Here’s what’s on our plate today:

• 🧪 Why OpenAI's superapp is really an IPO story.

• 📰 Opendoor's India exit, Fable dodges biology questions, Anthropic's CEO wants more regulation.

• 🛠️ Weekend To-Do: stress-test the superapp, compare enterprise contenders, audit switching costs.

• 🗳️ Poll: Will OpenAI's superapp bet actually work?

Let’s dive in. No floaties needed…

The Laboratory

TL;DR

• IPO first, product second: OpenAI filed confidentially for an IPO targeting a $730B to $850B valuation, and the superapp announcement arrived almost simultaneously, combining ChatGPT, Codex, and third-party services into one platform designed to make enterprise customers harder to lose.

• Anthropic forced the pivot: Ramp data tracking 50k+ U.S. companies shows Anthropic surpassed OpenAI in enterprise adoption in April 2026, from 32.3% to 34.4%, after sitting at 8% just a year earlier.

• The product graveyard is recent: OpenAI’s earlier ventures, including Atlas and Sora, launched with headlines, shipped with flaws, and faded.

• The stakes: If the superapp lands, OpenAI locks in enterprise infrastructure revenue before markets scrutinize its losses. If it follows the pattern, the IPO story becomes much harder to tell.

Why OpenAI’s superapp is really an IPO story

There is a common idea in product design that the more successful a tool becomes, the more tempted its creators are to keep adding new features. At first, that can make the product more useful. But beyond a certain point, it becomes cluttered, harder to use, and less effective at the things that made people like it in the first place.

Many technology companies have fallen into this pattern, assuming that success in one area gives them the right to expand into every other one. WeChat is one of the few exceptions, largely because it grew in a market with limited foreign competition and few comparable alternatives. In most cases, however, attempts to turn a single product into an all-purpose platform have produced mixed results, often weakening the original experience. OpenAI is now sprinting toward it anyway.

According to reporting by the Financial Times, the company is preparing the largest overhaul of ChatGPT since the product launched in 2022, converting it from a question-and-answer chatbot into what it is calling a “superapp”: a single platform combining coding tools, AI agents capable of executing autonomous tasks, image generation, and third-party services from partners like Canva and Booking.com.

The changes are expected to begin rolling out in weeks. While the announcement landed quietly, if you follow the trail of decisions OpenAI has made over the past 18 months, the superapp shifts from being a product revelation to being the conclusion of a financial argument.

The story behind the product

To understand why this is happening now, you need to understand what OpenAI is trying to become before it goes public. The company confidentially filed its IPO prospectus with the SEC on approximately May 22, with Goldman Sachs and Morgan Stanley leading the process and a target listing window between September and November 2026. The private-market valuation sits somewhere between $730B and $850B. At that price, OpenAI would rank among the largest IPOs in American history.

However, for OpenAI to achieve its IPO ambitions, it must first address the economics of its business. According to the Wall Street Journal’s projections, OpenAI generated $13.1B in full-year revenue in 2025 and spent approximately $22B to produce it, resulting in a net loss of roughly $9B. The burn rate is expected to remain at around 57% of revenue through 2026 and 2027, with operating losses projected to widen before any hoped-for turn toward profitability around 2029 to 2030. HSBC analysts have estimated that the company may need more than $207B in additional capital before it gets there—all this, even as OpenAI expects to record a $14B net loss in 2026 alone.

The challenge is convincing public-market investors that a loss-making AI chatbot deserves an $850B valuation. For that, you need a story that sounds more like a platform than a product, and a superapp could be that story.

Enterprise contracts, which account for roughly 40% of OpenAI’s revenue across its 2M business customers, offer a far more attractive business than consumer subscriptions because they are both stickier and more lucrative. A consumer paying $20 a month can cancel with little consequence. Still, an enterprise that has woven a single AI platform into its workflows, software stack, and internal processes faces a far more complex and costly decision: switching providers. Seen through that lens, the drive to turn ChatGPT into a superapp becomes more about making its products indispensable to businesses’ daily operations, deepening customer dependence while creating more opportunities to sell additional services than by adding new features or chasing user growth. With an IPO increasingly in view, building that kind of durable revenue base is not just a growth strategy but a financial imperative.

The competitive pressure underneath it

However, beyond the IPO, there is also a more immediate problem. According to spending data tracked by Ramp across more than 50k U.S. companies, Anthropic’s enterprise adoption reached 34.4% in April 2026, crossing OpenAI’s 32.3% for the first time. A year earlier, Anthropic had 8%. That reversal happened while OpenAI was building browsers, video generators, and consumer social features. Anthropic was building enterprise relationships.

To counter Anthropic, OpenAI’s strategy has been to consolidate by folding Codex, its coding agent, more deeply into ChatGPT, expanding it with new enterprise capabilities, including Sites (hosted web applications generated from documents), role-specific plugins connecting 62 business applications, and an in-place editing tool called Annotations.

Non-developers now make up 20% of Codex’s 5M weekly users and are adopting it at three times the rate of engineers. That number is the real signal inside the superapp thesis: the company believes it can extend AI assistance from developers into every white-collar role. If that transition works, Codex stops being a coding tool and starts being a general-purpose business infrastructure.

The pattern you need to see

In October 2025, OpenAI launched Atlas, a Chromium-based browser built around ChatGPT and pitched as a new gateway to the internet. The launch briefly rattled Alphabet investors, but security concerns and rollout delays quickly overshadowed the original vision. Atlas survives, but the transformative narrative surrounding it has largely faded.

Before Atlas, there was Sora, OpenAI’s text-to-video generator, launched in September 2025. The app reportedly reached 1M downloads within five days, and discussion quickly turned to its commercial potential, including reports of a possible $1B partnership with Disney. Yet the excitement faded almost as quickly as it arrived. According to OpenAI’s help center, the consumer Sora app was discontinued on April 26, 2026, with API support scheduled to end in September, less than a year after launch. The Disney partnership never materialized, and peak monthly revenue was estimated at roughly $540k, a tiny figure for a company operating at OpenAI’s scale.

Another ambitious project remains on the horizon. OpenAI is developing a screenless AI device with former Apple design chief Jony Ive, whose startup the company acquired for $6.4B in May 2025. OpenAI’s chief global affairs officer told Axios in January 2026 that the product remained on track for a second-half 2026 launch, while Sam Altman has described early prototypes as “jaw-dropping.” Yet the device will enter a market that has already witnessed the collapse of Humane’s AI Pin, which promised a similar vision of ambient AI computing and was discontinued within a year of release.

None of this requires a conspiracy to explain. Even unsuccessful launches can shape investor perception by reinforcing the idea that a larger platform is taking shape. For OpenAI, whose valuation has climbed to roughly $850B despite continuing losses, that narrative carries particular importance. The superapp strategy is the most consequential version of it yet, arriving just as the company nears an IPO.

What the superapp actually has to clear

The case for OpenAI’s superapp strategy is straightforward. ChatGPT’s 900M weekly active users give the company a distribution advantage that few technology firms have ever enjoyed. Codex’s adoption outside traditional software development suggests that AI tools can expand into a much broader range of work, and enterprise AI remains young enough that a successful platform could become deeply embedded in business operations, making it difficult for competitors to dislodge.

The counterargument is just as compelling. No major Western technology company has successfully built a true superapp. Meta spent years trying to turn Facebook, Messenger, and WhatsApp into a unified ecosystem that combined communication, payments, and commerce, only to abandon that vision and shift its attention to the metaverse. Elon Musk rebranded Twitter as X with the explicit goal of creating an “everything app,” yet that ambition remains largely unrealized.

And while WeChat’s success in China is often cited as a model, it emerged within a regulatory and competitive environment that does not exist in North America or Europe.

Friday Poll

Headlines You Actually Need

• Opendoor's India exit sparks AI debate: Opendoor pulling its operations out of India is fueling a wider conversation about how AI and automation are reshaping the global outsourcing model.

• Fable dodges biology questions: Anthropic's new Fable model reportedly won't answer basic biology questions, reigniting the debate over how far safety guardrails should go before they hurt usefulness.

• Anthropic's CEO wants more regulation: In an exclusive interview, the CEO is calling for stronger AI regulation, a notable stance from a frontier lab as Washington weighs how to govern the technology.

Weekend To-Do

• Stress-test the superapp idea: Try running a full workflow inside ChatGPT end to end, draft, code, image, booking, and see where it actually holds together versus where you'd still reach for a dedicated tool.

• Compare the enterprise contenders: Put the same business task through ChatGPT and Claude to feel why enterprise adoption is the real battleground, not consumer features.

• Audit your own switching costs: List every AI tool your team relies on and mark how painful each would be to replace; that's exactly the lock-in OpenAI is chasing.

Meme Of The Day

The Toolkit

• Regie.ai: AI sales agent that researches prospects, writes personalized outbound at scale, and handles follow-ups so reps can focus on closing.

• Replit AI: A browser-based coding environment with an AI agent that builds, debugs, and deploys apps from prompts, with no local setup required.

• Sourcegraph: Code intelligence platform with Cody, an AI assistant that understands your entire codebase so it can answer questions and write code that actually fits.

Rate This Edition