SK Hynix and Intel discuss the first US-made memory chips

Share
SK Hynix and Intel discuss the first US-made memory chips

Memory has been the tightest link in the AI supply chain all year. Now Seoul and Santa Clara are talking about moving part of it onshore.

SK Hynix is in talks with Intel to manufacture memory chips in the United States for the first time, a deal that would hand Intel's foundry business the marquee customer it has been chasing and put high-bandwidth memory on American soil. Under one scenario, SK Hynix would lease part of Intel's chipmaking facility in Ohio; the two companies could also form a joint venture with Intel and unnamed major cloud firms, and other structures remain on the table, according to the report. Intel and SK Hynix both declined to comment. Investors moved before either company did: Intel rose about 5% and SK Hynix's Nasdaq-listed shares about 3% in premarket trading.

Why it matters is the component, not the paperwork. HBM is what Nvidia's accelerators cannot ship without, and SK Hynix is the leading supplier of it. The data-center buildout has produced a memory shortage that lifted prices, and with them profits — SK Hynix's Korea-listed stock is up 400% over the past year, and it listed American depositary receipts on the Nasdaq in July. For Intel chief executive Lip-Bu Tan, whose turnaround rests on the foundry business, a memory tenant would validate capacity that has so far failed to attract large outside customers, and it would advance Washington's effort to bring semiconductor manufacturing home. We have tracked this squeeze from the demand side — AI demand sold out all 2027 DRAM and HBM capacity.

The complication is Seoul's, not Washington's. A deal covering advanced memory such as HBM could face opposition from the South Korean government, which treats the technology as sensitive, Reuters reported, citing three sources. The talks are exploratory and no decisions have been made — which is the correct way to read a Wednesday morning leak that sent both stocks up. What the report does establish is where the leverage sits: for now, the country that builds the memory sets the terms for the country that buys it.


Ant Group's health app Afu launched an "AI diet score" that grades a meal from a single photo, built on a nutrition rubric the company developed with the Chinese Nutrition Society's expert group. The score follows the society's Balanced Diet Index: the model identifies the dishes, estimates the balance of carbohydrate, protein, fat and fibre, and flags excess oil, salt and sugar as well as high-purine and high-glycemic foods, with the verdict adjusted for the user's body-mass index and activity level. A reporter's test on a two-meat-one-vegetable work lunch returned 81 points, with praise for protein and variety, a warning on sodium, and a suggestion to swap the drink for unsweetened tea.

The distribution plan is the more interesting design choice. Afu is pushing photo logging through its "lose 100 million jin scientifically" campaign by handing out 100 million eggs — three consecutive days of meal photos unlocks a 1-cent offer with free shipping. Ant says more than 2.75 million users have weighed in since June and collectively lost over 5 million jin, about 2.5 million kilograms, with users who also logged their diet losing roughly twice as much as those who did not. A national nutrition body's rubric plus a grocery incentive is a consumer-health play rather than a medical one, and it works only as long as people keep taking the photo.

What to watch: whether Intel and SK Hynix confirm the talks formally — and whether Seoul's technology-sensitivity review becomes the story instead.

If a memory shortage is what forces US onshoring, should Washington be paying for it rather than the buying companies? Tell us in the comments.

Read more

Akhetonics says its all-optical CPU reaches a customer in 2026

Akhetonics says its all-optical CPU reaches a customer in 2026

Light-based computing keeps promising more than it delivers — but one Munich startup has just put a date on its bet, and the interview laying it out is doing the rounds on Hacker News this week. Akhetonics says it will deploy its first commercial machine with a major customer by the end of 2026, with several more planned for 2027. The company, founded by Michael Kissner and Leonardo Del Bino, is building a computer where data enters as light, is switched as light, and circulates through memory

The Week in AI — October 5–11, 2026

The Week in AI — October 5–11, 2026

Every big claim this week turned out to rest on fine print more interesting than the headline: revenue only the company reporting it can define, safety tests sandboxed while the product keeps the web, and a Pentagon phase-out nobody would confirm until reporters kept asking. The week's top 5 1. OpenAI's revenue was $20 billion below the numbers everyone quoted — and the gap was definitional. The Financial Times reported Thursday that OpenAI's annualized revenue runs roughly $20 billion unde

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Russia's largest tech company is learning what the AI era's infrastructure war looks like from the receiving end — three data centers in four days, and with them much of the cloud layer Russian businesses run on. A Ukrainian drone strike knocked out Yandex's data center in Vladimir early Sunday morning, the third of the company's facilities hit since October 8. The site — reported at roughly 50 MW and designed for about 2,880 server racks — stopped operating completely after the attack, Yandex

Agent teams cost up to 5x more, barely score higher

Agent teams cost up to 5x more, barely score higher

The multi-agent hype train hit a benchmark this weekend — and the grid and the trucking regulators had quiet weeks of their own. Vals AI put agent teams head-to-head with single agents on its Vibe Code Bench, and the teams cost between 1.8 and 5.1 times more for almost no extra quality. The evals company ran GPT-6 Sol and Claude Opus 5.5 solo and in teams across 50 apps at two reasoning efforts; out of four comparisons, only one was statistically significant — Sol at medium effort, where the t