01 · Business

OpenAI targets September IPO with Goldman Sachs and Morgan Stanley

OpenAI is targeting a September 2026 IPO with Goldman Sachs and Morgan Stanley, planning to file confidential paperwork within weeks. The move sets up a direct public-market race with SpaceX—which now owns Elon Musk's xAI and is filing imminently—as two AI giants compete for investor capital and valuation dominance.

The details:
  • OpenAI crossed $3.4B in annualized revenue in Q4 2025, driven by API sales and ChatGPT subscriptions, though profitability status remains undisclosed.
  • SpaceX IPO filings were expected as soon as May 20, 2026, consolidating Musk's AI (xAI) and space businesses under one public company.
  • A court dismissed Elon Musk's lawsuit against OpenAI on statute of limitations grounds yesterday, removing a legal threat to OpenAI's corporate structure and financing.

Why it matters: OpenAI and SpaceX are now locked in a rare dual IPO sprint where execution and market timing matter as much as financials. The winner claims first-mover advantage in AI public markets, potentially anchoring investor appetite for the entire sector—and cementing which Altman-Musk vision of AI dominance gets bet on first.

Read the full story →

02 · News

Google AI Mode hits 1 billion monthly users as usage doubles quarterly

Google's AI Mode hit 1 billion monthly users in May 2026, with usage doubling every quarter since launch. The company is overhauling Search with a redesigned search box, AI Overviews on most queries, and agentic features launching this summer that will generate custom apps and interactive dashboards instead of traditional web links.

The details:
  • Google rolled out its biggest search box redesign in 25 years, using Gemini to autocomplete queries based on user context and intent across desktop and mobile.
  • Starting summer 2026, Search will generate full custom apps on demand—like a family outing planner with calendar integration and map embeds—powered by Gemini 3.5 Flash.
  • The new AI Overview nudge now sits at the bottom of overviews, obscuring top organic results and steering users away from traditional web listings by design.

Why it matters: Google is betting hard that AI-generated interfaces will replace the 10-blue-links model, and the 1 billion user milestone suggests the bet is working—at least in raw engagement numbers. But the strategy trades user choice for query inflation and marginalizes publishers at scale. Expect antitrust scrutiny to intensify as Google algorithmically buries organic results in favor of its own AI-generated outputs.

Read the full story →

03 · Models

Stability AI releases Stable Audio 3.0 with 6-minute song generation

Stability AI released Stable Audio 3.0 today, a family of four generative audio models that create full songs up to 6 minutes 20 seconds—more than double the 3-minute limit of last year's version. The company trained on licensed data from Warner Music Group and Universal Music Group, sidestepping the lawsuits hammering competitors Suno and Udio over unauthorized training.

The details:
  • Four models released at 459M, 1.4B, and 2.7B parameters; the small and medium versions ship with open weights for commercial use under $1M revenue.
  • Large model available only via paid API or self-hosting, with enterprise license required for companies above $1M annual revenue.
  • Stability AI hired Ethan Kaplan, former chief digital officer at Universal Audio and Fender, to lead a professional musician product suite around the new models.

Why it matters: Stability AI is trying to win the AI music wars by doing the thing its rivals refuse: paying for licenses upfront. A 6-minute song is still shorter than most album cuts, but the licensed-data moat matters more than the length ceiling—it lets Stability operate without the legal liability that's bleeding Suno and Udio. Watch whether major labels actually use these models or keep them at arm's length anyway.

Read the full story →

AI Box AI Box
Every AI model. One chat.
The latest models from ChatGPT, Claude, Gemini, Sora, ElevenLabs — 80+ models in a single chat. Compare answers side by side. Pick the best one every time.
  ChatGPT, Claude, Gemini, Grok, DeepSeek — in one chat
  Generate images & video with Sora, Veo, Ideogram
  Compare any two models side by side
  From $8.99/mo · 80+ models, all included
Try AI Box  → aibox.ai
Trusted by 3,000+ teams
04 · Business

IrisGo raises $2.8M seed to build a proactive AI desktop automation agent

IrisGo, backed by Andrew Ng's AI Fund, raised $2.8M to build a desktop agent that learns workflows by watching you perform a task once, then automates it forever without prompting. The stakes: automating repetitive work for knowledge workers at scale through preinstall deals with PC makers like Acer.

The details:
  • Co-founder Jeffrey Lai built the Chinese version of Siri at Apple and demonstrated the system ordering a coffee by recording one transaction, then repeating it autonomously.
  • Most processing happens on-device for privacy; cloud tasks require explicit user authorization with end-to-end encryption, a differentiator against cloud-heavy competitors.
  • IrisGo shipped beta macOS and Windows apps and secured a preinstall agreement with Acer to distribute the agent on new devices without acquisition spend.

Why it matters: Watch-once automation is the obvious next frontier for desktop AI—if it actually works. IrisGo's on-device architecture and Ng's credibility position it well, but the real test is whether single-example learning holds up across messy, multi-app workflows in the real world. OEM deals bypass consumer acquisition entirely, which matters in a crowded agent market where distribution wins or loses the game.

Read the full story →

05 · Business

Google embeds AI-generated product descriptions inside Search ads

Google is now using Gemini to write custom product descriptions for ads inside Search results, embedding sponsored items directly into conversational AI Mode responses. This expands ad real estate beyond traditional sidebars—ads now appear inline with AI-generated answers, full-screen product cards, and chatbot Q&As tied to advertiser sites, blurring the line between editorial content and paid placement.

The details:
  • A Nespresso Vertuo Up ad example shows Gemini synthesizing specs like 3-second heat-up time and 6 cup sizes into natural-language buying copy.
  • New ad formats include chatbot-driven Q&A buttons that open conversations powered by advertiser website data, plus full-screen sponsored product cards occupying the entire viewport.
  • Google first tested basic sponsored results in AI Mode last year; these new formats significantly expand ad surface area by embedding promotions structurally identical to organic results.

Why it matters: Google is monetizing AI Mode aggressively by making ads native to the conversational layer—not bolted on. This moves the conflict from sidebar ads to the core feed, where users already struggle to separate AI guidance from paid pitches. Advertisers get access to Gemini's copywriting; Google locks in revenue from its next-gen search interface. But user backlash over ad-editorial blur could intensify if sponsored items become indistinguishable from organic recommendations.

Read the full story →

Everything else in AI today

How was today's issue?

One tap helps me dial in tomorrow's send.

🔥
Nailed it
5 / 5
👍
Solid
3 / 5
😬
Missed
1 / 5

Share more feedback →

That's it for today.

Same stories, fuller context: catch the AI Chat podcast wherever you listen. Reply with what we missed — I read every one.

Browse all today's stories · forward to a friend.

Keep Reading