This week’s stories are about tech, AI and mobile, coming from Korea and the USA

California watchdog pushes rules to ensure data centers cover power costs

A California government oversight panel is urging lawmakers to adopt new regulations for the rapidly growing data center industry, warning that rising electricity demand from artificial intelligence (AI) could otherwise drive up power bills for households.

I saw this story at The Associated Press and the panel proposed more than a dozen recommendations aimed at managing the industry’s impact on the power grid, electricity prices and the state’s climate goals. The report comes as lawmakers in Sacramento prepare new proposals to regulate the expanding industry after similar efforts last year stalled amid opposition from technology companies and business groups.

A key concern is the scale of future electricity demand. Utility company Pacific Gas & Electric told regulators that data center projects seeking power could add roughly 10 gigawatts of demand over the next decade — about four times the generating capacity of the Diablo Canyon nuclear power plant, according to the story. By comparison, the Sacramento region consumes slightly more than 3 gigawatts of electricity during peak periods.

The report recommends requiring confidential reporting of facility-level electricity use by data centers to help regulators forecast demand, determine where grid capacity exists and evaluate potential reliability and environmental effects.

The commission recommends that large data centers pay the full cost of the infrastructure and grid services they require. It also proposes creating a special electricity rate category for extremely large power users that would require advance payments for grid upgrades, contributions to wildfire safety programs and commitments to pay for a share of the power capacity they request, the story noted.

California watchdog pushes rules to ensure data centers cover power costs

US weighs new rules linking AI chip exports to foreign investment

US officials are considering a new regulatory framework for exporting artificial intelligence (AI) chips that could require foreign governments or companies to invest in U.S. data centers or provide security guarantees in order to receive large shipments of the technology.

The proposal, which is still under discussion and may change, would mark the first major attempt by the administration of Donald Trump to regulate shipments of advanced AI chips to allied nations after it scrapped the so-called AI diffusion rules introduced under former President Joe Biden.

I saw this story at Reuters and under the draft plan, countries seeking to import 200,000 or more AI chips could be required to commit to investments in US AI data centers or provide other government-backed security assurances.

The measures would not change existing restrictions on blacklisted countries such as Russia. China, which had been subject to tight export controls, was cleared in December to receive second-tier AI chips from Nvidia, though deliveries have been delayed by national security reviews, the story noted.(By the way I have a story here about AI)

According to the document, even relatively small deployments of fewer than 1,000 chips could require export licenses. To qualify for exemptions, chipmakers such as Nvidia and AMD would be required to monitor the hardware, while customers would need to use software preventing the chips from being linked together into large computing clusters.

Foreign companies seeking up to 100,000 chips would need to provide government-to-government assurances, the document revealed. The Trump administration has already imposed such conditions on Saudi Arabia when approving purchases of advanced chips.

Installations involving up to 200,000 chips could also require on-site inspections by U.S. export control officials.

In a statement posted on the social platform X, the US Department of Commerce confirmed that it is evaluating new export controls.

US weighs new rules linking AI chip exports to foreign investment

Samsung accelerates Exynos 2700 rollout to reduce Qualcomm reliance

Samsung prepares production-ready samples of its next flagship smartphone processor, the Exynos 2700, by midyear as the company aims to expand use of its own chips in Galaxy devices and reduce dependency on Qualcomm.

I saw this story at Korea Herald and the Exynos 2700 is expected to power a substantial portion of the Galaxy S27 lineup, slated for release early next year. Mobile application processors like Exynos act as the central computing unit in smartphones and typically account for 15% to% 30 of flagship device costs. Samsung’s System LSI division designs the chips, while the company’s foundry unit handles manufacturing before supplying them to the Mobile eXperience business, which assembles Galaxy devices.

The accelerated timeline follows challenges in Samsung’s in-house chip program. The Exynos 2300 never reached mass production due to underperformance, and delays with the Exynos 2500 led Samsung to equip its 2025 Galaxy S25 and Z Fold 7 models exclusively with Qualcomm processors, according to the story.

Samsung returned to the flagship market this year with the Exynos 2600 in the Galaxy S26 and S26 Plus, while the S26 Ultra continues to rely on Qualcomm’s Snapdragon 8 Elite Gen 5 worldwide, per the story. Benchmarks suggest the two chips deliver broadly comparable performance, with Qualcomm leading in single-core processing and Exynos remaining competitive in multicore, graphics, and AI-related tasks.

Exynos 2700 will be built on Samsung’s second-generation 2-nanometer process (SF2P) and will feature refined packaging to improve heat management, building on techniques introduced in the Exynos 2600.

Analyst Park Yu-ak of Kiwoom Securities estimates Exynos 2700 could account for 50% of Galaxy S27 chips, up from 25% in the S26 lineup. Industry data indicate Samsung spent about $7.5 billion on external mobile processors in the first nine months of 2025, including $400 million for Qualcomm chips for the Galaxy S25 series.

Samsung accelerates Exynos 2700 rollout to reduce Qualcomm reliance