cloud computing

Amazon is weighing a plan to sell Nvidia chips deployed in its data centers and then rent that computing capacity back from buyers, according to a report picking up traction in business and technology circles. The arrangement would let Amazon free up capital while keeping AI computing power for its cloud customers. Commentators see it as part of the industry's broader scramble to finance expensive AI infrastructure.

A proposal circulating among tech commentators argues that the answer to growing data centre sprawl could be small, modular facilities built underground. The idea, framed around small modular data centres and edge computing, ties into debates over digital sovereignty and where cloud infrastructure should physically sit as demand for computing capacity keeps rising.

A new essay argues it is time for 'Personal Computing 2.0', a revolution in how individuals own and control their computing. The piece contends modern software has drifted from empowering users toward cloud services and platforms that hold the data and the leverage, and calls for tools that put individuals back in charge of their own machines and workflows.

South African technology coverage is discussing the end of a Cloud-only approach in the country, suggesting a shift toward hybrid or on-premises computing arrangements. The report, published on a South African gadget news site, frames the change as significant for local businesses and IT strategy. Details on which providers or policies are affected were not included in the available coverage.

An open-source tool has been highlighted for allowing users to run artificial intelligence models directly on their own machines, without relying on cloud services. Coverage in the open-source software community points to growing interest in local AI for privacy, cost savings, and independence from major providers. Enthusiasts say the approach puts control of data and computing back in users' hands.

Oracle has agreed to a five-year, $7 billion lease arrangement covering AI chips, a major commitment to the computing infrastructure behind artificial intelligence. The same recap notes that Tencent's Yuanbao AI assistant is having some features limited. The news feeds ongoing attention to how cloud providers are locking in scarce AI hardware capacity.

Amazon is reportedly pursuing an $8 billion deal to offload Nvidia Blackwell chips. The move, if confirmed, would involve transferring a large batch of Nvidia's latest AI processors, possibly to other companies seeking computing capacity. It would be a notable reshuffling of AI hardware among major cloud providers, though details on buyers and terms remain sparse.

An economist at New Mexico State University argues that investment in data centers is propping up the US economy enough to prevent a recession. The claim points to the massive construction and infrastructure spending driven by artificial intelligence and cloud computing demand as a key source of economic growth. The remarks highlight how central the tech building boom has become to the broader American economy.

A statewide moratorium on data center construction is drawing attention to the environmental footprint of the industry, including high water and electricity use. Supporters of the pause argue growth must be slowed until impacts are assessed, while data center development is surging across many regions to meet demand for cloud computing and artificial intelligence.

Amazon shares edged higher in premarket trading following reports that the company is looking to sell $8 billion worth of Nvidia chips to investors. The reported move suggests Amazon may be looking to restructure parts of its AI hardware commitments or recover costs on surplus inventory, drawing attention across the semiconductor and cloud computing sectors.

Programmers are setting up small local servers, jokingly called 'AI sheds', to run AI coding tools on their own hardware instead of renting expensive cloud capacity. The trend follows a suggestion by Ruby creator David Heinemeier Hansson, who argued developers could save money and keep code private by hosting models at home. Responses have been mixed, with some praising the cost savings and others calling it impractical.

Google has reportedly placed its first orbital data center into orbit, marking a step toward computing infrastructure in space. The development was highlighted by computer scientist Daniel Lemire on his blog, filed under space and technology. Orbital data centers could offer solar-powered, always-illuminated computing, though questions remain about maintenance, cooling and radiation resilience.

Google's first Chromebook, the Cr-48, was never sold in stores and did not even carry the Chromebook name. The unbranded, plain black laptop was sent to testers as a pilot device ahead of the 2011 Chromebook launch, serving as a proof of concept for Google's browser-based computing vision. Tech commentators are revisiting the device and its role in shaping Google's later cloud-first laptop lineup.

Researchers have built a quasilinear verifier for quantum computations, a tool designed to check that quantum devices produce correct results without requiring exponentially large classical resources. Verification is one of the central unsolved problems in quantum computing, since users cannot directly confirm outputs of machines they cannot simulate. The development is being reported as a step toward trustworthy quantum cloud services and practical certification of quantum hardware.