OpenAI and Anthropic embrace Australia's AI regulation as Chinese competition reshapes the market

Why are OpenAI and Anthropic cheering on regulation in Australia? The answer has global reach

The companies hope to follow in the footsteps of SpaceX, which raised $86bn and soared to a $2.1tn valuation after it listed on public markets in June Top US AI developers Anthropic and OpenAI cheered when Australia announced it would set new AI rules. Big tech celebrating limits on their Silicon Valley VC-funded free-for-all might seem counterintuitive but there's a much broader play than just what happens in one relatively small market. Anthropic and OpenAI aren't even on sharemarkets yet but this week lost billions in predicted value after being shown up by a Chinese startup. Moonshot AI on Friday launched a new model, Kimi K3, advertising it as the first open-source model of its size, allowing more user modification than Anthropic's Claude and OpenAI's ChatGPT. Analysts are judging K3 is competitive with both the leading US models, which are also less freely customisable. Moonshot had to temporarily pause new subscriptions after demand for the new model strained capacity. Implied values for Anthropic's sharemarket debut has slumped US$232bn to $1.56tn from Friday to Tuesday on IG's trading platform. OpenAI dropped US$160bn to $1.16tn. The two companies are still privately owned and valued at less than US$1tn each, so IG's numbers are market bets, not true values. But they hope to follow in the footsteps of SpaceX, which raised $86bn and soared to a $2.1tn valuation after it listed on public markets in June. Anthropic has already begun registering to list on US sharemarkets, which would allow it to raise billions from new investors. The better the story it can tell, the more money it can raise. Chinese competition does not make for a good story. Nor do surprise lawsuits from unhappy creatives, such as the one Anthropic had to settle for US$1.5bn on Monday (Tuesday AEST). The company will pay authors about US$3,000 for each of an estimated 500,000 books covered by the settlement. One of the authors whose work Anthropic scraped was Andrew Charlton, Australia's assistant technology minister. Charlton says the new regulation will mean Australian creatives can choose whether AI models are trained on their work and big tech companies will have to pay them. Yet those new rules could be a boon to big tech, Charlton says. "If we are upfront with what we expect in this Australian standard, we can give clarity to investors that in some ways enhances the attractiveness of Australia," he told Sky News on Monday. Malik Ahmed Khan, an equity analyst at Morningstar Equity, says technology companies would find it much easier to invest in Australia as they learn where regulation is heading. The companies will also have an easier job convincing investors to buy in at high prices if they can show they're focusing on building AI and managing the risk of running afoul of governments or local laws, Khan says. While Australia is one of few countries to now be regulating AI directly, with little direct impact on companies based overseas, Anthony Albanese hopes to set the global tone. In his speech announcing the regulations, he cited Australia's leadership on social media bans for under-16s as proof. Khan says more countries following Australia's lead on AI could boost to companies' prospects and market values. "A positive deal in Australia could enable them to pursue similar deals across the globe," he said. Avoiding a tech wreck Regulation would also give companies the prospect of favourable terms. Albanese signalled he would consult "closely" with industry and other countries to legislate regulations that would aim to enable AI's arrival, not restrict it. OpenAI and Anthropic were quick to position themselves as partners ready to help shape the new approach. Dario Amodei, Anthropic's chief executive and a former OpenAI vice-president, has already given an indication of what big tech thinks it should look like. In a June essay, he called for a coalition of democracies to dominate in AI and lock out China. While there are real national security concerns in this area, there's also a healthy dose of self-interest in blocking out competitors. With the launch of Kimi's K3, analysts believes it is increasingly likely that the US will block businesses from exclusively using Chinese-based models over national security fears. Amodei also called for domestic regulation, asking governments to test, evaluate and even block AI models if they presented "unacceptable risks" - but with protections against local "political favoritism or arbitrary decisions". Anthropic's feud with the Trump administration has left it on a US supply chain blacklist. Days after Amodei published his essay, the US government forced Anthropic to suspend access to its Mythos 5 models for foreign nationals. Anthropic's critics have accused it of self-interest as it nears its initial public offering (IPO) on sharemarkets. The Pentagon's Kirsten Davies said in June: "Some things are simply more important than revenue cycles, clickbait and pre-IPO valuation." Yet Amodei's warnings about security concerns have been echoed by Albanese and by Five Eyes intelligence agencies. Aruna Sathanapally, the CEO of Australia's Grattan Institute, met the Anthropic chief when he visited Canberra in April. She said Amodei's concerns seemed sincere. "He was seeing things within his own company that were giving him great concern about what Anthropic, as well as their competitors, may be capable of developing," Sathanapally said. "Just because they have vested interests in the particular shape of regulation, it doesn't mean that the calls for regulation themselves are not well founded. It just means ... you need to keep alive to competitive neutrality in how you design it."

Australia's AI gamble: Who wins, who loses, who leaves

James Thomson and Ronald Mizen on why CEOs are losing patience with the AI giants, what's driving the government's policy shift, and whether Anthony Albanese is overplaying his hand. In a reminder of how fast everything is moving in the AI era, Chinese start-up Moonshot AI released Kimi K3 last week - a model that is only slightly less capable than the latest offerings from Anthropic and OpenAI. It came as corporate leaders, like Palantir Technologies CEO Alex Karp, are starting to speak out about the unbridled power of the AI giants and as governments around the world are cracking down on the development of massive data centres.