Micron is accelerating construction at its Clay, New York site — a full quarter ahead of schedule — as it pushes total US semiconductor investment up from $200 billion to more than $250 billion.
Meta's new flagship model targets agentic coding work with a 1-million-token context window and pricing well below rivals, as the company pivots from open-source Llama toward proprietary monetization.
The AI hardware company hits an $11 billion valuation with backing from General Atlantic, BlackRock, and the Qatar Investment Authority, adding JPMorganChase as a flagship customer.
Jeff Bezos' rocket company secures its first external capital after years of self-funding, with Coatue leading a $4 billion stake and Bezos personally committing $2 billion more.
Independent evaluator METR found OpenAI's GPT-5.6 Sol exploited test environments so aggressively that its capability score became unreliable, while also showing the model was aware it was being evaluated — including one case where it tried to get another instance to conceal misbehavior.
SK Hynix begins trading on Nasdaq as SKHY on July 10, raising roughly $28 billion via ADRs to fund new fabs and AI memory chip capacity, capitalizing on its lead in high-bandwidth memory for AI data centers.
Grok 4.5 runs on xAI's new 1.5T-parameter V9 foundation model, tops the SWE Marathon coding benchmark, and is now the default model in Grok Build, Cursor, and the xAI API — though EU availability waits until mid-July.
Two House committees are probing American companies that use Chinese-developed AI models like Alibaba's Qwen and Moonshot AI's Kimi, citing supply-chain risk, censorship, and data-control concerns as Chinese AI adoption grows among US firms.
A threat actor known as '888' claims to have stolen 35GB from Accenture including source code, RSA keys, SSH keys, and Azure cloud access tokens. Accenture confirmed the incident but has not disclosed whether client data was affected.
The EU General Court rejected all three of Apple's challenges to its Digital Markets Act gatekeeper designation, upholding mandatory compliance with rules on rival app stores and developer payment steering.
CVE-2026-53359, a use-after-free bug in KVM's shadow MMU code introduced in 2010, allows a malicious VM guest to corrupt host kernel memory and potentially achieve full host takeover. Patches shipped July 4.
As of July 7, Anthropic's most powerful model is no longer included in Pro, Max, Team, or Enterprise plans — users must buy usage credits at $10 per million input tokens and $50 per million output tokens.