Amazon and Google will together own nearly 30% of Anthropic after its IPO, and the company billed as OpenAI's independent challenger just became a subsidiary of the cartel it claimed to challenge.

That number matters because AI is the next frontier of speech and information — and when two corporate giants hold a third of the 'independent' alternative, independence is a marketing slogan, not a fact. TechCrunch reported the details from Anthropic's disclosure: Amazon will hold 14.7% of the company, Google will hold 13.7%. Combined, that's 28.4% — a stake large enough to dictate terms without firing a shot.

The structure is designed to look harmless. Both Amazon and Google's shares are a special class of non-voting stock. But as TechCrunch noted, those shares "can receive dividends and convert to standard voting Class A shares when sold." Non-voting today, voting tomorrow — the moment either giant decides to cash out or flex muscle, the conversion mechanism is already built in. This isn't a firewall. It's a screen door.

Then follow the money. Amazon didn't just buy equity — it bought a revenue pipeline. Anthropic has committed to spend more than $100 billion on Amazon's AWS cloud services over the next decade. Amazon profits coming and going: it owns a huge chunk of the company and collects a nine-figure annuity from its cloud dependency. That's not investment. That's vertical integration dressed up as venture capital.

And Amazon could have gone even bigger. TechCrunch reports that Amazon holds billions in promissory notes it chose not to convert yet. If it converted all of them, it would own 23.8% of Anthropic — nearly a quarter of the company. It held back. For now.

Google's stake comes from the same playbook — note conversions plus a private placement sale as part of the IPO. A broader syndicate of tech heavyweights — AMD, Broadcom, Micron, Microsoft, Nvidia, and TSMC Partners — agreed to buy $26 billion in non-voting stock, with Google and Nvidia each committing $10 billion. The entire semiconductor and cloud establishment is parking capital inside Anthropic. When the whole industry owns a seat at the table, nobody at the table is a rival.

Anthropic was founded on the promise of AI safety and responsibility — a counterweight to OpenAI's speed-over-safety culture. That pitch raised billions. Now the bill comes due, and the people collecting are the same platforms that censor speech, deplatform dissidents, and throttle content they don't like. The 'responsible AI' company just sold a third of itself to the least responsible custodians of free expression in American business.

The real question isn't whether Anthropic's technology will stay independent. It's whether any AI company can stay independent when the only customers rich enough to train these models are the same ones who want to control what the models say.