Google Remakes Search, Anthropic Hits $10.9B, Musk Loses: AI News (May 22, 2026)

1. Google I/O 2026: Search Gets Its Biggest Redesign in 25 Years

Google I/O's keynote ran nearly two hours and touched every major product. The search changes are the ones that matter most.

AI Mode now has 1 billion monthly users. AI Overviews appear on 48% of all queries. The search box itself got redesigned for multimodal, conversational input. Information agents can search in the background 24/7, meaning users don't even need to visit google.com.

For content publishers, this accelerates the shift from "rank and get clicks" to "get cited inside AI answers." Sites with original data and first-person testing (like what we do here) get cited. Generic listicles get replaced.

On the model side, Gemini 3.5 Flash now beats 3.1 Pro on coding and agentic benchmarks while maintaining Flash pricing. That's the small model outperforming the big model at a fraction of the cost. Gemini Omni creates video grounded in real-world data, and Gemini Spark is a 24/7 personal AI agent rolling out to Ultra subscribers.

Google also restructured its pricing: new $100/mo (≈₹9,300/mo) AI Ultra tier with Spark access and 5x usage, former $250 tier cut to $200/mo (≈₹18,600/mo). The $20/mo (≈₹1,860/mo) Pro tier stays unchanged.

Full coverage: Gemini 3.5 Flash + OpenAI verification · AI Ultra $100 pricing · Gemini Review

2. Anthropic Projects $10.9B in Q2 Revenue, First Profit

Anthropic told investors it expects $10.9 billion (≈₹1.01 trillion) in Q2 2026 revenue, up 130% from Q1's $4.8 billion, with its first-ever operating profit of $559 million. The growth rate outpaces what Google and Facebook reported before their IPOs.

Claude Code is the engine, reportedly crossing $2.5 billion in annualized revenue. Enterprise clients account for 80% of sales with 300,000+ businesses on the platform. Anthropic is raising at $900 billion, which would put it above OpenAI's $852 billion valuation.

The caveat: Anthropic warned profitability may not last through 2026 due to scheduled compute costs. SpaceX's IPO filing revealed Anthropic pays $1.25 billion per month for GPU access through 2029.

Full coverage: Anthropic $10.9B + OpenAI IPO · Claude Review

3. OpenAI Files Confidential S-1 for September IPO

OpenAI is working with Goldman Sachs and Morgan Stanley on a confidential S-1 filing, targeting a September 2026 listing at $852 billion to $1 trillion. This creates a three-way IPO race: SpaceX in June ($1.75T), OpenAI in September, Anthropic in October.

The company missed internal revenue and user targets in early 2026, making the IPO timing interesting. Coinbase prediction markets show 85% probability that OpenAI lists before Anthropic.

What it means for users: Both companies need user growth for their IPO narratives, which means keeping ChatGPT and Claude at $20/mo and expanding free tiers. Enjoy it while the pre-IPO pricing lasts.

4. Musk Loses $134B OpenAI Lawsuit in Under 2 Hours

A California jury unanimously rejected all claims in Elon Musk's $134 billion lawsuit against OpenAI, Sam Altman, and Greg Brockman. The jury deliberated for less than two hours, finding the case blocked by the statute of limitations.

The case was always a long shot. Musk's core argument was that OpenAI abandoned its nonprofit mission, but the jury didn't even reach that question. The statute of limitations defense was enough. Legal analysts note this effectively closes the door on similar challenges to OpenAI's restructuring.

Full coverage: Musk Lawsuit + Google I/O Preview

5. Cursor Ships Composer 2.5, Matches Opus 4.7 at Fraction of Cost

Cursor released Composer 2.5, their most capable in-house coding model. It scores 79.8% on SWE-Bench Multilingual (nearly matching Opus 4.7's 80.5%) and 63.2% on CursorBench v3.1 (up from 52.2%), at standard pricing of $0.50/M input tokens (≈₹46.50).

The model uses the same Kimi K2.5 base as Composer 2 but with 25x more synthetic training data and 85% of compute budget on Cursor's own RL pipeline. Cursor also announced a partnership with SpaceXAI to train a significantly larger model from scratch on Colossus 2.

I've updated our Composer 2 vs Claude Sonnet comparison (the site's most-visited page) with Composer 2.5 benchmarks.

Full coverage: Composer 2.5 deep dive · Cursor Review · Composer 2 vs Claude Sonnet

6. OpenAI Launches AI Content Verification Tool

OpenAI announced a three-layer approach to content provenance: C2PA Content Credentials, Google DeepMind's SynthID watermarking (notably, a competitor's technology), and a public verification tool where anyone can check if an image was generated by OpenAI's models.

The tool only detects OpenAI-generated content for now, with cross-platform support planned. SynthID watermarks survive common edits like cropping and format changes. OpenAI was transparent about the limitation: if no watermark is detected, it doesn't mean the image isn't AI-generated.

7. Google Rolls Out May 2026 Core Update

Google announced the May 2026 core update on May 21, two days after I/O. It's the second broad core update this year and is expected to take up to two weeks to complete (through roughly June 4).

Google described it as "a regular update designed to better surface relevant, satisfying content for searchers from all types of sites." No new guidance was provided. SEO tracking tools reported intense ranking volatility starting during I/O, days before the official announcement.

For publishers: don't make reactive changes during the rollout. Wait until after June 4 to assess impact. Sites with original data and first-person testing tend to benefit from core updates.

Quick Hits

• White House AI EO postponed again. Trump said "I didn't like certain aspects" of the draft executive order for voluntary AI model access. Microsoft and xAI had already agreed to participate.
• Samsung Intelligent Eyewear announced at I/O. Gemini-powered smart glasses launching fall 2026 with Warby Parker and Gentle Monster partnerships.
• Google Flow + Flow Music creative apps now available on mobile. AI-powered creative tools expanding beyond the browser.
• Chinese AI models hit 60% of all usage on OpenRouter. A quiet but significant shift in the open-source AI ecosystem.

What I'm Watching Next Week

• Composer 2.5 real-world testing. I'm running the full 15-task benchmark to update the comparison scores.
• Gemini 3.5 Flash testing. If the keynote claims hold, the Gemini review score needs updating.
• Core update impact. Tracking ranking changes across all our pages through June 4.
• OpenAI S-1 details. If the confidential filing happens, investor reactions will signal whether the $852B+ valuation holds.
• Anthropic funding round close. The $900B round with 48-hour investor allocations should finalize soon.

This roundup covers May 16-22, 2026. Previous roundup: May 15. All prices at ≈₹93/USD.

More from RawPickAI: 2026 AI Tools Reality Check · Transparency Index · Compare AI Tools · Best AI Coding Tools 2026