Brief

DeepSeek is actively recruiting for an IDC Design and Planning Engineer, signaling a significant move towards building its own large-scale data centers. This initiative, likely fueled by a recent substantial funding round, indicates the company's ambition to control its computing infrastructure, moving beyond rented capacity. The role involves planning and constructing facilities ranging from megawatts to gigawatts, focusing on advanced technologies like liquid cooling and high-density GPU clusters to support future AI model development and agent product lines. AI

Japanese telecom giant NTT is reportedly planning to establish a $500 million fund focused on optical networks. The company aims to raise capital from investors in the US and South Korea for this initiative. Approximately 20 companies are expected to participate in the fund, including Samsung Electronics. AI

Alibaba is exploring the use of small modular nuclear reactors to power its data centers in Hangzhou. The company has engaged in discussions with a Chinese state-owned nuclear power enterprise regarding this potential energy solution. However, significant hurdles related to pricing and the integration of these reactors into the existing power grid must still be overcome. AI

Algebra AI, a startup based in Dubai, has successfully raised $7 million to advance its AI-as-a-Service model. The company aims to manage AI systems for 30,000 small and medium-sized enterprises that struggle with implementing automation independently. This funding will support their expansion and development efforts in the AI management space. AI

ArchAstro has launched with $6.2 million in pre-seed funding to develop AI agents. These agents are designed to automate software deployments and integrations across different companies. The startup was founded by former engineers from major tech firms including Microsoft, Stripe, Statsig, and Meta. AI

Google is reportedly making significant changes to its search engine, moving towards a more AI-driven product. This shift is part of a broader trend of product alterations within Google, influenced by their advancements in AI technology. The company's focus on AI, particularly with models like Gemini, is reshaping its core offerings. AI

Apple's revamped Siri now utilizes Foundation Models that were co-developed with Google. The most complex queries handled by Siri will be processed on Google's servers. This reliance on a competitor marks a notable change for Apple, which has historically positioned itself as a privacy-focused alternative to Google. AI

Fujikura is poised to exceed its financial targets due to strong demand for fiber-optic cables, which are critical for AI data centers. The company also plans to implement price increases to further boost its performance. This positive outlook is attributed to the essential role of its products in supporting the growing AI infrastructure. AI

China's lab-grown diamond industry is experiencing a surge in demand due to its application as a cooling material for high-performance AI chips. Chinese manufacturers have begun commercial shipments, with these synthetic diamonds playing a crucial role in managing the heat generated by increasingly powerful AI semiconductors. This unexpected demand highlights the material's importance in advancing AI technology. AI

Apple's recent WWDC26 keynote introduced new macOS systems, with a particular focus on integrating third-party AI applications locally. This approach aims to shift AI processing from the cloud to personal devices, enhancing privacy and reducing reliance on external services. Demonstrations included Perplexity, Draw Things, and LM Studio, showcasing how Macs can function as personal AI hubs. AI

Sandstone, a legal tech startup, has secured $30 million in funding to develop AI tools for in-house legal departments. The company aims to streamline legal operations and enhance efficiency for corporate legal teams. This funding round is expected to accelerate Sandstone's product development and market expansion. AI

LG is expanding its presence in Vietnam, driven by the growing global demand for AI, 5G, and semiconductors. This strategic move aims to strengthen its position within the evolving international supply chains for these critical technologies. AI

SpaceX is preparing for an IPO with a projected valuation of $1.77 trillion, significantly higher than its 2024 valuation of $350 billion. While rockets are a core part of its business, investors are increasingly viewing the company as a communications and infrastructure provider. The Starlink satellite internet service, which already has over 10 million subscribers, is a major driver of this valuation, generating over $1 billion in profit in early 2026. The development of Starship, a super heavy-lift reusable rocket, is expected to further reduce launch costs and enable the construction of large-scale infrastructure in orbit, such as data centers and solar power systems. AI

Employees at Amazon have urged Seattle officials to halt the construction of new data centers. The employees cite concerns about the environmental impact and the strain these facilities place on the local power grid. They are advocating for a moratorium on new data center development until these issues can be adequately addressed. AI

US Steel plans to invest up to $2.5 billion in a new hot-rolled strip mill in Braddock, Pennsylvania. This facility will replace an aging 87-year-old mill located nearby. The company released an economic impact analysis report detailing these plans. AI

The UK government has announced a significant investment of £1.1 billion to bolster its AI infrastructure. This substantial funding aims to accelerate AI development and adoption across the nation. The initiative is seen as a strategic move to enhance the UK's AI capabilities and promote technological sovereignty within Europe. AI

Two new research papers explore the challenges of running large language models (LLMs) efficiently. The first paper investigates the performance trade-offs of deploying LLMs on edge devices like smartphones and specialized NPUs, highlighting thermal constraints and memory bandwidth limitations. The second paper introduces a scalable framework using heuristic algorithms to optimize resource allocation for LLM inference in heterogeneous GPU cloud environments, aiming to meet service level objectives while minimizing costs. AI

Researchers have introduced HNTL (Hierarchical No-pointer Tangent-Local), a new framework for vector memory systems designed to improve the efficiency of approximate nearest neighbor searches. This method partitions high-dimensional space into local segments, representing vectors using tangent spaces and a pointerless layout to reduce memory overhead and enhance CPU performance. Benchmarks show HNTL achieves high recall rates with a smaller candidate pool and offers a significant speedup over traditional pointer-chasing methods. AI

Researchers have developed and benchmarked a convolutional sparse coding implementation using the Locally Competitive Algorithm (LCA) on Intel's Loihi 2 neuromorphic hardware. This work represents the first known implementation and evaluation of convolutional LCA on this platform. The study aims to determine the conditions under which this approach becomes advantageous for sparse inference on neuromorphic systems, positioning it as a benchmark for structured sparse inference. AI

The AI chip supply chain is at a critical juncture, marked by a developing conflict between Broadcom and Google. This competition is poised to significantly reshape the landscape of AI hardware. AI

The stock market showed resilience as AI-related companies began to rebound from a recent downturn. Companies specializing in the production of computer chips and memory components, which are crucial for the AI industry, were among the top performers. This recovery suggests a stabilization in the market following a period of volatility. AI

Apple is expanding its Private Cloud Compute (PCC) infrastructure beyond its own data centers, partnering with Google and NVIDIA. This expansion allows Apple Intelligence workloads to run on Google Cloud, utilizing NVIDIA GPUs and Google's confidential computing technologies. The move aims to extend Apple's stringent privacy and security commitments to third-party cloud environments for more complex AI tasks. AI

Apple is reportedly planning to use Google's cloud infrastructure to power its "Private Cloud Compute" feature for AI tasks. This move would allow Apple to process sensitive user data on remote servers while maintaining a level of privacy. The exact details of the partnership and the scope of data processed remain unclear. AI

The world's first wind-powered underwater data center has begun operations off the coast of Shanghai, China. This innovative facility is designed to be more energy and water-efficient than traditional land-based data centers. Separately, the open-source photo management software digiKam has released version 9.1, introducing support for Pixel motion photos and enhanced timezone capabilities. AI

Singapore is increasingly relying on underground district cooling systems to combat rising temperatures and reduce its reliance on air conditioning. This concept, which uses less electricity than individual air conditioners, is being expanded across various neighborhoods, including the large Marina Bay network. Experts believe Singapore can become a model for energy-efficient cooling solutions in hot and humid climates, with the local market for this technology projected to double in the next decade. AI

Researchers have developed a novel five-phase pipeline utilizing LLM-based agents to translate legacy Fortran code into the JAX framework. This system automates the process of code migration, including dependency analysis, autonomous error correction, and numerical parity enforcement. The pipeline was successfully applied to CLM-ml-v2, a 19,000-line land surface model, resulting in a differentiable version that significantly speeds up computation and parameter recovery. AI

A research group, including Huawei and institutions from Shenzhen, claims to have successfully completed full-parameter post-training on DeepSeek's 1.6 trillion parameter V4-Pro model. This was achieved using a cluster of at least 1,000 Huawei Ascend 910C AI chips. This development is seen as a significant step towards China's AI self-reliance, particularly in overcoming challenges with training complex models on domestic hardware, though specific performance benchmarks are currently absent. AI

Vodafone has launched a new 5G Broadband service in the UK, aiming to provide fiber-like internet speeds without the need for cable installation. The service utilizes a 5G router and offers speeds up to 150Mbps, with pricing starting at £19 per month for a 50Mbps connection. This launch is supported by a survey indicating that slow or unreliable broadband is a major frustration for consumers, impacting daily life and even relationships. AI

NVIDIA has open-sourced parts of its cuDNN library, a significant move after 12 years of it being closed-source. This release includes over 20 Mixture-of-Experts (MoE) kernels and NSA sparse attention kernels. The codebase for these kernels is largely written in Python CuTe-DSL, with public documentation now available. AI

Alibaba's Qwen team has released FlashQLA, a new set of high-performance linear attention kernels developed using TileLang. These kernels are designed to improve the efficiency of attention mechanisms in large language models. The team also shared benchmark results for their Qwen models, showcasing performance across various configurations. AI

Google DeepMind has introduced Decoupled DiLoCo, a novel approach to training advanced AI models that enhances resilience and flexibility across data centers. This system can train models like Google's 12B Gemma model across geographically dispersed regions using low-bandwidth networks and can even mix different generations of hardware, such as TPU6e and TPUv5p. Decoupled DiLoCo is designed to be self-healing, isolating and continuing training through artificial hardware failures and reintegrating units when they come back online, addressing the synchronization issues that typically stall AI training. AI

OpenAI has secured $110 billion in private funding, with Amazon contributing $50 billion and Nvidia and SoftBank each adding $30 billion, valuing the company at $730 billion pre-money. This significant investment includes substantial infrastructure partnerships, with OpenAI expanding its AWS collaboration by $100 billion and committing to significant compute usage. The funding round is still open, and OpenAI anticipates further investor participation as it focuses on scaling infrastructure to meet the growing demand for AI services. AI

Apple has acquired the Israeli AI startup Q.ai for nearly $2 billion, aiming to bolster its capabilities in audio processing and machine learning. The startup, founded in 2022, specializes in technologies that can interpret whispered speech and enhance audio in noisy environments. This acquisition is Apple's second-largest to date and follows previous AI-focused feature integrations in products like AirPods and the Vision Pro headset. AI

Nvidia's latest financial results indicate a continued strong demand for AI infrastructure, with significant revenue generated from its AI chip sales. The company's performance highlights the ongoing substantial investment in hardware necessary to support the rapidly expanding AI sector. This robust spending suggests that the development and deployment of advanced AI models remain a top priority for many organizations. AI

The rapid expansion of AI is creating a significant bottleneck in the United States due to the limitations of its power grid, contrasting sharply with China's robust energy infrastructure. While U.S. AI growth is hampered by debates over data center power consumption and grid stability, China has proactively addressed this by overbuilding its power capacity over decades. This strategic oversupply allows China to integrate AI data centers as a means to absorb excess energy, a situation unimaginable in the U.S. where grids often operate with minimal reserve margins, leading to concerns about the sustainability of AI development. AI

AI companies are aggressively crawling open-source infrastructure, causing significant outages and disruptions for projects like SourceHut, KDE GitLab, and GNOME. These AI scrapers often disregard robots.txt and mimic legitimate user agents, making it difficult to implement effective defenses. As a result, some projects have resorted to implementing challenging proof-of-work systems to block these bots, which can also impact legitimate users. AI

Intel has effectively dismantled Habana Labs, an Israeli AI chip startup it acquired for $2 billion, marking a significant failure in its attempt to compete with Nvidia. Despite initial optimism and a deal with Amazon for its Gaudi chips, Intel's internal issues and integration problems led to key personnel departing and the cancellation of next-generation products like Falcon Shores. This outcome represents a rare misstep for Habana's founder, Avigdor Willenz, who has a history of successful ventures in the semiconductor industry. AI

The White House has issued an executive order aimed at bolstering U.S. leadership in AI infrastructure. The order focuses on expanding access to computing resources, developing AI talent, and promoting responsible AI innovation. It also emphasizes the importance of international collaboration and the development of safety standards for AI technologies. AI

Vultr, a cloud computing provider focused on AI workloads, has secured $333 million in funding at a $3.5 billion valuation. The investment round was led by existing investor Thoma Bravo. The company plans to use the funds to expand its global infrastructure and enhance its AI-specific offerings. AI

OpenAI has entered into a new agreement to utilize Oracle Cloud Infrastructure (OCI) for its artificial intelligence workloads. This partnership aims to expand OpenAI's existing AI platform, which is primarily hosted on Microsoft Azure. The collaboration will leverage OCI's high-performance computing capabilities to support OpenAI's growing demand for AI training and inference. AI

The AI sector's expenditure on Nvidia chips significantly outpaced its revenue generation, with a reported 17x difference. This highlights a substantial investment phase in AI infrastructure, potentially indicating a focus on future growth and capability development over immediate profitability. The data suggests a considerable capital outlay is being made to acquire the necessary hardware for training and deploying advanced AI models. AI

Researchers have developed a software stack called 'this http URL' to enable fast and lossless inference of 1-bit Large Language Models (LLMs) like BitNet b1.58 on CPUs. This new infrastructure achieves significant speedups, ranging from 2.37x to 6.17x on x86 CPUs and 1.37x to 5.07x on ARM CPUs, depending on model size. The goal is to make LLMs more efficient and deployable on a wider range of devices. AI