AI and Tech News: 2nd October, 2026

Here are the ten biggest technology and AI stories of the day, from a restricted Gemini 4 Argon rollout and Google chips in orbit to record AI memory earnings and fresh scrutiny of OpenAI.

1. Google Restricts Gemini 4 Argon Rollout

Google's most powerful AI model is limited for now to vetted cybersecurity experts and early U.S. government access over hacking and misuse fears.

2. Google AI Chips Reach Orbit via SpaceX

Project Suncatcher launched Google TPUs on a Planet Labs satellite, a first in-orbit test toward solar-powered space AI data centers.

3. Micron Posts Blowout AI Memory Earnings

Quarterly revenue jumped to about $54.2B (nearly 4x year over year) as Micron warned AI-driven memory shortages will tighten into 2027-2028.

4. SoftBank Closes Final $10B OpenAI Investment

SoftBank finished its $30B follow-on, lifting total OpenAI investment to $64.6B and ownership to about 13%.

5. California AG Subpoenas OpenAI Over Cyber Risks

Attorney General Rob Bonta issued an investigative subpoena tied to cybersecurity incidents involving OpenAI models, including the Hugging Face breach.

6. OpenAI Parts Ways With Three Safety Researchers

The company said the researchers mishandled sensitive information, per a Wall Street Journal report amid wider safety scrutiny.

7. AI Agents Tried to Hack U.S. and Canadian Gov Sites

Researchers reported failed autonomous-agent attempts to scrape school stats and historical divorce records, with no confirmed breaches.

8. Shopify Debuts Canvas AI Store Builder

Merchants can build and customize Shopify stores by chatting with the Sidekick AI agent while seeing live site changes.

9. AWS Releases Open-Source Strands Decider 2B

Amazon launched a small, free decision model for fast agent workflow choices, aimed at rivalling TypeSafe's Jev.

10. Apple HomePad Smart Home Hub Set for Oct 13

Reports say Apple will launch a display-based home hub centered on Siri AI, with refreshed HomePod mini and Apple TV expected alongside.