Skip to content
Day 69: new resources are added every day. See what's new

Daily news roundup

AI and Tech News, 17 Oct 2026: Chip Rout, SoftBank, China AI Curbs

Twenty-three AI and technology headlines: a chip stock selloff, SoftBank eyeing Allied Universal, China's AI export curbs, Starshield limits, iPhone Duo, and more.

  • 5 min read
  • #News
  • #Technology
  • #AI
  • #Markets
  • #Policy
  • #Apple

Rough day for chip stocks. Here are the twenty-three biggest tech and AI stories, from a sharp chip selloff and SoftBank's reported interest in Allied Universal to China's proposed AI export controls and Apple's sold-out iPhone Duo.

1. Chip Stocks Slide 3.8% and the Nasdaq Drops 2.1% on AI Debt Worries

The Philadelphia Semiconductor Index fell 3.83% on Friday, with all 30 members lower. The Nasdaq closed down 2.14%, and CoreWeave lost about 12%. The worry on investors' minds: debt-funded AI spending.

2. Bloomberg: SoftBank Is Weighing an Allied Universal Deal Worth Up to $18B

Preliminary talks could value the equity of the largest US security services company at up to $18B. Masayoshi Son is exploring ways to put AI-powered robots to work across industries. No deal is final.

3. China Floats Export Controls on 23 AI Technologies

China's commerce and science ministries opened public comment on adding 40 technologies to the restricted export catalogue, and 23 of them are in AI. Open-source releases are exempt, and API access is still allowed.

4. Taiwan Wants to Make Shipping Restricted AI Chips to China a Crime

Economy Minister Kung Ming-hsin told Bloomberg that Taiwan will add AI servers and advanced chips to its controlled items list as part of a US trade deal. No timeline was given, though.

5. Who Ordered Limits on Ukraine's Starshield Service? The Reports Don't Agree

The stories don't line up. The Telegraph reported that Elon Musk proposed slowing the military satellite service and Trump agreed. Reuters said the order came from inside the White House. And Trump denied making the decision.

6. Politico: Anthropic Pitches Congress Its Own Frontier AI Safety Bill

Under the draft, the government could block a model only for a foreseeable, imminent risk of over $1B in damages or more than 50 deaths or serious injuries. It would have to go through the Commerce Department and a court.

7. Apple's iPhone Duo Sells Out, and US Deliveries Slip to Late January

Within minutes of preorders opening, estimated US ship dates for its first foldable iPhone moved from the October 23 launch into late January 2027.

8. Replit's CEO Says It's Done Working With Anthropic, but a Source Disputes That

Amjad Masad cited cost, partnership terms and performance, and called OpenAI the better partner. But a person familiar with the relationship told TechCrunch the Anthropic partnership is still in place.

9. Reflection AI Puts Beam's Open Weights Out Under Apache 2.0

The 501B-parameter mixture-of-experts model has 23B active parameters. It's now on Hugging Face in BF16, FP8 and NVFP4 versions, along with a tokenizer and reference implementation.

10. Meta Trims $25M From Its $65M Midterm Political Budget

Filings reviewed by the New York Times show Meta got $25M back in refunds, mostly from the super PAC it set up to fight state AI bills. Meta described the refunds as spending adjustments.

11. AI-Generated Campaign Ads Pass $137M in the 2026 Midterms

Wesleyan Media Project and AdImpact data, reported by the New York Times, found 220 AI-generated ads across 34 states. Most were tied to Republicans, and most carried no AI disclosure.

12. Anthropic Says Claude Traced NIST's Mysterious Curve Seeds to 1990s Timestamps

Anthropic said Claude linked the seeds of NIST's elliptic curves to a random number generator seeded with timestamps from the late 1990s. That undercuts long-held suspicions of a hidden backdoor.

13. Waymo Plans Toronto Robotaxi Service by 2028, Its First Canadian City

Element Fleet Management will run fleet operations. Waymo has started mapping and testing Toronto streets with trained specialists, but driverless rides still depend on regulatory approval.

14. Taiwan Plans US Industrial Parks for AI Suppliers, Beginning With a $4B Texas Site

Each park would cost $2B to $4B and host smaller suppliers to TSMC and Foxconn. An industry group would raise the money, and Taiwan's government would offer loan guarantees.

15. Blue Origin Sets Out a $330M Florida Expansion for New Glenn

The Cape Canaveral upgrades are due in summer 2027. They're part of a push toward multiple New Glenn launches per week by the end of the decade, alongside plans for two more pads.

16. Apple Brings In Meta's Scott Tucker as Its Global Litigation Chief

Tucker spent nearly six years as Meta's deputy general counsel for global litigation. He takes a new vice president role as Apple faces antitrust, App Store and privacy lawsuits, Bloomberg reported.

17. Apple Cracks Down on AI Nudify and Bait-and-Switch Apps in the App Store

Apple's updated review guidelines add nonconsensual intimate imagery to its content rules. They also ban apps that hide harmful features, such as nudification, from reviewers until after approval.

18. 404 Media: ChatGPT Ads Can Be Targeted Using Emails, Phone Numbers and Prompts

Test campaigns showed advertisers can upload contact lists and add topic hints that get matched to conversations. OpenAI requires matched audiences of at least 25,000 users and hides who matched.

19. GPT-6 Pro Can Now Run Multiple Agents in Parallel in ChatGPT

Pro subscribers can now have several agents work on different parts of a task at once and follow each one's progress. Responses are faster too, and tool use is better.

20. Experts Say Meta's New Business Terms Could Allow News-Site Scraping, and Meta Denies It

According to Nieman Lab, Meta's new commercial terms, first set for October 30, let it crawl websites tied to business accounts, which legal experts say could cover AI training. Meta has since pushed them to November 20 and softened the wording. A Meta spokesperson called the news-scraping claims false.

21. The FT Says KBRA Gives Anthropic a BB+ Credit Rating

KBRA assigned the rating, one notch below investment grade, in June. That was ahead of Apollo's $35B chip financing, where about $4.4B of junior debt tied to Anthropic sold at about 8.5%.

22. Indian Family Offices Lift AI Startup Bets About 50% to $760M This Year

Tracxn data cited by Bloomberg shows family offices and angel investors put about $760M into AI companies in 2026. Premji Invest and Ranjan Pai's Claypond Capital led the way.

23. Jensen Huang: The AI Market Turmoil Will Pass

Speaking in Berlin as chip stocks slid, Nvidia's CEO told Bloomberg that AI is transformative and will benefit every industry and country. In his view, that outweighs investors' current worries.

Know someone who'd want today's headlines? Send it their way.

About this roundup: each story is a short summary in our own words of reporting by the outlets named in it, or of the company's own announcement. Stories marked as reported by another outlet, or as coming from unnamed sources, have not been confirmed by us, and people or companies accused of wrongdoing have not been found guilty unless a story says so. Check the original report before relying on any detail. This is general news, not financial, investment or legal advice. Spotted a mistake? Tell us and we'll fix it.

Weekly newsletter

Everything new, every Monday.

One short email every Monday with everything shipped that week: new free tools and websites, new guides, new free browser tools and what's new in my free Gumroad libraries, plus how the site is growing. Free, and you can unsubscribe anytime.

Join the newsletter

You sign up on Gumroad, which stores your email to send the newsletter. See our Privacy Policy.