Apple vs OpenAI, Anthropic Leads on Revenue, Meta Goes Custom Silicon: AI News

Sean Ebeling
Jul 14
4 min read

TL;DR

Apple filed a federal lawsuit against OpenAI alleging coordinated trade secret theft tied to hardware development, Anthropic surpassed OpenAI in annualized revenue at roughly 47 billion dollars driven by enterprise adoption of Claude, and Meta confirmed its custom Iris AI chip enters production in September as it targets 14 gigawatts of compute by 2027. These three moves together show that the AI race has shifted from model benchmarks to hardware ownership, legal IP control, and enterprise contract lock-in.

AI competition is now being fought on legal, financial, and silicon battlegrounds, not just model leaderboards. Enterprise revenue has become a more reliable signal of market position than user counts, and every major lab is now vertically integrating into hardware to reduce GPU dependency and cut inference costs. Meanwhile, Googles flagship Gemini 3.5 Pro remains days away from a rumored July 17 launch, keeping competitive pressure high across the entire sector.

Apple Sues OpenAI Over Alleged Trade Secret Theft Tied to Hardware Push

Apple filed a federal lawsuit in the Northern District of California on July 10 2026, accusing OpenAI of orchestrating a coordinated campaign to extract confidential hardware designs, component specs, and vendor data through the recruitment of more than 400 former Apple employees. The suit specifically names OpenAI Chief Hardware Officer Tang Tan, a 24-year Apple veteran, for allegedly coaching departing employees to evade Apples exit security checks and bring physical hardware components to OpenAI interviews. The lawsuit matters because it puts OpenAIs first hardware product directly in legal jeopardy and sours what had been a high-profile partnership since ChatGPT was integrated into Apple Intelligence in 2024.

Anthropic Overtakes OpenAI in Annualized Revenue at 47 Billion Dollar Run Rate

Anthropic hit a 47 billion dollar annualized revenue run rate by May 2026, exceeding OpenAIs self-reported figure of 25 to 33 billion dollars, driven almost entirely by enterprise API contracts rather than consumer subscriptions. The engine behind the surge is Claude Code, which grew from a 1 billion dollar annualized run rate to over 2.5 billion dollars in roughly two months earlier this year, and Anthropic reported operational profitability in Q4 2025 while OpenAI has not yet provided a profitability timeline. The revenue reversal signals that enterprise contract stickiness and developer tooling are now outpacing the consumer subscription model that made ChatGPT a household name.

Meta Iris Chip Enters Production in September as Infrastructure Spend Hits 145 Billion Dollars

An internal Meta memo reviewed by Reuters confirmed that its custom AI chip code-named Iris will enter mass production in September 2026 after clearing six weeks of testing with no major issues. The chip is designed with Broadcom and manufactured by TSMC, and it is intended to handle recommendation ranking and inference workloads across Facebook and Instagram while reducing reliance on Nvidia and AMD GPUs. Meta expects to spend up to 145 billion dollars on AI infrastructure this year and is targeting a total compute capacity of 14 gigawatts by 2027, doubling its current footprint.

Gemini 3.5 Pro Targets July 17 Launch With 2 Million Token Context Window

Leaked plans and multiple reports point to a July 17 2026 release date for Google DeepMinds Gemini 3.5 Pro, a model that was originally due in June before Google scrapped the existing 2.5 Pro architecture for a full rebuild. The rebuilt model is expected to ship a 2 million token context window, a Deep Think Reasoning Layer for complex multi-step problem solving, and autonomous workflow capabilities for coding and tool management, though Google has not officially confirmed the date or final specs. The delay reflects both internal performance concerns and the intensifying competition from OpenAIs GPT-5.6 and Anthropics Fable 5, both of which are also targeting mid-July availability.

UN Scientific Panel Warns Safeguards Are Falling Behind AI Adoption Globally

The Independent International Scientific Panel on Artificial Intelligence, a 40-expert body established by the UN, published its first preliminary report and fed its findings into the UN Global Dialogue on AI Governance held in Geneva on July 6 and 7 2026. Panel members including Yoshua Bengio noted that AI is approaching or surpassing human capabilities in many domains while outpacing both scientific understanding and governments ability to adapt, and participants flagged that frontier AI development is concentrated in just two countries, the US and China, raising concern about an AI divide that leaves developing nations behind. The panel called for universally accepted global guardrails, framing inaction as a rising cost rather than a neutral default.

The pattern across this weeks stories is consolidation: of legal IP, of hardware supply chains, of enterprise revenue, and of regulatory pressure. Companies that win the next phase will not simply be those with the best models but those that control the hardware, the enterprise contracts, and the governance frameworks around their deployments. The Apple versus OpenAI lawsuit and the Anthropic revenue lead are the two data points most worth watching, because together they define where the economic and legal gravity in AI is shifting.