For much of the past two years, OpenAI has lived inside a halo. Its chatbot, ChatGPT, became shorthand for artificial intelligence itself, the way “Google” once did for search. But the mood inside the company has shifted. According to people familiar with the internal climate, chief executive Sam Altman has declared what staff describe as a “code red”: not a panic, but a recalibration.
The phrase arrives at a moment when the map of power in artificial intelligence is being redrawn. Externally, rivals are no longer content simply to imitate. Internally, OpenAI is confronting the friction that comes when brilliance meets scale: engineering at the frontier is intoxicating; operating at planetary size is another discipline altogether.
Competition has grown less abstract and more structural. Google’s Gemini programme is not just a model but a distribution strategy, weaving itself into search, email and productivity software. The company does not merely release tools; it folds them into habits — a pattern closely examined in coverage by the Financial Times. Control in technology has always belonged to those who own the doorway, not just the device.
Alongside Google sits a quieter challenger. Anthropic has advanced deliberately through enterprise partnerships, positioning its models less as spectacle and more as infrastructure. Reporting from The Wall Street Journal suggests that large firms are now weighing reliability and governance as carefully as raw performance.
If OpenAI built the spark, its rivals are laying the cable. The distinction matters. One defines a moment; the other shapes a decade.
There is also the arithmetic. Running a model at global scale consumes electricity and capital with a relentlessness that does not make headlines. Each query has a real-world cost; each surge in demand tests the limits of data centres. Partnerships have helped — particularly with Microsoft — but reliance narrows the room for manoeuvre. Autonomy and subsidy rarely coexist for long.
Inside the company, emphasis is said to have shifted from novelty to steadiness: fewer launches, more optimisation; less theatre, more plumbing. It is the sort of pivot that signals maturity — or vulnerability — depending on who is telling the story.
Then there is the politics. What began as software has become social infrastructure. Users ask systems about health, grief, money and meaning. Lawmakers across the Atlantic are preparing frameworks that will eventually dictate what is permitted and what is restrained.
Altman now occupies one of the most difficult positions in modern technology: the architect asked to become custodian. A year ago, OpenAI was spoken of in the tone once reserved for Google in its early years. Today the conversation is colder, more pragmatic.
The question is no longer whether artificial intelligence will change the world — it already has. The question is who will control it when novelty expires and regulation arrives.
For more analysis and reporting on technology and global innovation, read more from our Technology Desk at Swikblog.
