Singapore’s sovereign wealth fund GIC has invested in Anthropic three times in the past year, backing the AI company at a valuation that has climbed to S$1.2 trillion, about $900 billion, according to people familiar with the matter. The latest round puts Anthropic within reach of the trillion-dollar club, a threshold reached by only a handful of companies in the world, and, until now, by no AI startup.
GIC’s repeated investments, three rounds in twelve months, make it one of Anthropic’s largest institutional backers. The fund, one of the world’s largest sovereign wealth funds, is known for patient, long-horizon investing, and its conviction is a signal: sovereign money, which can afford to wait out cycles, is betting that Anthropic’s models will be central to the next decade of computing.
Anthropic’s valuation growth has been extraordinary even by the standards of the AI boom. The company, founded in 2021 by former OpenAI researchers Dario and Daniela Amodei, has ridden its Claude model family to the front rank of AI labs, and its investors include Amazon and Google, which have poured billions into the company and its cloud infrastructure. The S$1.2 trillion valuation implies investors expect Claude to be a pillar of enterprise AI for years to come.
The same day the funding news surfaced, Anthropic upgraded its Claude voice mode, moving the product to its Opus and Sonnet models. The update improves the speed and naturalness of voice conversations, part of Anthropic’s push into consumer products to compete with OpenAI’s voice features. The pairing of a funding round and a product launch on one day showed the two tracks of the company’s strategy: build the most capable models, and put them in front of users as quickly as possible.
The controversy around the company is also growing. The United States has accused Chinese AI company Moonshot of stealing technology from Anthropic’s Fable model through knowledge distillation, a technique in which one model is trained on the outputs of another. The allegation, if substantiated, would be one of the first high-profile cases of claimed model theft between American and Chinese AI companies.
Knowledge distillation is common in AI research, models learn from other models all the time, and frontier labs use the technique internally to make smaller, faster versions of their biggest models. What separates legitimate use from theft, regulators and companies argue, is authorization: training on a model’s outputs without permission, and at scale, crosses a line that the industry is only beginning to define.
The case highlights how the U.S.-China technology rivalry has extended to the models themselves, the most closely guarded assets in the AI industry. Washington has already restricted the export of advanced chips to China and has debated limits on American AI models being made available there. The distillation accusation adds a new front: protecting the intellectual property inside the models.
GIC’s position in the middle is delicate. A Singaporean fund is one of the biggest backers of an American AI champion at a time when Washington is restricting Chinese access to American AI technology. Singapore has positioned itself as neutral ground in the rivalry, and GIC’s investments in both U.S. and Chinese technology companies have been scrutinized from both sides.
The valuation question is whether $900 billion is justified. Analysts are split. Bulls point to enterprise adoption accelerating, with Claude winning corporate contracts across banking, insurance, and software; bears note the cost of training frontier models, the intensifying competition from OpenAI, Google, and Meta, and the possibility that the market is pricing AI like a utility while the companies spend like startups.
Anthropic’s revenue has grown quickly, driven by corporations that see Claude as a safer choice than some rivals for regulated industries, according to people familiar with the company’s finances. The company has emphasized safety in its product design, a posture that has distinguished it in the AI race and that its customers say is a factor in their buying decisions.
Anthropic’s closest rival, OpenAI, has also raised money at valuations in the hundreds of billions, and the two labs have become the poles of the AI investment boom. The comparison is instructive: both companies face the same capital treadmill, spending heavily on compute to stay ahead, and both are betting that revenue from enterprise and consumer products will eventually cover the cost.
For GIC, the three rounds are a statement about where the next decade of computing value will be created. The fund has historically favored diversified, defensive positions, and its willingness to concentrate in a single AI company is unusual. The strategy appears to be a bet on the category as much as the company: whoever wins the AI race, GIC wants to hold a seat at the table.
Anthropic’s path to a trillion-dollar valuation runs through enterprise adoption, product expansion, and a legal and geopolitical environment that remains unsettled. GIC’s three rounds suggest its money is on the company navigating all three. The next test is whether the market agrees the price is right, and whether the U.S.-China dispute over its models becomes a problem the balance sheet cannot solve.
