Anthropic cuts unauthorised share platform list to four

A warning from a hot AI startup can move money before anyone trades a single share.

That is what happened after Anthropic, the maker of Claude, publicly named platforms it said were not authorised to trade its private shares. The company first listed eight names. It has now cut that list to four.

For Indian investors watching the AI rush from afar, this is not just Silicon Valley drama. It is a reminder that private shares are not like buying Reliance or Infosys on the stock exchange.

Anthropic trims its warning list

Anthropic has revised its public warning on secondary-market trading of its shares. The company now says Open Door Partners, Unicorns Exchange, Pachamama and Upmarket are not authorised to buy or sell Anthropic stock.

The earlier warning had named eight platforms. That wider list unsettled investors and drew pushback from firms involved in private-share trading.

Anthropic said any sale or transfer through unauthorised firms would be void. In plain English, the buyer may think they own exposure to Anthropic, but the company may refuse to recognise it.

That matters because Anthropic is still private. Its shares do not trade freely on a public exchange. The company controls who can buy, who can sell, and how transfers get recorded.

Why private AI shares are messy

A secondary market is a place where existing shareholders sell shares to new buyers. In listed companies, this happens every day on exchanges.

In private companies, the rules are tighter. Founders, employees and early investors often cannot sell without company approval.

The reason is simple. A private company wants control over its shareholder list. It also wants to stop unknown investors from entering through side deals.

This becomes tricky when demand explodes. AI companies now attract huge interest from funds, wealthy individuals and structured products.

Many investors want exposure before an IPO. They fear missing the next Microsoft or Nvidia-style run. That fear creates a market for indirect access.

But indirect access can come with fine print. Some products may not offer actual shares. Some may offer rights linked to shares. Some may depend on future approvals.

For a retail investor in India, the lesson is blunt. If the underlying company does not recognise the transfer, the fancy document may not mean much.

Hiive pushes back hard

Hiive was among the platforms named in Anthropic’s first warning. The company has now been removed from the updated list.

Hiive chief executive Sim Desai said his platform does not allow share transfers without company approval. He also said Anthropic’s earlier public position created uncertainty and hurt Hiive’s reputation.

Desai said Hiive would have worked with Anthropic on a clear market message, had the company approached it first.

Anthropic has not offered a fresh public comment on the dispute. But its revised list suggests the company heard at least some of the market reaction.

This is where the story gets important. Publicly naming trading platforms is not common. Companies usually handle transfer disputes quietly through lawyers and shareholder agreements.

Anthropic chose a louder route. That sent a signal to the market that it wants tighter control over its cap table.

A cap table is simply the company’s ownership record. It shows who owns what. In a private company, that record is everything.

The valuation makes scrutiny sharper

Investor interest in Anthropic is intense because the numbers are staggering. The company announced a new funding round that raised $65 billion and valued it at $965 billion, including fresh capital.

That valuation puts Anthropic near the top of the global AI race. It also places it ahead of OpenAI by some recent measures.

For context, valuations at this scale are not just about current revenue. Investors are paying for expected dominance in AI models, enterprise tools and future computing demand.

That is a big bet. It assumes companies will keep spending heavily on AI software. It also assumes model makers can turn usage into durable profits.

This is where private markets can mislead casual investors. A headline valuation does not work like a stock price on the National Stock Exchange.

A listed share moves every trading day. Buyers and sellers discover price in public. Regulators require regular disclosures.

Private valuations come from funding rounds. A small group of investors sets the price. The terms may include special protections that ordinary holders do not get.

So when a startup says it is worth hundreds of billions, that does not mean every share carries the same risk. It also does not mean every investor gets the same deal.

What Indian investors should watch

Indian investors have become much more curious about global private tech. Wealth platforms, offshore funds and feeder products have widened that access.

That access can be useful. It also demands more caution than buying a listed mutual fund.

The first question is simple. Does the company itself recognise the share transfer or interest being sold?

The second question is just as important. Are you buying actual shares, a fund unit, or only economic exposure through another structure?

The third question is about exit. If you want to sell later, who will buy it, and at what discount?

These questions sound boring during a boom. They become painful when a company challenges the trade.

The Anthropic episode shows how fast confidence can shift. The company’s warning hit publicly traded funds with exposure to Anthropic. Private-share brokers also faced disruption.

That is the hidden risk in private markets. The price may look attractive, but liquidity can vanish quickly.

Liquidity means the ability to sell without taking a big loss. In private shares, liquidity often depends on company approval and buyer appetite.

AI is real, and the opportunity is large. But the market around private AI shares is still uneven.

For Indian families and professionals building wealth, the safer instinct is not to avoid ambition. It is to know exactly what one owns. In this AI boom, the fine print may matter as much as the model.