AI Daily: OpenAI scales back, reworks $500B ‘Stargate’ Project

Catch up on the top artificial intelligence news and commentary by Wall Street analysts on publicly traded companies in the space with this daily recap compiled by The Fly.

Claim 55% Off TipRanks

STARGATE INITIATIVE: OpenAI is scaling back and reworking its $500B Stargate data center initiative, halting or pausing several planned projects while shifting toward a more flexible model that relies on leasing compute capacity from third-party providers, The Financial Times reports. The change reflects a strategic pivot away from large-scale owned infrastructure toward faster access to computing resources, even as competition for AI infrastructure and talent continues to intensify. Stargate was initially announced in early 2025 by Donald Trump as a joint venture between OpenAI, Oracle (ORCL), Abu Dhabi fund MGX and SoftBank (SFTBY).

PHOTO-EDITING OVERHAUL: Apple (AAPL) is working on a major overhaul of the built-in photo-editing features for its iPhone, iPad, and Mac products, leaning heavily on AI to rival Android (GOOGL) devices, Bloomberg’s Mark Gurman reports. The tech giant is planning a new suite of tools powered by its Apple Intelligence platform for iOS 27, iPadOS 27, and macOS 27, all set for release this fall, the author says, citing people familiar with the matter. Shares of Apple are up about 1% in afternoon trading.

AI AGENTS: Amazon (AMZN) has introduced new AI-driven hiring software designed to streamline seasonal recruiting by reducing reliance on traditional face-to-face interviews, alongside a new design philosophy called “humorphism” aimed at making AI systems adapt more naturally to human workflows, Reuters’ Greg Bensinger reports. The announcements were made at an AWS event featuring leadership from Amazon Web Services and appearances by industry figures including OpenAI executives.

CLAUDE: Goldman Sachs (GS) has restricted its Hong Kong bankers from using Anthropic’s Claude models, reflecting growing sensitivity around U.S.-China tensions in AI deployment, The Financial Times’ Arjun Neil Alim, Zijing Wu, and Tim Bradshaw report. The move reportedly follows a strict interpretation of the firm’s contract with Anthropic and highlights how access to Western AI tools in Hong Kong can be shaped by corporate policy rather than local regulation, the report notes. A spokesperson for Anthropic said its Claude models had never been officially “supported” in Hong Kong but declined to comment further.

ISAAC SIM, BLACKWELL: Arrive AI (ARAI) is accelerating AI and robotics development using Nvidia (NVDA) Isaac Sim and high-performance GPU workstations powered by Nvidia Blackwell architecture. The company is leveraging simulation-driven AI training to rapidly improve computer vision systems used in real-world automation, robotics, and autonomous delivery environments.