AI compute

OpenAI has announced what it describes as three new discoveries spanning mathematics, computer science, and physics, saying they were made by building scientific understanding together with academic communities. The announcement, published on OpenAI's site, drew heavy attention on Hacker News, topping the front page with over a thousand points of engagement. Readers are debating the substance and significance of the claimed results.

Edsger W. Dijkstra's 1978 essay 'The foolishness of natural language programming' is drawing renewed attention online. In it the Dutch computer scientist argued against the idea that programming languages should resemble human language, warning of the ambiguity inherent in natural language. Readers are debating whether his warnings hold up in the era of large language models and code generation.

A project called GPU-unlock is drawing attention for exploring the FP/BF16 instruction issue rate on Nvidia's RTX 3000 series graphics cards, examining whether performance limits on these consumer GPUs can be bypassed or better understood. Discussion is focused on what this means for compute workloads like machine learning, where BF16 throughput matters, and how much hidden performance Nvidia's cards may hold.

OpenAI says its artificial intelligence systems have produced new discoveries in mathematics, computer science and physics. The claim, reported by Al Jazeera, suggests AI is moving beyond assisting research to generating genuine scientific findings. Details of the discoveries and how they were verified have not been fully disclosed, prompting debate among researchers about the significance and credibility of the announcement.

OpenAI is announcing three new discoveries spanning mathematics, computer science, and physics, covering harmonic analysis, complexity theory, and quantum field theory. The announcement, published on the company's research blog, frames the work as building scientific understanding with academic communities, and is drawing attention from researchers discussing what the claimed breakthroughs actually involve.

SpaceX has struck a $40 billion agreement to lease GPUs, backed by private credit investors, according to Morningstar. The deal points to a major financing push for the company's computing and AI-related infrastructure ambitions, using private credit markets rather than traditional bank lending. It ranks among the largest such arrangements reported, and is drawing attention for the scale of capital private credit funds are now directing into tech hardware.

Micron's DRAM roadmap is drawing attention for its shift toward High-NA EUV lithography, wafer bonding, and vertically stacked memory, aimed at AI systems that need data positioned closer to compute. The approach signals how memory makers are rethinking scaling as conventional cell shrinking slows, with hybrid bonding and 3D stacking emerging as key levers for next-generation AI memory performance.

The creator of OpenClaw says he spent $10 million worth of AI tokens over the past month, according to a report by The Register. The claim has drawn attention because of the sheer scale of compute spending involved for a single developer running an open-source project, prompting discussion about the costs of running AI agents at this intensity.

OpenAI announced that its artificial intelligence systems have made breakthroughs on major open problems in mathematics, computing, and physics. The claim, reported by Quartz, suggests the company's models are now capable of contributing original research-level work rather than only answering routine questions. Details about which specific problems were cracked, and whether outside experts have verified the results, were not provided in the initial report.

Researcher and interviewer Dwarkesh Patel is exploring the long-term future of semiconductor technology, asking what computer chips might look like two decades from now. The question touches on how far Moore's Law can stretch, the limits of silicon, and what manufacturing breakthroughs would be needed to keep advancing computing power toward the mid-2040s.

Ocean Power Technologies has signed a letter of intent for a demonstration of an offshore AI data center. The project would pair the company's ocean-based power systems with computing infrastructure at sea, reflecting growing interest in using ocean space to meet rising data center and AI energy demands. Details on the partner, location, or timeline have not been disclosed.

Oracle has expanded its chip supply agreement with Nvidia, while also adding offerings based on chips from Google and Cerebras, according to a Reuters report. The move widens Oracle's access to advanced AI processors from multiple suppliers as it builds out cloud computing capacity to meet surging demand for artificial intelligence workloads.

Researchers and policymakers are examining the environmental footprint of data centers, which power cloud computing and artificial intelligence but consume large amounts of electricity and water. Discussion focuses on energy use, carbon emissions, and local impacts of new facilities, as governments weigh regulation and companies pledge greener operations.

A science commentary circulating online argues that DNA outclasses billionaire-funded data centers as an information system: just four chemical letters encode the instructions for building and running a human body, from eye color and heartbeat to the senses and brain. The comparison is being shared as a reflection on how compact and efficient biological storage is compared with the vast industrial infrastructure of modern computing.

German AI researcher Klaus Mainzer argues that today's artificial intelligence is nothing compared with what is coming next, pointing to quantum computing-inspired 'quantum brains', photonic chips that process data with light, and the prospect of superintelligence. The claims are drawing attention among German-speaking audiences debating how quickly the next leap in AI hardware and capability could arrive.

Attention is turning to plans for orbital data centres, with the booming energy and computing demands of artificial intelligence driving interest in placing infrastructure in space. The idea has been described as a new space race, as governments and companies weigh the potential of orbit-based data processing against Earth's growing power constraints.

Dozens of data center projects are seeking approval to build on public land in the western United States. The proposals come as demand for computing power, driven largely by artificial intelligence, pushes developers toward large, cheap sites. Conservation groups, including The Wildlife Society, are flagging concerns about the impact on wildlife habitat and publicly owned landscapes.

Tens of thousands of Nvidia GPUs in Meta's Prometheus supercluster are reportedly sitting idle, essentially being held as spares until failures occur elsewhere in the system. The arrangement highlights the scale of Meta's AI infrastructure buildout, where massive hardware inventories are provisioned for redundancy rather than active use, drawing criticism about wasteful capacity and the economics of hyperscale AI computing.

OpenAI says its models have produced new discoveries in mathematics, computer science and physics, according to an announcement reported by Engadget. The claim, if verified, would mark a rare instance of an AI system contributing genuinely novel scientific findings rather than reproducing existing knowledge. Researchers and commentators are expected to press for details and independent verification of the results.

OpenAI says it has produced three discoveries generated by artificial intelligence in mathematics, computer science and physics. The announcement, reported by Anadolu Agency, highlights the company's claim that its AI systems can contribute original scientific findings rather than only assist researchers. Details of the discoveries, their validation and independent verification were not immediately provided, leaving experts and observers awaiting more information about the significance of the claimed breakthroughs.

OpenAI has brought in outside experts, including a Nobel laureate, to review proofs its AI systems claim to have produced in mathematics, computer science, and physics. The move signals the company is treating the results seriously enough to seek independent verification rather than relying on internal checks, as researchers weigh what verified AI discoveries would mean for scientific work.

OpenAI is reported to have produced new mathematical proofs with implications spanning mathematics, computing and quantum physics. The announcement, flagged by technology coverage, suggests the company's AI systems contributed to results in these foundational fields. Details of the proofs, their verification and peer review have not yet been widely reported, so experts' reactions remain limited for now.

SpaceX has obtained a $40 billion finance lease commitment for GPUs, according to Yahoo Finance. The deal suggests the rocket company is making a major push into large-scale computing power, likely to support data center or AI-related ambitions alongside its launch and satellite businesses. Few further details about the lenders, hardware providers or intended use of the chips were included in initial reporting.

Neocloud firm Argentum AI is in talks to secure $5 billion in financing for AI chips, according to sources cited in an exclusive Reuters report. The deal would fund large-scale purchases of processors used for AI computing capacity, a costly but fast-growing market as companies race to expand data-center infrastructure. Details of the lenders and terms have not been disclosed.

An artificial intelligence system has disproved a mathematical assumption that computer scientists had treated as true for roughly 30 years, according to New Scientist. The result suggests AI tools can now settle open conjectures that resisted human proof, with implications for how researchers use machine learning in theoretical work and how the discredited assumption shaped parts of the field's reasoning.

Broadcom has arranged $63 billion in financing from Wall Street to fund Anthropic's purchases of AI chips, according to a Quartz report. The deal would give the OpenAI rival a major funding line to secure the custom semiconductors it needs to train and run its AI models, underscoring how the AI boom is drawing unprecedented financing from banks into the chip supply chain.

Major Wall Street banks have committed $63 billion to finance Broadcom AI chips for Anthropic, one of the leading artificial intelligence companies. The deal underscores how investment banks are increasingly bankrolling the computing infrastructure behind the AI boom, extending massive credit to chipmakers and AI labs as demand for specialised hardware accelerates.

Meta is preparing a stock sale of nearly $100 billion, described as a turning point in how Big Tech finances its artificial intelligence ambitions. The move suggests major technology companies are increasingly turning to equity markets, rather than cash flow alone, to cover the enormous costs of AI infrastructure and computing power.

Broadcom has arranged a $63 billion financing deal to fund Anthropic's chip purchases, one of the largest semiconductor financing arrangements on record. The agreement ties the AI lab's computing expansion directly to Broadcom's custom chip business, underscoring how far suppliers are going to lock in major AI customers as demand for specialised accelerators keeps climbing.

A group of banks together with Blackstone have assembled roughly $63 billion in financing to fund Broadcom chips that will be deployed for Anthropic. The arrangement is one of the largest debt packages yet tied to AI computing hardware, reflecting how lenders are increasingly bankrolling the infrastructure behind frontier AI models.

The Economist argues that AI compute has become a strategic resource akin to oil, but that the market for buying and selling it is underdeveloped. Computing power is concentrated in a handful of cloud and chip providers, and prices are opaque, so the piece calls for new market mechanisms and trading infrastructure to allocate compute efficiently across the AI economy.

Agile Robots and NVIDIA have announced the Agile ONE SN, a humanoid robot designed for research purposes. The launch pairs Agile Robots' robotics hardware with NVIDIA's computing technology, targeting research institutions developing humanoid and embodied AI applications. The announcement is drawing attention as competition in the humanoid robotics sector intensifies among hardware makers and AI platform providers.

Amazon CEO Andy Jassy says the company's cloud computing capacity is fully sold out through 2027, underscoring overwhelming demand for AI-driven computing power. The remarks point to severe constraints in data centre availability, with customers securing capacity years in advance. Commenters are debating whether this signals genuine demand or a bottleneck Amazon cannot build fast enough to relieve.

XDS Datacentres is showcasing how it cools its chips at its facilities in Saudi Arabia. Bloomberg carried a look at the company's cooling methods, highlighting the infrastructure being built in the kingdom to support energy-intensive datacentre operations as demand for computing power grows across the region.