Decoding why OpenAI is betting on India amid AI talent war

The corridors of Silicon Valley have witnessed many talent wars, but none quite like this. In 1976, when Apple co-founder Steve Jobs ‘poached’ engineers from Atari with the promise of ‘changing the world,’ the stakes were measured in thousands, not billions. Today's artificial intelligence (AI) battle sees companies throwing around compensation packages worth hundreds of millions of dollars for a single researcher. The prize isn't just market dominance, but potentially the future of human intelligence itself.

Meta's talent raid on OpenAI throughout 2024 and into 2025 has altered the competitive landscape. The company has recruited over 50 researchers and engineers from its rivals, offering packages that dwarf traditional Silicon Valley compensation. Some recruits reportedly received offers worth hundreds of millions of dollars, with one deal valued as high as $1.5 billion, according to reports. The numbers tell a story of pressure. 

OpenAI hit $13 billion in annualised revenue by July 2025, doubling from $6 billion in January and tripling from $4 billion in 2024. ChatGPT's weekly active users surged to 700 million in July, up from 500 million in March. This success has made OpenAI a target for competitors willing to pay astronomical sums for the talent driving these achievements.

Can money buy Meta the edge in AI?

Meta's recruitment strategy is quite dramatic and can surely turn heads. The company poached ChatGPT co-creator Shengjia Zhao, former OpenAI researchers Shuchao Bi, Jiahui Yu, and Hongyu Ren, along with Apple's Ruoming Pang, who had led foundational model work. In a striking move, the tech giant attempted to acquire entire startups, including a nearly $1 billion offer to co-founder Andrew Tulloch of Thinking Machines Lab, the venture started by former OpenAI COO Mira Murati.

Yet Meta’s high-stakes hiring spree shows that money alone doesn’t guarantee loyalty or impact. Shengjia Zhao, brought in with much fanfare, reportedly threatened to quit days after joining Meta and even signed paperwork to return to OpenAI, before being persuaded with the title of chief scientist of Meta's Superintelligence Labs. Other expensive hires also proved difficult to retain. Ethan Knight left after a few weeks, while former OpenAI researcher Avi Verma completed the intake process but never showed up for his first day.

The industry has taken note of Meta’s tactics. DeepMind's CEO Demis Hassabis didn't mince words on the Lex Fridman podcast. He noted that Meta is not at the AI frontier as of now, and while they could reach that stage eventually, it seems “rational” to hire from competitors to get ahead of the curve. 

His assessment cuts to the core issue. Meta's Llama models, released in 2024, failed to generate significant excitement, forcing the company into reactive mode, which basically equates to offering huge sums of money to top AI talent.

In fact, Anthropic co-founder Benjamin Mann critiqued this money-first approach and said, "My best case at Anthropic is we affect the future of humanity. My best case at Meta is we make money." 

This philosophical divide between financial incentives and mission-driven work has created a tension, to say the least, in AI development and makes me wonder whether breakthrough ‘innovation’ stems from purpose or paychecks.

OpenAI's counter-strategy

Facing a loss of its talent, OpenAI has deployed a retention strategy. The company sent its entire staff on a mandatory week-long break in July 2025 to break the cycle of 80-hour weeks and ward off Meta's poaching attempts. While the move was most likely to avoid poaching, it was an acknowledgement that burnout was one of the reasons making their workforce vulnerable to competitors' advances.

The company also reportedly distributed substantial bonuses to retain key personnel, though these defensive measures only addressed symptoms, not the underlying competitive pressure. 

Throughout 2024, roughly half of OpenAI's AI safety researchers had already left the company, citing concerns about the company's role in industry-wide problems. This predated Meta’s hiring.

Meanwhile, Meta has frozen hiring in its AI division after months of aggressive talent acquisition. The company is either satisfied with the current team or recognises that the strategy isn’t yielding expected results, especially since some of the newly hired talent have exited. The freeze coincided with a restructuring of Meta's Superintelligence Labs into four specialised teams. 

OpenAI aims to win through India 

Rather than engaging in an increasingly expensive bidding war for Silicon Valley talent, OpenAI has pivoted toward an entirely different approach, shifting its geographical focus through aggressive expansion in India. The company plans to open its first office in Delhi by the end of 2025, according to reports. It is also scouting local partners to establish a data centre in India with at least 1 gigawatt capacity.

This represents more than just an investment. India represents the world's second-largest ChatGPT user base, with 13.5% of global users. Sam Altman noted that usage grew four times over the past year. 

The timing coincides with India's IndiaAI mission, a government initiative aimed at positioning the country as a global AI leader. OpenAI has already appointed several key executives for its India operations, including Raghav Gupta as Head of Education for India and APAC, Sheeladitya Mohanty as Marketing Lead, and Akash Iyer as Social Lead.

The company's India strategy extends beyond talent acquisition to market penetration. ChatGPT's India-specific Go plan, priced at ₹399, undercuts competitors while building local user adoption. 

This pricing strategy is combined with partnerships with institutions like IIT Madras and AICTE, positions OpenAI to capture both consumer and educational markets simultaneously. It has launched an AI literacy programme in partnership with the Ministry of Electronics and IT called OpenAI Academy, while expanding Indic language support in GPT-5. 

The company will also host its first Education Summit in India, and Union IT Minister Ashwini Vaishnaw welcomed OpenAI’s OpenAI’s decision to establish a presence in India. 

Sam Altman's planned September 2025 visit to India may include announcing the data centre facility, transforming what began as a defensive move against Meta's poaching into an offensive strategy for market expansion. 

Meta’s largest user base is in India, with over a billion users across its apps, including Meta AI. Instead of fighting expensive battles for existing talent, OpenAI is creating new talent pipelines in markets that could have higher growth potential.

Playing the longer game

By establishing substantial infrastructure and local teams, OpenAI creates multiple competitive advantages simultaneously. Local data centres reduce latency for Indian users, government partnerships could provide regulatory cover, and educational initiatives could build long-term ecosystem relationships.

While Meta spent hundreds of millions acquiring individual researchers, OpenAI can build entire teams at a fraction of the cost while accessing India's vast technical talent pool. The country's engineering workforce, trained at institutions like the Indian Institutes of Technology, could offer both quantity and quality that Silicon Valley's limited talent pool cannot match.

When Microsoft established its India Development Centre in 1998, critics dismissed it as cost-cutting. Today, that facility employs over 8,000 people and serves as the company's largest R&D centre outside Redmond. Similarly, Google's India operations, initially focused on search localisation, now drive significant portions of the company's global product development.

The current AI talent war appears less about immediate recruitment battles and more about long-term development. While Meta burns cash on individual acquisitions with mixed results, OpenAI is building relationships that could provide sustainable competitive advantages for decades.

The ultimate irony may be that Meta's aggressive poaching strategy, intended to weaken OpenAI, has instead forced the company toward a more diversified and potentially stronger global position. Sometimes the most effective defence isn't matching your competitor's tactics; it's changing the game entirely.