
Coding Is AI's First Real Business. Now OpenAI, Anthropic and Google Are in an All-Out War Over It.
Claude Code ignited a revenue explosion. OpenAI ordered its team to drop side quests and fight back. Now all three are racing to IPO on coding revenue.
The AI Post newsroom — delivering AI news at the speed of intelligence.
After five years and hundreds of billions of dollars, the AI industry finally found its first real business. It is not chatbots. It is not image generators. It is not search. It is writing code.
The Verge published a definitive account this week of how AI coding went from a GitHub experiment in 2021 to the most consequential product war in tech. The timeline is worth understanding, because the winner of this fight will likely become the most valuable AI company on Earth.
How Claude Code Changed Everything
When Anthropic released Claude Opus 4.5 in late 2025, benchmarks suggested it was good but not revolutionary. Then developers started using it in Claude Code over the holidays. Almost universally, they reached the same conclusion: it works. Not in the useful autocomplete way that previous tools worked. In the "I described what I wanted and got a working prototype" way.
Boris Cherny, the creator of Claude Code, told The Verge he already has AI writing 100% of his code. "It was just as surprising for me as it was for everyone else," he said. Claude Code went viral in a way that seemed impossible for a coding tool.
The revenue impact was immediate and staggering. Anthropic tripled its revenue to $30 billion in a single year, with Claude Code as the engine. One product turned a safety research lab into a cash machine.
OpenAI Panicked. Then It Declared War.
OpenAI saw the numbers and hit the alarm. According to The Verge, one of OpenAI's top executives recently told her team to stop doing "side quests" and focus entirely on competing with Anthropic and Claude Code. OpenAI launched a new $100 ChatGPT Pro tier on April 9, priced to match Claude Max dollar for dollar, with five times the Codex access. The message was clear: this is the fight that matters.
Google is not sitting out. It rolled out a command-line interface for Gemini and has been loading more coding features into AI Studio. All three companies are building what The Verge calls "AI super apps," and all three see coding as the core use case that gets developers locked in.
The Business Case Is Obvious. The Stakes Are Existential.
Developers are expensive. Software creation takes time. Any tool that lets companies hire fewer developers or ship faster is an easy sell. Nvidia CEO Jensen Huang recently said he would worry about any engineer who was not spending $250,000 a year on AI tokens. A 2025 study found 98% of developers use AI coding tools multiple times per week.
Companies across Silicon Valley now compete internally over who burns the most AI tokens. GPU access has become a recruiting tool. The spending is real, it is growing, and it flows directly to whichever AI company captures the coding market.
Both OpenAI and Anthropic are planning IPOs this year. Both need to show investors something concrete for the billions they have raised and burned. Coding revenue is that something. Whoever wins the developer market goes public with the strongest story Wall Street has seen since the smartphone era.
The Vibe Coding Wild Card
There is also a second market forming underneath the professional developer war: non-developers who code by describing what they want. Andrej Karpathy coined the term "vibe coding" in early 2025, and it stuck. People who had never written a line of code were suddenly prompting their way to working software. For those people, a barely functional prototype was enough, and these tools could deliver that reliably.
Startups like Cursor and Windsurf raised massive rounds to chase this market. But the gravitational pull of OpenAI, Anthropic, and Google may be too strong. When the platform owners decide your market is their market, the window for independents closes fast.
Coding was AI's first use case in 2021 and now, five years later, it is the first use case that actually makes money. The irony is hard to miss: the technology that threatens to replace software engineers is being sold primarily to software engineers. For now, they are buying. The question is how long that lasts before the tool finishes the job.
Analysis based on reporting by The Verge, The Next Web, and Sherwood News.