*Update*: Anthropic files for IPO — AI costs burden users and startups

Anthropic has filed a formal application for an initial public offering today. Spiegel reports that the company was most recently valued at around 965 billion dollars – surpassing OpenAI and making it the world's most valuable AI company. Anthropic did not comment on the exact contents of the filing, but market observers expect the placement to occur later this year. The move follows a 65 billion dollar funding round that the company completed just days earlier.

N-tv reports that Anthropic is thus following in the footsteps of Elon Musk's SpaceX and could pave the way for further AI IPOs – OpenAI harbors similar plans. The IPO would give Anthropic access to public capital, which is urgently needed for the enormous computational power required by modern AI models.

Rising costs burden users and startups

Yet this very computational power is driving costs for everyone. El Espectador explains that prices are rising across all AI segments – AI agents are cited as the main driver, consuming significantly more computing power than simple chatbots. Users feel this directly: The West reports that more than 150,000 Australians already pay for AI subscriptions like ChatGPT or Claude – a massive increase compared to three years ago.

On the losing side of the boom are startups founded before the ChatGPT breakthrough in 2022. CNBC describes how the AI boom has channeled more than 250 billion dollars into companies like OpenAI and Anthropic – while drying up hundreds of older startups. Investors, talent, and attention flow almost exclusively to large frontier model providers, while former darlings stagnate or close.

That cost savings are nonetheless achievable is demonstrated by a technical example from practice: Hacker News points to a report showing that a team reduced its Spark computing costs by 44 percent through AI-powered monitoring – a hint that AI tools can at least partially offset their own cost pressure when deployed strategically.

The overall picture of today is thus contradictory: one company is heading to the stock market and signaling maturity, while the industry simultaneously shows that the boom produces its own costs – for users, for infrastructure teams, and for an entire generation of startups that built too early.