US data center

A Wall Street Journal analysis examines projections that US investment in artificial intelligence could reach roughly 9% of GDP, a level of spending comparable to historic infrastructure booms. The piece notes that the technology industry is counting on this wave of capital continuing, raising questions about whether such unprecedented outlays on data centers and computing power are sustainable or realistic.

Tencent is reportedly leasing 100,000 Nvidia H100 AI chips from Oracle, housing them in Southeast Asian data centers including Malaysia. Commenters describe it as a strategic and costly move by a Chinese tech giant to build offshore computing capacity that sidesteps US export controls on advanced chips to China.

Trading community TopGTraders is hosting a live forex trading session on October 2 centered on the US Non-Farm Payrolls release, with analysis of gold and Bitcoin markets. The session marks day 513 of the group's ongoing live trading series, drawing strong engagement as traders watch how the jobs data moves currency pairs and crypto prices.

Aleph Alpha has launched Kolibri, a new open-weight language model positioned as a sovereign European alternative to models from US and Chinese providers. The release emphasizes data sovereignty and transparency, letting organizations inspect and run the weights themselves. Discussion online centers on whether a European open-weight model can credibly compete with dominant frontier models.

GlobalFoundries has received an additional $300 million award under the CHIPS Act from the Trump administration, strengthening US chip manufacturing support. Separately, Nvidia CEO Jensen Huang said silicon photonics is 'essential' to the future of AI, highlighting growing demand for optical interconnect technology in data centers. The moves underline Washington's continued push to onshore semiconductor production while Nvidia frames photonics as key infrastructure for scaling AI systems.

A new report from Wired says rural data centers are in line for a significant federal tax break. The story is being shared and discussed as part of the ongoing debate over whether US policy should keep subsidizing the infrastructure behind the AI boom, and at what cost to taxpayers and local communities.

A news roundup flags three developing stories: natural gas prices, a slowdown in hiring, and negotiations around data center projects. US natural gas prices remain a concern for consumers and industry, while a cooling labor market is raising economic questions. Meanwhile, governments and tech companies are increasingly treating data center siting as a matter of diplomatic and economic strategy.

The US National Cancer Institute's Center for Global Health is examining whether gaps in Caribbean cancer registry data are limiting cancer control research in the region. Incomplete registries make it harder to track cancer incidence, plan prevention programmes and evaluate treatment outcomes across Caribbean nations. The discussion highlights concerns about health data infrastructure in small island states.

LinkedIn cofounder Reid Hoffman said investment in AI infrastructure is the 'only reason we're not in a recession,' arguing that massive spending on data centers and computing capacity is propping up the US economy. His comments add fuel to a growing debate about whether AI-driven capital expenditure is masking broader economic weakness.

Amazon plans to move about US$8 billion of semiconductor chips off its own balance sheet, according to a Financial Times report cited by the Taipei Times. The arrangement would shift ownership of the chips, likely used in its cloud and AI data centers, to suppliers or partners, reducing Amazon's capital costs. The move highlights how intensely tech firms are managing spending on the hardware powering the AI boom.

Google has announced Project Suncatcher, an initiative aimed at eventually running AI data centers in space, using solar-powered satellites to meet the growing energy demands of artificial intelligence. The announcement is being reported widely across US public radio outlets. Details on timelines, costs and feasibility remain limited, but the project signals how major tech firms are exploring unconventional ways to power AI infrastructure as terrestrial energy needs surge.

Donald Trump has pushed back against suggestions that he is minimizing the dangers of artificial intelligence, while claiming there is a conspiracy working against data centers. The remarks add to an ongoing debate in the United States over AI safety regulation and the rapid expansion of data center infrastructure, which has drawn resistance from some communities and politicians.

Georgia voters headed to the polls to pick a new state representative in a race shaped by two major local and national issues. Candidates debated their positions on the US conflict with Iran and on a proposed moratorium on data center construction, reflecting concerns about foreign policy and the strain large data facilities place on local power and land resources.

Sherrod Brown, campaigning in Toledo, criticized Jon Husted over data center development and economic policy, framing the issue as part of his broader economic pitch to Ohio voters. The two are rivals in Ohio's US Senate race, and Brown used the stop to draw a contrast on jobs, energy demands and who benefits from data center projects. Coverage centered on his attacks and the campaign's economic messaging.

Security firm Trail of Bits has published SequenceHash, a tool it describes as multihashing "for the rest of us." The announcement introduces an approach to computing multiple hash functions over the same data more easily, aiming to make a technique common in cryptography and security engineering accessible to ordinary developers. Discussion among engineers is centered on the tool's practical uses and design.

A US Treasury official said the massive buildout of technology, including data centers and related infrastructure, is creating cost pressure, according to Fox Business. The comment adds to a growing debate in Washington and on Wall Street about how spending on artificial intelligence and computing capacity is straining budgets, supply chains and borrowing costs across the economy.

America is racing to build new power plants to meet the surging electricity demands of artificial intelligence data centers. Commentators argue there is another way to meet this demand, pointing to alternatives to large-scale plant construction, though the specifics of that alternative approach are not detailed in the headline.

US construction spending in August was driven largely by home improvement projects and data center building. The figures point to continued strength in residential renovation and a boom in technology infrastructure, even as other parts of the construction sector have cooled. Analysts view data center demand as a key engine of overall construction activity this year.

A startup is seeking approval to place more than 450 small nuclear reactors on public land in Utah, with the sole purpose of powering a large data center complex. The proposal stands out for its sheer scale, far beyond any existing deployment, and would tie a massive computing facility directly to new nuclear generation in the US West.