queueing community can offer in improving LLM inference systems and call for more interdisciplinary development.

This paper introduces a novel federated learning framework termed LoRa-FL designed for training low-rank one-shot image detection models deployed on edge devices. By incorporating low-rank adaptation techniques into one-shot detection architectures, our method significantly reduces both computational and communication overhead while maintaining scalable accuracy. The proposed framework leverages federated learning to collaboratively train lightweight image recognition models, enabling rapid adaptation and efficient deployment across heterogeneous, resource-constrained devices. Experimental evaluations on the MNIST and CIFAR10 benchmark datasets, both in an independent-and-identically-distributed (IID) and non-IID setting, demonstrate that our approach achieves competitive detection performance while significantly reducing communication bandwidth and compute complexity. This makes it a promising solution for adaptively reducing the communication and compute power overheads, while not sacrificing model accuracy.

-- This study explores alternative framework configuration s for adapting thermal machine learning (ML) models for power converters b y combining transfer learning (TL) and federated learning (FL) in a piecewise manner . This approach inherently addresses challenges such as varying operating conditions, data sharing limitations, and security implications. The framework starts with a base model that is incrementally adapted by multiple clients via adapting three state - of - the - art domain adaptation techniques: Fine - tuning, Transfer Component Analysis (TCA), and Deep Domain Adaptation (DDA). The Flower framework is employed for FL, using Federated Averaging for aggregation. Validation with field data demonstrates that fine - tuning offers a straightforward TL approach with high accuracy, making it suitable for practical applications. Benchmarking results reveal a comprehensive comparison of thes e methods, showcasing their respective strengths and weaknesses when applied in different scenarios. L ocally hosted FL enhances performance when data aggregation is not feasible, while cloud - based FL becomes more practical with a significant increase in the number of clients, addressing scalability and connectivity challenges.

In a temporal graph the edge set dynamically changes over time according to a set of time-labels associated with each edge that indicates at which time-steps the edge is available. Two vertices are connected if there is a path connecting them in which the edges are traversed in increasing order of their labels. We study the problem of scheduling the availability time of the edges of a temporal graph in such a way that all pairs of vertices are connected within a given maximum allowed time $a$ and the overall number of labels is minimized. The problem, known as \emph{Minimum Aged Labeling} (MAL), has several applications in logistics, distribution scheduling, and information spreading in social networks, where carefully choosing the time-labels can significantly reduce infrastructure costs, fuel consumption, or greenhouse gases. The problem MAL has previously been proved to be NP-complete on undirected graphs and \APX-hard on directed graphs. In this paper, we extend our knowledge on the complexity and approximability of MAL in several directions. We first show that the problem cannot be approximated within a factor better than $O(\log n)$ when $a\geq 2$, unless $\text{P} = \text{NP}$, and a factor better than $2^{\log ^{1-ε} n}$ when $a\geq 3$, unless $\text{NP}\subseteq \text{DTIME}(2^{\text{polylog}(n)})$, where $n$ is the number of vertices in the graph. Then we give a set of approximation algorithms that, under some conditions, almost match these lower bounds. In particular, we show that the approximation depends on a relation between $a$ and the diameter of the input graph. We further establish a connection with a foundational optimization problem on static graphs called \emph{Diameter Constrained Spanning Subgraph} (DCSS) and show that our hardness results also apply to DCSS.

Autonomous AI systems reveal foundational limitations in deterministic, human-authored computing architectures. This paper presents Cognitive Silicon: a hypothetical full-stack architectural framework projected toward 2035, exploring a possible trajectory for cognitive computing system design. The proposed architecture would integrate symbolic scaffolding, governed memory, runtime moral coherence, and alignment-aware execution across silicon-to-semantics layers. Our design grammar has emerged from dialectical co-design with LLMs under asymmetric epistemic conditions--creating structured friction to expose blind spots and trade-offs. The envisioned framework would establish mortality as a natural consequence of physical constraints, non-copyable tacit knowledge, and non-cloneable identity keys as cognitive-embodiment primitives. Core tensions (trust/agency, scaffolding/emergence, execution/governance) would function as central architectural pressures rather than edge cases. The architecture theoretically converges with the Free Energy Principle, potentially offering a formal account of how cognitive systems could maintain identity through prediction error minimization across physical and computational boundaries. The resulting framework aims to deliver a morally tractable cognitive infrastructure that could maintain human-alignment through irreversible hardware constraints and identity-bound epistemic mechanisms resistant to replication or subversion.

Node Assigned physics-informed neural networks for thermal-hydraulic system simulation: CVH/FL module Jeesuk Shin a,1, Cheolwoong Kim b,1, Sunwoong Yang c, Minseo Lee a, Sung Joong Kim b,, Joongoo Jeon a,d,e, a Department of Applied Plasma and Quantum Beam Engineering, Jeonbuk National University, Jeonju-si, Republic of Korea b Department of Nuclear Engineering, Hanyang University, Seoul, Republic of Korea c Cho Chun Shik Graduate School of Mobility, Korea Advanced Institute of Science and Technology, Daejeon, Republic of Korea d Department of Quantum System Engineering, Jeonbuk National University, Jeonju-si, Republic of Korea e Graduate School of Integrated Energy-AI, Jeonbuk National University, Jeonju-si, Republic of KoreaAbstract Severe accidents (SAs) in nuclear power plants have been analyzed using thermal-hydraulic (TH) system codes such as MELCOR and MAAP. These codes efficiently simulate the progression of SAs, while they still have inherent limitations due to their inconsistent finite difference schemes. The use of empirical schemes incorporating both implicit and explicit formulations inherently induces unidirectional coupling in multi-physics analyses. The objective of this study is to develop a novel numerical method for TH system codes using physics-informed neural network (PINN). They have shown strength in solving multi-physics due to the innate feature of neural networks--automatic differentiation. We propose a node-assigned PINN (NA-PINN) that is suitable for the control volume approach-based system codes. NA-PINN addresses the issue of spatial governing equation variation Corresponding author Corresponding author Email addresses: sungjkim@hanyang.ac.kr (Sung Joong Kim), jgjeon41@jbnu.ac.kr (Joongoo Jeon) 1 These authors contributed equally to this work. In this phase, we evaluated the accuracy of the PINN methods for the hydrodynamic module. In the 6 water tank simulation, PINN and NA-PINN showed maximum absolute errors of 1.678 and 0.007, respectively. It should be noted that only NA-PINN demonstrated acceptable accuracy. To the best of the authors' knowledge, this is the first study to successfully implement a system code using PINN. Our future work involves extending NA-PINN to a multi-physics solver and developing it in a surrogate manner Keywords: FDM, PINN, Thermal-hydraulics, Control-volume approach1. INTRODUCTION Due to the extremely low frequency of severe accident (SA) in nuclear power plants (NPPs) and the limited availability of real-world accident data, SA-related research inevitably relies on the use of system codes to simulate hypothetical accident scenarios and assess the potential safety concerns. Widely used system codes, such as RELAP5/SCDAP, MAAP, and MEL-COR, model the physical behavior of NPP components and simulate accident progression by accounting for complex thermal-hydraulic (TH) and physicochemical interactions arising under SA conditions.

--Quantum computing has the potential to improve our ability to solve certain optimization problems that are computationally difficult for classical computers, by offering new algorithmic approaches that may provide speedups under specific conditions. In this work, we introduce QAOA-GPT, a generative framework that leverages Generative Pretrained Transformers (GPT) to directly synthesize quantum circuits for solving quadratic unconstrained binary optimization problems, and demonstrate it on the MaxCut problem on graphs. T o diversify the training circuits and ensure their quality, we have generated a synthetic dataset using the adaptive QAOA approach, a method that incrementally builds and optimizes problem-specific circuits. The experiments conducted on a curated set of graph instances demonstrate that QAOA-GPT, generates high quality quantum circuits for new problem instances unseen in the training as well as successfully parametrizes QAOA. Our results show that using QAOA-GPT to generate quantum circuits will significantly decrease both the computational overhead of classical QAOA and adaptive approaches that often use gradient evaluation to generate the circuit and the classical optimization of the circuit parameters. Our work shows that generative AI could be a promising avenue to generate compact quantum circuits in a scalable way. Quantum computing is rapidly emerging technology with significant potential across various domains, including finance [1], chemical simulations [2], material science [3], combinatorial optimization [4], and machine learning [5], among others. V ariational quantum-classical algorithms represent one of the most promising classes of quantum algorithms in different domains, showing potential for both fault-tolerant quantum computers and near-term noisy intermediate-scale quantum (NISQ) devices. The Quantum Approximate Optimization Algorithm (QAOA) [6] and many of its subsequent versions and customizations [7] belong to this class and demonstrate great potential due to their problem/application flexibility and compatibility with various quantum architectures. The original QAOA framework employs a fixed ansatz structure, which can limit expressibility and hinder performance, particularly on near-term quantum devices where circuit depth is limited. This rigid design may not capture the problem-specific features needed for efficient optimization. Such methods as ADAPT -QAOA [8] address this challenge by iteratively constructing the ansatz in a problem-informed manner. At each step, ADAPT -QAOA selects operators from a predefined pool based on their gradient with respect to the cost function, incorporating only those that contribute most significantly to improving the objective.

This paper presents the design and implementation of an AI vision-controlled orthotic hand exoskeleton to enhance rehabilitation and assistive functionality for individuals with hand mobility impairments. The system leverages a Google Coral Dev Board Micro with an Edge TPU to enable real-time object detection using a customized MobileNet\_V2 model trained on a six-class dataset. The exoskeleton autonomously detects objects, estimates proximity, and triggers pneumatic actuation for grasp-and-release tasks, eliminating the need for user-specific calibration needed in traditional EMG-based systems. The design prioritizes compactness, featuring an internal battery. It achieves an 8-hour runtime with a 1300 mAh battery. Experimental results demonstrate a 51ms inference speed, a significant improvement over prior iterations, though challenges persist in model robustness under varying lighting conditions and object orientations. While the most recent YOLO model (YOLOv11) showed potential with 15.4 FPS performance, quantization issues hindered deployment. The prototype underscores the viability of vision-controlled exoskeletons for real-world assistive applications, balancing portability, efficiency, and real-time responsiveness, while highlighting future directions for model optimization and hardware miniaturization.