OpenAI Is Considering Suing Apple — Here’s Why

This one caught nearly everyone off guard. OpenAI has reportedly hired outside legal counsel to explore suing Apple over their ChatGPT integration deal — the agreement that placed ChatGPT inside Siri and Apple’s Visual Intelligence feature starting in 2024.

So what went wrong? According to internal reporting, OpenAI expected billions of dollars in revenue from the deal. The idea was simple: having ChatGPT baked into every iPhone would drive massive subscriber growth. Instead, the integration turned out to be hard to find, users rarely triggered it by name, and the financial returns have reportedly been “disappointing.”

“Unlike the multi-billion-dollar deal Apple maintains with Google for search, the OpenAI partnership involved no large direct payments, with both sides instead relying on mutual strategic benefits that OpenAI now claims have not materialised.” — LatestLY, May 15, 2026

OpenAI’s frustration goes beyond the numbers. The company also feels Apple didn’t do enough to promote the ChatGPT integration — no prominent placement, no advertising push, no real effort to get people actually using it.

The situation is further complicated by OpenAI’s own ambitions in hardware. The company recently acquired a device startup co-founded by Jony Ive — the legendary Apple designer behind the iPhone — and has been actively poaching Apple engineers. Understandably, that hasn’t gone over well in Cupertino.

Why This Matters for Regular Users

If you use an iPhone and rely on AI features, a messy OpenAI–Apple breakup could affect which AI model powers your Siri in the short term. It could also accelerate Apple’s push toward other providers — which brings us to story three.

Codex Is Now on Your Phone — And It’s Kind of a Big Deal

On May 14, OpenAI announced that Codex — its AI coding agent — is now accessible directly from the ChatGPT mobile app on both iPhone and Android. The update is currently in preview and available across all plans.

Here’s what that actually means in practice: Codex still runs on your Mac as a standalone desktop app, but you can now monitor it, steer it, approve or reject its next steps, and even start new tasks — all from your phone, while the actual computation keeps running on your laptop back home.

“This is more than the ability to remotely control a single task or dispatch new tasks to your computer. From your phone, you can work across all of your threads, review outputs, approve commands, change models, or start something new.” — OpenAI announcement, May 14, 2026

This is a meaningful quality-of-life upgrade for developers. Long-running AI coding tasks — the kind that take 20–40 minutes to complete — no longer chain you to your desk. You can kick something off, go for a walk, attend a meeting, and check in from your phone when Codex needs input.

How to Set It Up

• Update the Codex Mac app to the latest version
• Update the ChatGPT mobile app on your iPhone or Android
• Open Codex on Mac and find the new “Codex Mobile” section
• Scan the QR code with your phone — you’re connected instantly

Tools to Check Out:

• OpenAI Codex
• ChatGPT Mobile
• Claude Code

iOS 27 Will Let You Choose Claude or Gemini Instead of ChatGPT

This is the most significant structural change in consumer AI in months. Apple is set to open its Apple Intelligence platform in iOS 27 to third-party AI models — meaning you’ll be able to choose Anthropic’s Claude, Google’s Gemini, or other providers to power the AI features you use every day on your iPhone.

The system, internally called “Extensions,” will work through App Store applications. Users will be able to set their preferred AI model as the default for features like Writing Tools, Image Playground, and other Apple Intelligence capabilities.

This is a notable shift for Apple, a company whose entire identity is built around keeping its ecosystem tightly integrated and controlled. The decision to open up — even partially — reflects both competitive pressure and, likely, the deteriorating relationship with OpenAI.

What’s Changing in iOS 27:

• Users can set Claude, Gemini, or other approved models as their default AI
• Google’s Gemini is reportedly being tested to power a revamped Siri
• Apple holds its WWDC developer conference June 8–12 — full details expected there
• The iPhone 18 and a rumored foldable device are also in the pipeline for later this year

GPT-5.5 Is Out — What’s Actually Different

Amid all the drama, OpenAI also quietly shipped GPT-5.5, described as the company’s most capable model for agentic and multi-step work.

The headline claim: GPT-5.5 can plan, use tools, and verify its own outputs with less hand-holding from the user. OpenAI says it’s particularly strong at agentic coding, computer use, and early-stage scientific research — tasks where the AI needs to make sequential decisions over long time horizons without constant correction.

This isn’t a ChatGPT chatbot upgrade in the traditional sense. GPT-5.5 is primarily aimed at developers building autonomous AI systems — tools that don’t just answer a question but complete an extended workflow on your behalf.

GPT-5.5 vs Previous Models:

Anthropic Is Committing $200 Billion to Infrastructure

The less-discussed story of the week is arguably the most consequential long-term: Anthropic has reportedly committed over $200 billion toward cloud infrastructure and AI chips, largely in collaboration with Google Cloud.

This isn’t an abstract number. It reflects a bet that the bottleneck in AI going forward isn’t models or talent — it’s raw compute power. Whoever can run more inference, faster and cheaper, wins.

Anthropic’s Mythos model — a more advanced version of Claude currently in limited preview — has also made headlines after reportedly identifying critical security vulnerabilities in decades-old legacy financial systems. This positions AI not just as a creative tool, but as a serious systems auditing capability at a scale humans never achieved before.