You probably used ChatGPT this morning. Most of the internet did too — OpenAI remains the most-used AI product on the planet by a comfortable margin. It's the most-funded, most-hyped company in the space. Everything's fine. The king is comfortable on the throne.

Except the throne is on fire. In the days since OpenAI closed a $122 billion funding round at an $852 billion valuation — the largest private raise in human history — its competitors didn't exactly sit around. Anthropic is locking in enterprises with a full agent-hosting platform. Google is handing out models that rival paid ones for free. The gap between "most popular" and "most profitable" is getting wider by the day.

So what did OpenAI do with its first week as the world's most expensive startup?

The double move: April 8–9

On April 8, OpenAI published a blog post titled "The next phase of enterprise AI" — a corporate strategy declaration disguised as a blog. Key number: enterprise customers now generate 40% of OpenAI's revenue, up from a minority share a year ago, with 9 million paying business users. Sam Altman celebrated 3 million weekly Codex users — Codex being OpenAI's AI coding agent (think: a robot programmer that writes code inside your project) — by resetting everyone's usage limits as a gift. "We will do this for every million users up to 10 million," he wrote. Generous? Sure. Also a growth hack.

The next day, April 9, came the consumer punch: a $100/month ChatGPT Pro tier. Not a random price point. Anthropic charges exactly $100/month for Claude's top consumer tier. CNBC didn't even pretend otherwise — their headline read "OpenAI looks to take on Anthropic with $100 per month ChatGPT Pro subscriptions." The tier gives you 5x the usage limits of the $20/month Plus plan and 10x Codex usage (promotional, through May 31). There's also a $200/month tier with 20x limits for the truly addicted.

What this actually reveals

Forget the pricing chess for a second. The interesting part is where OpenAI is spending its attention.

Codex — the coding agent — is the real weapon here. It went from 2 million to 3 million weekly users in under a month, according to Codex head Thibault Sottiaux. OpenAI also rolled out Codex-only seats on April 2 with pay-as-you-go pricing — token-based billing (tokens are word-chunks that AI processes, roughly ¾ of an English word) instead of flat message rates. No seat fees, no rate limits. This is OpenAI trying to become the default IDE — the tool where developers actually write code — not just a chatbot.

Meanwhile, the Frontier platform (launched February 5) is OpenAI's enterprise play — a place where companies like Uber, State Farm, and Intuit build and deploy AI agents with shared business context and governance. The twist: Frontier supports agents from Google, Microsoft, and Anthropic. OpenAI is betting that owning the platform matters more than owning every model on it.

The valuation question, revisited

We covered the raw math when the $122 billion round closed three days ago: $25 billion annualized revenue, $852 billion valuation, a multiple that makes nosebleed tech stocks look conservative. That analysis stands. What the April 8–9 moves add is a clearer picture of how OpenAI plans to grow into that number.

The strategy is a deliberate revenue-mix shift. Token-based Codex billing targets developers who'll weave it into daily workflows — sticky usage that compounds. The $100 and $200 consumer tiers extract more from power users who won't flinch at price hikes. Enterprise at 40% of revenue is growing fast, with Dresser (OpenAI's CRO, ex-Slack CEO) saying she's "never seen this level of conviction spread so quickly" in enterprise AI adoption.

But "growing toward" isn't "already there." Anthropic runs closer to 80% enterprise and just launched a managed-agents platform with session-hour pricing — purpose-built infrastructure for the contracts that don't churn when a competitor drops prices by $5/month. Dresser is probably right about conviction spreading. The question is whether that conviction translates into OpenAI contracts or just AI contracts in general.

What this means for you

If you're building on OpenAI's API — the programmatic interface that lets your app talk to GPT models — the Codex-only seats and token-based pricing are genuinely good news. Lower barriers, flexible billing, no commitments. If you're paying for ChatGPT Plus at $20/month, the new $100 tier is worth considering only if you're a heavy Codex user who hits limits regularly. For everyone else, the $20 plan didn't get worse.

If you're choosing between OpenAI and Anthropic for enterprise deployment, the Frontier platform's multi-vendor approach is a smart hedge — you're not locked into OpenAI's models alone. But smart hedges are also admissions that your own models might not always win.

The real game

Remember when the AI race was about who had the smartest model? That was 2024. In April 2026, it's a three-front infrastructure war. Google commoditizes the model layer by giving Gemini away. Anthropic locks in the enterprise layer with managed agents and deep integrations. And OpenAI — the company you used this morning — is trying to fight on all three fronts simultaneously with $122 billion in fresh ammunition.

The $122 billion bought time, not technology. And judging by the pace of this week's moves, OpenAI knows exactly how fast that clock is ticking.

Sources: