A two-week training pause has ended only in part: OpenAI says its biggest planned run remains blocked by stricter security and monitoring gates.
Companies
AI companies
Strategy, platforms, partnerships, pricing, competition, and business-model shifts.
19 sourced postsThe company is linking machine-led triage to bounded responses while keeping people responsible for its highest-impact security decisions.
The August 11 snapshot points to heavy voice, visual and image-generation use, but provides no methodology or independent audit for the figures.
The company is building a zero-day close and continuous forecasting around approved data, human validation and source-linked outputs, but reports no achieved target.
The publisher says its current and archived reporting will become available across Meta apps and devices, but the announcement leaves attribution and rollout details open.
The services group is embedding Claude into engineering and operations systems while scaling a trained delivery workforce for client deployments.
The planned program combines AMD compute and software with Korean chips, research, talent programs, and a proposed hub for locally developed AI assets.
Announced June 25, GA July 2026: Agentforce Help Agent bills $2 per autonomous resolution; no charge on human escalation or negative feedback; resolution blocks sold from 1,000 units; Data 360 and Agentforce unmetered during interactions. S
DeepSeek emailed API subscribers June 29–30 announcing an "official V4 version" for mid-July plus the first time-based pricing from a major LLM API: 2x peak rates during Beijing business hours (9–12, 14–18). deepseek-chat / deepseek-reasone
Bloomberg (Julia Love, Davey Alba), citing 10 current/former employees, reported July 16 that Gemini 3.5 Pro — promised by Pichai at I/O in May for June — is months behind internal targets on coding performance; a late-June training-data up
July 15: China's CAC registered Apple Intelligence for launch in mainland China; Alibaba confirmed Qwen will be integrated into Apple Intelligence across iOS/iPadOS/macOS/visionOS (text + image understanding/generation); Baidu confirmed sup
A July re-verification confirms OpenAI's May-launched educator plan is still no-cost with a FERPA-aligned workspace — a settled landscape item, not a new launch.
On July 13, 2026, Google, Microsoft, Salesforce, Snowflake, and ServiceNow agreed to back a shared agent-connectivity standard built on Google's Agent-to-Agent (A2A) protocol — explicitly positioned as an alternative to Anthropic's MCP, the
Reuters interview (~July 12) with CEO K Krithivasan: TCS will convert 1–1.5% of associates (~5,900–8,900 people) into forward-deployed AI engineers and is evaluating acquisitions in AI, data security, and cybersecurity — a break from its or
July 10: Apple filed a 41-page complaint in N.D. Cal.
Anthropic shares trade on secondary markets at an implied ~$1.2T (+550% YoY), above OpenAI's ~$908B secondary mark; Anthropic confidentially filed for an IPO in June after a $65B raise at $965B (May). Fortune reports ~$47B annualized revenu
July 9: Fidji Simo announced she is leaving her full-time role (CEO of AGI deployment / Applications; Altman's de facto No. 2) to become a part-time advisor, after three months of medical leave for a severe exacerbation of a chronic neuroim
On July 6, 2026, xAI rebranded as SpaceXAI (X handle @xai → @SpaceXAI), formalizing the Feb 2, 2026 all-stock SpaceX–xAI merger (~$1.25T combined) after SpaceX's record ~$75B June IPO (SPCX, $135/share, +19% day one). Grok products, SuperGr
Under a January 2026 Master Relationship Agreement (> $20B, 750MW of Cerebras wafer-scale capacity 2026–2028, expandable to 2GW by 2030), OpenAI serves GPT-5.6 Sol on Cerebras at up to 750 tokens/second (vs ~40–120 on GPU clusters; Cerebras