AI IPO Race Heats Up as Apple and Anthropic Unveil Major AI Products

This week marks an AI “super week” with simultaneous capital and technology breakthroughs. OpenAI and Anthropic entered an IPO sprint, each submitting confidential S‑1 filings to the SEC, with OpenAI’s filing indicating a target valuation of about $852 billion and Goldman Sachs and Morgan Stanley as joint bookrunners (source: OpenAI official statement).

Anthropic disclosed a privacy issue: user prompts and code entered into Claude were collected for AI training, contradicting its “AI safety‑first” branding and sparking industry‑wide concerns about privacy policies (source: DoNews).

Apple announced at WWDC 2026 the most significant Siri AI upgrade since its 2011 debut, introducing the Apple Intelligence system built on a hybrid architecture of a self‑developed foundation model and Google Gemini, featuring screen awareness, personal context understanding, and cross‑app task execution, now packaged as a standalone app (source: Apple official).

Anthropic launched two new models: Claude Fable 5 for public use and Claude Mythos 5 for restricted access, both based on the same underlying architecture. Claude Fable 5 achieved an 80.3% score on the SWE‑Bench Pro benchmark and is priced at $10 per million input tokens and $50 per million output tokens (source: Anthropic official blog).

Meta open‑sourced the Scout and Maverick models from its Llama 4 series, which employ a mixture‑of‑experts (MoE) design, support multimodal image‑text input, and offer a context window up to 10 M tokens, surpassing GPT‑4o and Claude Sonnet 3.7 on several STEM and code benchmarks (source: Sohu Tech).

Google released Gemini 3.5 Live Translate, a real‑time speech translation model that supports over 70 languages while preserving speaker tone, speed, and pitch; developers can access it via the Gemini Live API and Google AI Studio preview (source: IT Home).

Security incidents surfaced across the ecosystem: Google Gemini’s voice assistant was found to contain a hidden vulnerability allowing crafted notifications via WhatsApp or SMS to trigger privileged actions, potentially hijacking smart‑home devices (source: Pacific Tech). Microsoft’s GitHub abruptly revoked access to dozens of open‑source repositories after hackers injected malware, which could steal passwords when AI coding tools like Claude Code or Gemini CLI opened compromised code (source: Security Brief).

Funding activity intensified: the Chinese startup 月之暗面 sought up to $20 billion in a new financing round, aiming for a post‑money valuation of $30 billion—a seven‑fold increase from its $4.3 billion valuation six months earlier, with ARR surpassing $200 million in April (source: Bloomberg via iHeima). Additional Chinese AI firms such as 阶跃星辰 filed IPO applications targeting valuations up to $12 billion, reflecting a broader acceleration of capitalisation in the domestic large‑model market (source: Sina Finance).

Collectively, these developments illustrate the AI sector’s transition from pure technological competition to a comprehensive race involving capital, ecosystem integration, privacy, and security considerations.