The Ferrari and the Dirt Road, India’s AI Summit, the Ten Potholes, and the Uncomfortable Truth About Building a Vishwaguru on a Fractured Foundation
In a few days, New Delhi will host the AI Impact Summit, the fourth in a series of global gatherings that began at Bletchley Park in the United Kingdom in 2023. That first summit was consumed by anxieties about AI safety—the existential risks, the alignment problem, the prospect of machines escaping human control. Today, those worries have been largely tossed aside as the world witnesses a US-China race for technological dominance, with the European Union scrambling to play referee. The focus has shifted from safety to supremacy, from caution to competition.
India’s chosen theme for the summit is a deliberate and pointed departure. It is drawn from ancient Sanskrit: sarvajana hitaya, sarvajana sukhaaya —welfare for all, happiness for all. This is not merely a rhetorical flourish; it is a declaration of intent. It asserts that the ultimate measure of AI’s success is not the size of its models, the speed of its chips, or the market capitalisation of its developers, but its contribution to human flourishing.
As Bhaskar Chakravorti argues in the accompanying analysis, this is a tall order. AI is not ushering in much sukha these days. The early delights of playing with ChatGPT have faded. Investors fret that the race for the most powerful models will trigger a stock-market crash of unprecedented proportions. College graduates and workers wonder which jobs will disappear. A multitude of “goddamners” worry about human extinction. The mood has shifted from euphoria to anxiety.
Yet Chakravorti insists that India has a genuine shot at leading the world—not in the race for ever-more-powerful models but in the design of AI for purpose. The potential applications are vast. Seventy per cent of the world’s food is produced by low-productivity smallholder farmers. Four and a half billion people lack access to essential healthcare. Seven hundred and thirty-nine million adults and two hundred and fifty million children cannot read or write. AI can help with each of these challenges. India already has examples: the Kisan E-Mitra chatbot handling 20,000 daily queries in 11 languages; Telangana’s Saagu Baagu project doubling chilli farmers’ earnings; the eSanjeevani telemedicine platform conducting 389 million virtual consultations; AI-powered learning platforms reaching millions.
But examples are not systems. Pilot projects are not scaled solutions. And India’s ambitions will remain unrealised unless it confronts the ten structural potholes that Chakravorti identifies—the gaps in connectivity, energy, talent, supply chains, governance, and infrastructure that threaten to swallow the Ferrari before it can leave the garage.
The Potholes: Ten Structural Constraints on India’s AI Ambitions
Chakravorti’s list of ten “potholes” is not a random catalogue of complaints; it is a systematic diagnosis of the constraints that will determine whether India’s AI aspirations are realised or frustrated.
First, the internet connectivity gap. Only 24 per cent of rural households have internet access, compared to 66 per cent in urban areas. Of the 125 countries that Chakravorti’s Digital Planet research centre studies, India is close to the bottom in digital gender parity. The AI revolution cannot benefit those who are not connected. Closing this gap is not merely a matter of social justice; it is a precondition for inclusive growth.
Second, the energy challenge. AI is energy-intensive. Training large models and running inference at scale require reliable, abundant electricity. India’s energy infrastructure is inadequate, its grid is unreliable, and its transmission capacity is weak. These problems will be stressed further by the growing demands of AI. Harmonising energy, climate, and AI goals is not a luxury; it is a necessity.
Third, the talent deficit. For every ten AI roles in India, there is only one qualified engineer. This is not a problem that can be solved overnight. Building a workforce with the skills to develop, deploy, and maintain AI systems requires sustained investment in education and training. Talent is a constraint, and it is binding.
Fourth, supply chain dependence. India imports over 90 per cent of its semiconductor chips, along with high-purity chemicals, gases, and silicon wafers. The US-China rivalry has already fragmented the global tech ecosystem, and this fragmentation will only intensify. Shifting to hardware manufacturing and building domestic compute infrastructure will not be easy.
Fifth, regulatory fragmentation. AI governance in India is characterised by uncertainty and complexity. Customs clearance and documentation requirements vary across states, slowing innovation and raising costs. Streamlining these processes is essential if India is to become a competitive location for AI development.
Sixth, the need for “good enough” AI. The race for ever-more-powerful models is driven by the logic of the frontier. But AI for purpose does not need to be at the frontier; it needs to be versatile, adaptable, and appropriate to local conditions. Engineering “good enough” AI is a different challenge from engineering the most powerful AI, and it requires a different set of skills and priorities.
Seventh, infrastructure security. India’s AI ambitions will be built on its digital public infrastructure—Aadhaar, UPI, DigiLocker, and others. This infrastructure must be secured against cyber threats, data breaches, and system failures. The consequences of a major breach would be catastrophic.
Eighth, the capital chasm. Seed funding for AI start-ups exceeds Series A through C investments, when companies need resources to scale. While early-stage funding has improved, start-ups still rely heavily on foreign capital beyond Series B. Bridging this chasm requires patient, long-term investment that Indian capital markets have not yet demonstrated.
Ninth, the actual dirt road. Inadequate physical infrastructure—roads, ports, logistics—adds costs and slows operations. An AI-powered supply chain is still a supply chain; if the underlying physical infrastructure is broken, the AI cannot fix it.
Tenth, the need for a new metric. Beyond the US-China-EU trinity, India must articulate a fourth possibility. It must measure success not in terms of model size or compute power but in terms of the yield of the average smallholder farm, the life expectancy of the poorest, the percentage of those who can read this page. That is the AI race that India can—and ought to—win.
The Opportunity: AI for Purpose
The potholes are real, but so is the opportunity. India has demonstrated that it can use technology to serve underserved communities at scale. The examples Chakravorti cites are not isolated anecdotes; they are proof points of what is possible.
The Kisan E-Mitra chatbot handles 20,000 queries daily in 11 regional languages. This is not a futuristic vision; it is a functioning system. Telangana’s Saagu Baagu project doubled chilli farmers’ earnings while reducing pesticide and fertiliser use. This is not a laboratory experiment; it is a field-tested intervention. The eSanjeevani telemedicine platform has conducted 389 million virtual consultations. This is not a pilot; it is a scaled programme. AI-powered learning platforms have reached millions of learners, with impressive penetration in Tier II and Tier III cities and rural areas.
These successes share common characteristics. They are designed for purpose, not for power. They address specific, well-defined problems. They leverage India’s digital public infrastructure. They are built by teams that understand local conditions and constraints. And they are evaluated rigorously, with attention to what works and what doesn’t.
The challenge is to move from examples to systems, from pilots to scale. This requires confronting the potholes head-on. It requires investment in connectivity, energy, talent, supply chains, and infrastructure. It requires regulatory reform and capital market development. It requires a sustained commitment to building the Ferrari while fixing the dirt road.
The Summit’s Challenge: From Chest-Thumping to Problem-Solving
Chakravorti’s warning against the summit being “co-opted into a platform for chest-thumping” is timely and necessary. The temptation to celebrate India’s achievements, to showcase its successes, to project an image of technological prowess is strong. But celebration without self-criticism is empty. Showcasing without problem-solving is performative.
The summit must be a moment of honest reckoning. It must acknowledge the potholes as well as the achievements. It must bring together not only those who have succeeded but also those who are struggling. It must foster collaboration across sectors—government, industry, academia, civil society—to address the structural constraints that limit AI’s potential.
The theme of sarvajana hitaya, sarvajana sukhaaya is not a slogan to be displayed on banners; it is a standard to be met. It demands that AI be judged by its contribution to human welfare. It insists that the benefits of AI reach those who are most in need. It requires that the AI race be measured not in teraflops but in lives improved.
Conclusion: The Race India Can Win
The US-China race for AI supremacy will continue, and India cannot ignore it. The EU will continue to play referee, seeking to set global standards. But India’s path need not be defined by this trinity. There is a fourth possibility: a model of AI development focused not on power but on purpose, not on dominance but on welfare, not on the frontier but on the farm.
This is the race that India can—and ought to—win. It is a race measured by the yield of smallholder farms, the health of the poorest, the literacy of children. It is a race in which the starting line is not determined by the number of GPUs or the size of models but by the depth of understanding of local conditions and the strength of commitment to serving the underserved.
Winning this race will not be easy. The potholes are real and numerous. But India has advantages that no other country possesses: a deep bench of technology talent, a world-class digital public infrastructure, a government that has demonstrated commitment to technology-led development, and a scale that makes experimentation meaningful and impact significant.
The AI Impact Summit is an opportunity to articulate this vision, to acknowledge the challenges, and to build the partnerships needed to overcome them. If it succeeds, India will have taken a significant step toward realising its ambition of becoming a vishvaguru in AI impact. If it fails, it will be remembered as another missed opportunity in a long history of missed opportunities.
The choice is India’s. The potholes are mapped. The Ferrari is waiting. The road must be fixed.
Q&A Section
Q1: What is the significance of India’s choice of theme for the AI Impact Summit—sarvajana hitaya, sarvajana sukhaaya—and how does it differ from the focus of previous global AI summits?
A1: The theme—drawn from ancient Sanskrit and meaning “welfare for all, happiness for all”—is a deliberate and pointed departure from the focus of previous global AI summits. The first summit at Bletchley Park in 2023 was dominated by concerns about AI safety: existential risks, the alignment problem, the prospect of machines escaping human control. Subsequent summits have been shaped by the US-China race for technological dominance and the EU’s efforts to play referee. India’s theme shifts the focus from power and safety to purpose and welfare. It asserts that the ultimate measure of AI’s success is not the size of its models, the speed of its chips, or the market capitalisation of its developers, but its contribution to human flourishing—the yield of smallholder farms, the health of the poorest, the literacy of children. This is not merely a rhetorical flourish but a declaration of intent: India aims to lead the world not in the race for ever-more-powerful models but in the design of AI for purpose.
Q2: What are the ten “potholes” that Bhaskar Chakravorti identifies as structural constraints on India’s AI ambitions, and why is the metaphor of the “Ferrari and the dirt road” appropriate?
A2: The ten potholes are: 1) the internet connectivity gap (only 24 per cent of rural households have internet access, and India is near the bottom in digital gender parity); 2) the energy challenge (inadequate infrastructure, grid unreliability, weak transmission capacity); 3) the talent deficit (only one qualified engineer for every ten AI roles); 4) supply chain dependence (over 90 per cent of semiconductor chips imported); 5) regulatory fragmentation (complex, state-varying requirements); 6) the need for “good enough” AI (versus frontier models); 7) infrastructure security (protecting digital public infrastructure); 8) the capital chasm (seed funding exceeds later-stage investment); 9) the actual dirt road (inadequate physical infrastructure); and 10) the need for a new metric (measuring success by human welfare, not model size).
The metaphor of the “Ferrari and the dirt road” is appropriate because it captures the fundamental contradiction of India’s AI ambitions. India is trying to build a world-class AI ecosystem (the Ferrari) on a foundation of inadequate infrastructure, talent shortages, and regulatory fragmentation (the dirt road). The Ferrari cannot perform on a rutted, potholed road; it will break down before it can demonstrate its capabilities. Building the Ferrari while fixing the dirt road is the challenge—and it requires simultaneous investment in advanced technology and basic infrastructure, in frontier research and foundational systems.
Q3: What examples does Chakravorti cite of successful AI applications in India, and what common characteristics do these successes share?
A3: Chakravorti cites several examples: the Kisan E-Mitra chatbot, which handles 20,000 queries daily in 11 regional languages; Telangana’s Saagu Baagu project, which doubled chilli farmers’ earnings while reducing pesticide and fertiliser use; the eSanjeevani telemedicine platform, which has conducted 389 million virtual consultations; Qure.ai‘s TB detection algorithms, which have reached millions; and AI-powered learning platforms like FutureSkills PRIME (1.6 million learners, 41 per cent women, 85 per cent from Tier II/III cities) and DIKSHA (275 million learners, 70 per cent rural penetration).
These successes share several common characteristics. First, they are designed for purpose, not for power—they address specific, well-defined problems rather than chasing frontier capabilities. Second, they leverage India’s digital public infrastructure—Aadhaar, UPI, and other platforms provide the foundational layer. Third, they are built by teams that understand local conditions and constraints—they are not imported solutions but indigenous innovations. Fourth, they are evaluated rigorously—with attention to what works and what doesn’t, enabling continuous improvement. These characteristics offer a blueprint for scaling AI for social good in India and beyond.
Q4: What does Chakravorti mean by the “fourth possibility” beyond the US-China-EU trinity, and what metrics should define success in this alternative approach?
A4: The “fourth possibility” is a model of AI development focused not on power but on purpose. The US-China race is about technological dominance—who can build the largest models, the fastest chips, the most powerful systems. The EU’s role is to set standards and play referee, seeking to shape global norms. India’s alternative is to focus on AI’s contribution to human welfare—on the yield of smallholder farms, the health of the poorest, the literacy of children. This is not a rejection of technological ambition but a redefinition of its purpose.
The metrics that should define success in this approach are fundamentally different from those used in the US-China race. Success should be measured not by teraflops or parameter counts but by: the increase in smallholder farm yields; the reduction in infant mortality and increase in life expectancy among the poorest; the percentage of children who can read this page; the number of previously excluded individuals who gain access to healthcare, education, and financial services. These are the metrics that matter for sarvajana hitaya, sarvajana sukhaaya. They are the metrics by which India’s AI leadership should be judged.
Q5: What warning does Chakravorti offer about the AI Impact Summit, and what does he argue is necessary for India to realise its AI ambitions?
A5: Chakravorti warns that the summit must avoid being “co-opted into a platform for chest-thumping.” The temptation to celebrate India’s achievements, to showcase its successes, to project an image of technological prowess is strong, but celebration without self-criticism is empty. The summit must be a moment of honest reckoning that acknowledges the ten potholes as well as the achievements.
To realise its AI ambitions, India must do several things. First, it must confront the structural constraints—invest in connectivity, energy, talent, supply chains, and infrastructure. Second, it must streamline governance—reduce regulatory fragmentation and uncertainty. Third, it must bridge the capital chasm—develop patient, long-term investment mechanisms for late-stage start-ups. Fourth, it must secure its digital public infrastructure—protect against cyber threats and system failures. Fifth, it must engineer “good enough” AI—versatile, adaptable systems appropriate to local conditions. Sixth, it must measure success by the right metrics—the yield of farms, the health of the poor, the literacy of children.
The summit can catalyse these efforts by fostering collaboration across sectors, sharing lessons from successes and failures, and building the partnerships needed to overcome the potholes. But the real work will begin after the delegates depart. It will be measured not in headlines but in lives improved. That is the race India can—and ought to—win.
