Connect with us

AI

OpenAI Turns ChatGPT Into a Super App Before Its IPO

OpenAI plans to turn ChatGPT into a super app by merging Codex and the Atlas browser into one platform, ahead of a confidential IPO filing targeting a valuation near $852 billion.

Published

on

OpenAI is planning the biggest ChatGPT overhaul in the product’s history, building a super app that merges its coding platform Codex and the Atlas browser into a single interface, according to a Financial Times report published June 7 citing more than a dozen current and former employees. The redesign, set to begin rolling out in coming weeks, will surface partner apps from Canva and Booking.com alongside OpenAI’s own tools and give AI agents the access to handle multi-step tasks across a user’s desktop. “Chat is dead,” a senior OpenAI employee told the Financial Times. OpenAI did not respond to a Reuters request for comment on the report.

Reuters reported in May that OpenAI was preparing a confidential U.S. IPO filing, with Goldman Sachs and Morgan Stanley advising, targeting a listing as early as this year at a private-market valuation of $730 billion to $852 billion. The super app redesign starts rolling out before OpenAI’s first public prospectus has cleared the confidential phase.

Three Products in One Shell

The super app consolidates three products that currently require separate downloads and logins. ChatGPT handles conversational queries and image generation. Codex, OpenAI’s autonomous coding platform, can write new features from a text description, fix bugs, run tests in cloud environments and propose pull requests; it has grown to more than 2 million weekly active users. Atlas, a Chromium-based AI browser OpenAI launched in October 2025, puts ChatGPT inside the browsing layer so the AI can read pages, answer questions about what it sees and take actions on the web on a user’s behalf.

Product What It Does Current Status Role in the Super App
ChatGPT Conversational AI, image generation, general queries Live on web, iOS, Android Front door; routes users toward paid and partner products
Codex Autonomous coding agent: writes, tests and deploys code Live; 2M+ weekly active users Commercial engine; primary driver of enterprise contracts
Atlas Chromium AI browser; reads pages, takes web actions macOS only; Windows in development Native web layer; enables agents to operate across the internet

Thibault Sottiaux, who runs OpenAI’s core product and platform and previously led Codex, described the end state to the Financial Times:

What we’re building towards is where you have your own personal agent that is capable of helping you across everything in your life, be it personally or at work. You can connect through it on your mobile, desktop or web. When you’re in the car, you can talk to it.

Desktop users are the initial target; OpenAI confirmed to CNBC that the mobile ChatGPT app stays separate. Greg Brockman, OpenAI’s president, will temporarily oversee the product integration and organizational changes while Fidji Simo, the company’s chief of applications who joined from Instacart in May 2025, leads the commercial rollout for enterprise sales. Within the unified platform, Codex functions as an agent the other components can invoke, Atlas provides the native web layer for browsing and live information gathering, and the shared session context means a user can research a coding problem in Atlas, generate a solution in Codex and discuss it in ChatGPT without any of the three tools losing the thread.

The Enterprise Ground Claude Already Owns

The overhaul follows a market-share reversal in business AI spending. The May 2026 Ramp AI Index, which tracks corporate card and invoice data across more than 50,000 U.S. businesses, found that Anthropic’s Claude had overtaken ChatGPT in business adoption for the first time. A year before that crossover, fewer than one in 25 businesses on the platform paid for Anthropic. By April 2026, nearly one in three did.

  • 34.4% of U.S. businesses paid for Claude in April 2026 per the Ramp AI Index; ChatGPT fell to 32.3%, down from a peak near 36.5% in mid-2025
  • 70% of first-time enterprise AI purchases went to Anthropic in head-to-head matchups against OpenAI, per Ramp’s March 2026 index
  • $2.5 billion estimated annualized run rate for Claude Code, Anthropic’s agentic coding tool, per Reuters
  • 0.3% growth in OpenAI’s business adoption over the 12 months Anthropic quadrupled its share

OpenAI retains structural advantages. The company reported 9 million paying business users and 1 million distinct business customers in its most recent disclosures, and Microsoft’s Copilot products, built on OpenAI’s models, cover an estimated 15 million paid enterprise seats through Microsoft 365. On raw enterprise user count, OpenAI’s installed base is still substantial. But the Ramp data tracks where businesses direct new AI spend, and on that measure the trend was running toward Anthropic before OpenAI announced anything about a super app. First-purchase decisions shape the contracts that compound into multi-year commitments, which is why the 70% loss rate on new enterprise deals is the figure that matters most here.

The Wall Street Journal reported that Simo told employees at an all-hands meeting in March that the company couldn’t afford side quests given Anthropic’s gains in enterprise and coding customers. Her X post that week put the same message publicly: “Companies go through phases of exploration and phases of refocus; both are critical. But when new bets start to work, like we’re seeing now with Codex, it’s very important to double down on them and avoid distractions.”

Codex as the Commercial Engine

The reorganization centers on Codex because enterprise coding contracts carry fundamentally different economics. A ChatGPT Plus subscription costs $20 per month. An enterprise Codex deployment can run to millions annually; Anthropic now has more than 1,000 customers paying over $1 million per year for Claude, a figure that doubled in under two months after its February 2026 Series G funding round per Reuters and CNBC. OpenAI’s enterprise revenue exceeded 40% of total as of its spring 2026 disclosures and is on track to reach parity with consumer revenue by year-end.

The Canva and Booking.com partner integrations work as a conversion funnel. ChatGPT has 900 million weekly active users and 50 million paid subscribers, meaning the vast majority pay nothing. A partner app that routes a user to Canva for a design project or Booking.com for a hotel search brings a revenue share and generates behavioral data that supports premium plan pitches. OpenAI executives described ChatGPT to the Financial Times as a way to introduce customers to higher-value products.

Hotels and travel platforms have been moving into ChatGPT since late 2025. IHG Hotels launched conversational hotel search and direct booking inside ChatGPT on June 3, routing guests to IHG’s own reservation channels without leaving the AI interface. Booking.com’s role in the super app follows the same model: a commerce partner available inside ChatGPT rather than as a separate destination.

Codex drives the enterprise upgrade path specifically because it requires a paid plan at the professional or enterprise tier. A developer who queries ChatGPT about a coding problem and gets directed toward Codex’s full agentic capabilities needs an upgrade to proceed; the free tier becomes the discovery layer. Ramp’s lead economist Ara Kharazian noted in the May 2026 index that Codex performs comparable tasks at lower cost than Claude Code, with minimal switching friction between the two platforms. Uber’s CTO disclosed separately that the company burned through its entire 2026 AI budget in four months, largely on Claude Code, with per-engineer monthly API costs between $500 and $2,000. That cost pressure is the opening Codex is built to step into.

The IPO Equation

A $14 Billion Loss Year

OpenAI generated $13.1 billion in revenue in 2025 and spent approximately $22 billion to do it. Internal projections cited by the Financial Times put the 2026 net loss at $14 billion. CFO Sarah Friar confirmed in a January 2026 post that annualized revenue had crossed $20 billion by year-end 2025; Sacra Research, a private market data firm, estimated the figure climbed to $25 billion by February 2026, driven by subscriptions and API adoption. OpenAI’s internal timeline puts profitability in 2030.

The company closed a $122 billion funding round in March at an $852 billion post-money valuation, with commitments from Amazon, NVIDIA and SoftBank. HSBC analysts have estimated OpenAI may need more than $207 billion in additional funding by 2030, even accounting for projected revenue growth. Reuters reported in May that OpenAI was preparing a confidential U.S. IPO filing with Goldman Sachs, JPMorgan and Morgan Stanley advising. The 60-to-90-day window between a confidential filing and a public S-1 points to the earliest prospectus arriving in late summer.

What Public Investors Will Scrutinize

The S-1 will be the first time investors outside private rounds can see audited full-year revenue, gross margin after compute costs and the contract structure behind the valuation. Consumer user count, which OpenAI has emphasized most publicly, doesn’t address any of those questions directly. A platform with 900 million weekly active users, the majority on free plans, shows a different margin profile than one built around multi-year enterprise contracts.

Enterprise deals carry better unit economics: bulk API pricing and multi-year terms generate higher margins than $20-per-month subscriptions competing with a free tier. If the super app accelerates Codex adoption among enterprise teams before the prospectus goes public, the margin composition investors see at the S-1 stage is a better picture of the commercial platform the company has argued it can become. The gap between OpenAI’s $852 billion private-market valuation and the audited financials a public filing must show is what the prospectus will have to bridge.

What the Overhaul Still Hasn’t Solved

Atlas runs on macOS only. The Windows version is still in development, a gap that matters because most corporate computing environments run Windows. A business evaluating the super app as a productivity platform won’t have the full product until the Windows build ships, and OpenAI hasn’t given a public timeline for that release.

The rollout is staged rather than simultaneous. OpenAI said “the coming weeks” for initial changes to ChatGPT’s website and mobile apps, but the full merger of ChatGPT, Codex and Atlas proceeds in phases: Codex expands its capabilities first, then the three products merge. No firm date for the complete integration has been announced.

OpenAI’s product history offers context. Sora, the company’s video generation tool, launched as a standalone app in September 2025 with a TikTok-style feed, briefly hit number one in the Apple App Store, then saw usage stall. OpenAI is now folding Sora into the main ChatGPT app as part of the same streamlining. That arc, from dedicated launch to reabsorption in under a year, is the backdrop for Simo’s “refocus” framing: the 2025 product expansion spread resources before building depth in any single area. The super app is the correction.

Competitive pressure doesn’t wait for a phased rollout. Anthropic’s enterprise crossover happened in April, before OpenAI made the super app public. Claude Code’s gains in coding, the specific category this consolidation targets, preceded the response. Whether the unified platform ships fast enough, and on enough operating systems, to shift the enterprise mix before the S-1 surfaces is the question the next few months answer.

OpenAI’s public prospectus, when it clears the confidential phase, will show the gross margin that 900 million weekly users alone don’t justify. The super app has a narrow window to change what that number looks like.

Logan Pierce is a writer and web publisher with over seven years of experience covering consumer technology. He has published work on independent tech blogs and freelance bylines covering Android devices, privacy focused software, and budget gadgets. Logan founded Oton Technology to publish clear, no nonsense tech news and reviews based on real hands on testing. He has personally tested and reviewed dozens of mid range and budget Android phones, written extensively about app privacy, and built and managed multiple WordPress publications over the past decade. Logan holds a bachelor's degree in English and studied digital marketing at a certificate level.

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Trending