OpenAI hands Southeast Asia and Australia to Meta's former India VP

Every six weeks or so, one of these labs takes a country-level executive off an American tech company's bench in Asia. Nine moves in twelve months. The nine is a minimum. The supply is the problem. Arjun Jaggi, an applied AI researcher and executive adviser to Fortune 500 companies, told Rest of World that the number of executives in India who have actually run a country-level P&L (profit and loss) for a global tech company, negotiated with Indian regulators, and built enterprise trust at scale is small, maybe a few hundred people across all of Big Tech India.

You can read the job off the resumes. OpenAI's first hire in India, Pragya Mishra, came from Truecaller, and earlier was the first WhatsApp hire in India. Its Asia-Pacific director, Kiran Mani, joined from JioStar in March after 13 years at Google. Devanathan's decade at Meta covered consumer platforms, gaming, commerce, business messaging, payments, advertisers and partnerships. "The prized asset is not generic familiarity with American technology but dual institutional literacy: knowing how a U.S.-headquartered company allocates capital and risk, and knowing how Indian customers, governments, and partners turn global strategy into local reality," Sanchit Vir Gogia, chief analyst at Greyhound Research, told Rest of World.

Recruiters put it the same way. Deepali Vyas, global head of data and AI executive search at ZRG Partners, said that once a company decides a region is a strategic growth market, it starts looking for executives who can essentially operate as mini-CEOs. The people who fit already have jobs. "At the leadership level, the scarcity is more pronounced because experience cannot be created overnight. Most of the large technology companies are increasingly competing for the same relatively small group of proven executives," Neeti Sharma, CEO of TeamLease Digital, told Rest of World.

The hiring comes without a number. Neither OpenAI nor Anthropic would discuss the strategy or make the new hires available, and the reporting carries no adoption, revenue or retention figures for either lab in India or Singapore. So the case that these labs are now limited by policy access and enterprise selling instead of model quality rests on where the recruiting money is going, and on analysts reading that spend. It is a reasonable read, but not a measured one.

For an operator deciding whether a new country team changes their own rollout, two things settle it: whether the deployment needs a local answer on data handling or sectoral rules, and whether it needs a contracting counterparty who can escalate above the global support queue. If either is true, the new office is worth a test, and the test is cheap. Ask for one artifact only an in-country team could produce, such as a written data-handling position signed by someone local, or a named escalation contact with a phone number. If the reply comes back from the same queue as before, the hire was made for policymakers and enterprise pipelines.