Inside Anthropic's Singapore Expansion and Its Pivotal 2026

Anthropic is making a decisive move into Southeast Asia. As of early June 2026, the AI safety and research company behind the Claude model family has listed its first dedicated Singapore-based positions, translating its surging global valuation into a tangible local hiring footprint. The expansion is unfolding against a dramatic backdrop: unprecedented capital injections from Singapore’s sovereign wealth funds and a high-stakes ethical standoff with the US military that could cost the company billions.

The Singapore Jobs: Who Anthropic Is Hiring and What They’ll Earn

On June 4, 2026, Anthropic’s careers page advertised four open positions in the city-state, spanning finance, product support, and economic research [15, 21]. While a specific office address has yet to be disclosed, the company expects successful candidates to be physically present at the workspace for at least 25% of their time, confirming that a local office is being set up .

The known Singapore-based openings are:

• Regional Research Economist (Economic Research): This role carries a published salary range of SGD 307,200–331,200 per year (approximately USD 228,000–246,000) and requires a PhD [17, 18].
• APAC Accounting Lead: This Singapore-based head of regional corporate accounting will report to the head of international accounting in the company’s Dublin office, with responsibility for end-to-end finance operations across Asia-Pacific [15, 21].
• Product Support Specialist (Singapore): First listed in April 2026, these positions are designed to be the foundation of Anthropic's global support brand. They are open to candidates holding a bachelor's degree or equivalent experience .
• An additional finance- or product-support role referenced by The Straits Times, for which a specific title was not immediately available .

These hires are a fraction of a broader global hiring spree. Anthropic currently has over 800 open positions worldwide, with the greatest concentration in San Francisco, New York, and Seattle [14, 16].

The GIC Connection: $65 Billion in Two Funding Rounds

The scale of Anthropic’s expansion is directly tied to its funding, and Singapore’s GIC has been at the center of it. The sovereign wealth fund has led or co-led two massive rounds in 2026:

• February 2026: GIC co-led a $30 billion Series G round alongside Coatue Management, placing Anthropic’s post-money valuation at $380 billion. GIC and Coatue each contributed roughly $1.5 billion to this raise [5, 8].
• May 2026: GIC and Temasek anchored a staggering $65 billion Series H round, with GIC acting as a co-lead investor. This round valued the company at $965 billion post-money, cementing its status as the most highly valued AI startup [2, 13].

Beyond equity financing, GIC and Anthropic deepened their operational ties with a joint venture on May 4, 2026, alongside Blackstone, Hellman & Friedman, and Goldman Sachs. The new firm is a standalone entity dedicated to helping enterprises rapidly deploy Claude into their core business operations . GIC and Anthropic also co-hosted their first joint event in Singapore on April 23, bringing together about 150 senior leaders from the region’s tech and investment communities .

This torrent of funding is supported by Anthropic’s breakneck revenue growth. The company’s revenue doubled from $4 billion to $9 billion in the six-month period ending in early 2026, a trajectory that has attracted investors like Microsoft and Nvidia alongside sovereign funds [6, 5].

The Model Too Dangerous for Public Launch

On April 7, 2026, Anthropic unveiled Claude Mythos Preview. The company simultaneously published a 244-page System Card for its most powerful model to date and announced it would not be making it available to the general public [39, 46].

Launched under the security initiative Project Glasswing, Mythos Preview achieved a 93.9% score on the SWE-bench benchmark . It was provided exclusively to approximately 50 partner organizations — including AWS, Apple, Google, Microsoft, NVIDIA, CrowdStrike, and JPMorgan Chase — for strictly defensive cybersecurity applications [37, 38].

The model’s core capability is its agentic skill at identifying and exploiting software vulnerabilities at immense speed and scale. Anthropic stated that Mythos Preview’s “large increase in capabilities has led us to decide not to make it generally available” . Over its first month of operation, the model found 23,019 vulnerabilities across 1,000+ open-source projects, with 90.6% confirmed as genuine upon independent sampling . A separate independent assessment by the UK’s AI Security Institute found the model succeeded at expert-level hacking tasks 73% of the time .

The decision sparked immediate controversy. Within weeks, reports surfaced of unauthorized access via a third-party vendor environment, and Anthropic confirmed it was investigating the breach . Critics argued that entrusting such a powerful system exclusively to a coalition of the world’s largest corporations — without public oversight — constituted “putting the calamity makers in charge” . As of May 2026, prediction markets assigned only a ~7% probability to a public release occurring by June 30, and the company’s official position remained that Mythos-class models “could reach the public once the defensive infrastructure is proven” — a condition with no fixed date [36, 37].

The Pentagon Standoff and Its Cost

While GIC and Temasek were writing billion-dollar checks to Anthropic, the company's relationship with its home government was imploding. A dispute over military use of its AI models escalated into an unprecedented public clash in early 2026.

• February 24–27, 2026: Defense Secretary Pete Hegseth gave CEO Dario Amodei an ultimatum to allow the Department of Defense (DoD) unrestricted use of Anthropic’s models or face severed ties. When Anthropic refused, Hegseth publicly designated the company a national security “supply chain risk” [27, 31].
• March 4–5, 2026: The Pentagon formally notified Anthropic of the designation, barring it from federal contracting. The US military terminated its existing ~$200 million contract with the company and ordered its contractors to stop using Anthropic products [28, 33, 29].
• March 9, 2026: Anthropic filed a lawsuit in federal court in California, arguing the designation was legally unsound and a violation of due process .

The crux of the dispute was Anthropic’s refusal to accept contract language that would permit its technology to be used for “all lawful purposes,” a standard clause that could have enabled mass domestic surveillance or fully autonomous weapons systems . In a public statement, Amodei noted that while Anthropic was the “first frontier AI company to deploy our models in the US government’s classified networks,” the company had to draw a red line at applications it deemed unethical .

Anthropic estimates it could lose “multiple billions of dollars” in revenue this year from the fallout . The episode has become a landmark test of whether AI safety and ethics frameworks can survive friction with national security imperatives .

The juxtaposition is stark: as Singapore’s state-backed funds propel Anthropic to a near-trillion-dollar valuation, the company is simultaneously suing its own home government and voluntarily locking away a model it says is too dangerous to give to the world.