Incentive wars: Why walking away can be wiser

By backing out of the race for Google's AI data centre, Karnataka has chosen prudence and restraint over optics

In a world that glorifies relentless competition, Karnataka’s refusal to outbid Andhra Pradesh for Google’s $15 billion AI data centre may seem like a loss, but in game theory – and governance – wisdom often lies in knowing when not to play. In game theory, a war of attrition describes a contest where two players keep spending resources to outlast each other. The winner is the one who holds out longer, but often ends up worse off than if it had walked away early.

I see this dynamic play out every year in my MBA classroom during the Rupee Auction Game. I auction a Rs 500 note in an open-cry format. The rules are simple: the highest bidder wins the note, but both the highest and the second-highest bidders must pay their respective bids. What starts as a playful experiment soon turns serious. As bids rise, ego overtakes logic. I have seen students bid Rs 1,000 for that Rs 500 note – driven not by logic, but by the urge to “win”.

That same logic seems to underpin the current debate over Google’s decision to invest $15 billion in a 1-GW AI data centre in Andhra Pradesh instead of Karnataka. Almost immediately, commentators framed it as a “missed opportunity” and a “policy failure” for Bengaluru. Yet, beneath this narrative lies a nuanced story.

Reports suggest that Andhra Pradesh offered incentives worth Rs 22,000 crore, including 25% subsidies on land and water, free electricity, and reimbursement of state GST. On paper, this looks like a major victory. But as with the Rupee Auction Game, the real question is: at what cost?

Data centres are among the most energy- and water-intensive infrastructures in the digital economy. They consume vast quantities of power and potable water for cooling. Navi Mumbai, which hosts nearly 60% of India’s data centres, has seen its average water demand surge from 350 million to 460 million litres per day – even as residents face periodic water shortages for drinking and household purposes. The benefits of attracting such projects must, therefore, be weighed against their long-term ecological and social costs too.

Around the world, multinational firms have mastered the art of triggering incentive wars between governments, extracting concessions that can erode public value. Consider Amazon’s HQ2 contest in 2018, seeking a location for its $5-billion second headquarters in North America. Newark, New Jersey, one of the 20 finalists, offered an extraordinary $7 billion in tax breaks, grants, and rebates. Economists and activists labelled the frenzied contest “anti-competitive,” urging governments to collaborate instead of undercutting each other. When Amazon finally chose Arlington, Virginia, and Queens, New York, the latter’s $1.5 billion incentive package sparked such public outrage that the company withdrew months later. The supposed prize turned into a public relations debacle.

In 2014, Tesla invited US states to bid for its $5-billion Gigafactory. Nevada won by offering $1.4 billion in incentives, including free land, tax abatements, and electricity discounts. Later disclosures revealed that the company had engineered a war of attrition-like contest among states, ensuring it would profit, no matter who won. These episodes illustrate a pattern: corporations gain leverage by exploiting political competition. Governments, fearing the optics of “losing” investment, overcommit public resources for short-term headlines.

Over the past few years, India has emerged as one of the world’s most promising data centre markets. In January, Maharashtra announced an $8.3 billion investment from AWS, part of the company’s larger $12.7 billion plan to develop cloud infrastructure in India by 2030. Reportedly, more than 15 major players – from Japan’s NTT and Singapore’s STT GDC to India’s Reliance and Adani – are competing for space in this fast-growing market.

India’s appeal is clear: lower power costs, abundant land, and expanding renewable energy capacity make it attractive for energy-hungry infrastructure. But this digital gold rush comes with ecological and governance challenges that are easy to overlook.

The right, rational choice

A single 20 MW data centre using conventional water-cooling systems can consume over 1.3 million litres of potable water daily. As cities like Navi Mumbai, Bengaluru, and Chennai host ever more data centres, their already strained water supplies are being pushed to the brink. Experts warn that 40% of Indian cities could run out of drinking water by 2030.

Globally, the environmental impact of data centres is becoming impossible to ignore. Google’s own emissions have risen 51% since 2019, despite its renewable energy investments. The UN’s International Telecommunication Union reports that operational emissions from major tech firms – Amazon, Microsoft, Meta, and Alphabet – have risen between 138% and 182% since 2020. In effect, the infrastructure powering our digital future is quietly becoming one of its largest climate challenges.

The competition for data centres is the new frontier in India’s industrial policy. With cloud computing, AI, and digital infrastructure expanding rapidly, states are eager to host these billion-dollar facilities. Entering an incentive war for data centres – which bring limited direct employment and heavy environmental costs – risks turning public competition into a negative-sum game, where every player ends up worse off.

Here, quitting early can be the smartest move. Yet, in a society that glorifies victory and frowns upon retreat, walking away is often mistaken for weakness. It is not. It is a recognition that not every prize is worth the price.

If India’s states learn to cooperate instead of undercutting each other – setting common benchmarks, environmental standards, and fiscal limits – the country could turn these corporate contests into truly sustainable partnerships. Recognising when not to play – when the cost of winning outweighs the benefits – is a mark of foresight that policymakers must exercise.

Karnataka’s decision might appear cautious, even defeatist. But in game theory terms, it could reflect an understanding that rationality sometimes lies in restraint, not escalation.

(The writer is an assistant professor at IIM Bangalore)

Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.