Comrades Editorial · 01 Sep 2026 (PDF)
In this edition
- Breaking the AI Hegemony: India’s Path to AI Sovereignty Science & Technology
Science & Technology
Breaking the AI Hegemony: India’s Path to AI Sovereignty
Source This editorial is based on “The battle for AI hegemony” published in the Financial Express on 01/09/2026. The article highlights that US-China tech rivalry threatens emerging markets. It urges India to build indigenous semiconductors, sovereign compute infrastructure, and localized foundation models to secure digital sovereignty and counter global hegemonic dominance. The US-China contest for AI hegemony spans the entire value chain (from chips and energy to foundation models and governance), prompting emerging economies like India to fiercely pursue AI sovereignty amidst widening technological divides. While the US dominates private capital and frontier models, China leverages narrowing performance gaps, open-source adoption, and its new World Artificial Intelligence Cooperation Organization (WAICO) to court the Global South. For India, balancing strategic pragmatism against upstream foreign dependencies remains vital to safeguarding its domestic ecosystem. What is AI Hegemony?- About:
- AI hegemony refers to the concentration of power over artificial intelligence, its data, computing infrastructure, semiconductor chips, foundational models, talent, platforms and governance standards, in the hands of a few technologically advanced countries and large corporations.
- It enables dominant actors not merely to lead technological development but also to shape the economic opportunities, security choices, public discourse, cultural values and political autonomy of other societies.
- Thus, AI hegemony represents a new form of power operating through algorithms and technological dependence, rather than direct territorial control.
- Major Dimensions:
- Geopolitical Hegemony:
- Leadership in AI has become central to strategic competition, particularly between the United States and China.
- Countries possessing superior AI capabilities can influence global alliances, technology transfers, sanctions, trade rules and security arrangements.
- Corporate Concentration:
- A small number of technology companies control advanced foundation models, cloud platforms and AI distribution channels.
- Their resources allow them to set prices, conditions of access and acceptable-use policies for businesses, researchers and governments.
- Computational Hegemony: Training advanced AI models requires large computing capacity and specialised accelerators. Countries and firms controlling high-end chips, data centres and cloud services therefore act as technological gatekeepers.
- Data Colonialism: Global platforms extract enormous quantities of behavioural and linguistic data from developing societies, convert them into commercially valuable AI systems and retain most of the resulting economic benefits.
- Surveillance Dominance: Facial recognition, predictive policing and mass-data analysis allow states and corporations to monitor individuals at an unprecedented scale. When these systems lack accountability, technological capacity may translate into authoritarian control.
- Military Hegemony: AI enhances autonomous weapons, cyberwarfare, battlefield surveillance, target identification and intelligence analysis. Countries possessing superior systems can gain asymmetric military advantages and accelerate an international AI arms race.
- Economic Hegemony: AI-leading economies capture high-value activities such as chip design, model development and cloud services. Developing countries may remain dependent consumers or suppliers of raw data and low-paid annotation labour, reinforcing existing global inequalities.
- Geopolitical Hegemony:
- AI Leadership versus AI Hegemony:
- AI leadership turns hegemonic when technological superiority is weaponized to restrict access, extract unfair benefits, impose non-inclusive rules, manipulate behavior, and create permanent technological dependence that concentrates gains while offloading risks onto weaker societies.
- Therefore, technological advancement itself is not hegemonic; unaccountable and exclusionary control over it is.
- AI hegemony is not merely technological superiority; it is the ability to shape how other societies produce knowledge, make decisions and imagine their future.
- Concentration of Financial Resources:
- Developing frontier AI models requires billions of dollars for chips, data centres, electricity, research and skilled personnel. Only wealthy governments and technology corporations can sustain investment at this scale.
- According to the Stanford AI Index 2026, private AI investment in the United States reached approximately USD 285 billion in 2025.
- The United States also recorded 1,953 newly funded AI companies in 2025, more than ten times the number in the next-ranked country.
- Unequal Access to Advanced Computing Power:
- Frontier AI depends upon massive clusters of specialised processors. Countries without affordable access to such computational infrastructure cannot train competitive foundation models and must depend on foreign cloud platforms.
- Global AI computing capacity has reportedly increased around 30-fold since 2021, illustrating the enormous scale of the ongoing infrastructure race.
- The World Bank identifies computing capacity as one of the four foundations, along with connectivity, data and skills, required to participate meaningfully in the AI economy.
- Semiconductor Supply-Chain Concentration:
- AI development depends on advanced GPUs, memory chips and semiconductor-manufacturing equipment, whose design and production are concentrated among a limited number of firms and locations.
- Companies such as NVIDIA (USA) dominate advanced AI accelerators, while cutting-edge chip manufacturing relies heavily on foundries such as TSMC (Taiwan).
- Advanced lithography equipment is supplied predominantly by a few companies, particularly ASML (the Netherlands).
- This enables technologically powerful States to use export controls and supply restrictions as instruments of geopolitical influence.
- Dominance of Cloud and Digital Infrastructure:
- A few corporations, particularly Amazon Web Services, Microsoft Azure and Google Cloud, provide much of the computing infrastructure used to train, host and distribute AI models.
- Major model developers have formed close investment and infrastructure relationships with cloud providers, linking access to capital, computing and market distribution that smaller developers generally cannot afford.
- Vertical Integration by Big Technology Companies:
- Leading technology corporations increasingly control several stages of the AI ecosystem simultaneously like Chips → Cloud infrastructure → Foundation models → Applications → Consumer platforms.
- Such vertical integration makes it difficult for independent firms to compete and allows incumbents to favour their own products.
- Concentration of Research, Patents and Intellectual Property:
- Advanced AI research is dominated by a relatively small group of universities, corporations and countries possessing laboratories, funding and computing resources.
- UNCTAD’s Technology and Innovation Report 2025 noted that just 100 companies accounted for approximately 40% of global business-funded research and development.
- Patents, trade secrets, proprietary datasets and closed model weights prevent technological diffusion and force less-developed economies to pay licensing fees or accept external dependence.
- State Support and Geopolitical Competition:
- AI leadership is no longer driven only by markets; it is actively supported by national industrial policies, defence funding, public procurement and diplomatic strategy.
- The US’s CHIPS and Science Act, 2022 authorised approximately USD 52.7 billion for US semiconductor manufacturing, research and workforce development.
- According to the Stanford 2026 AI Index Report, Chinese government guidance funds invested an estimated USD 184 billion in AI firms between 2000 and 2023.
- Enormous Commercial Value of AI:
- The prospect of exceptional profits encourages corporations and governments to secure control over AI resources. UNCTAD projects that the global AI market could grow from approximately USD 189 billion in 2023 to USD 4.8 trillion by 2033.
- Such economic importance makes control over AI comparable to earlier struggles for dominance over oil, industrial machinery and global finance.
- Dependence on Imported Advanced Semiconductors:
- India does not yet manufacture the most advanced GPUs and AI accelerators required to train frontier models. It remains dependent on foreign chip designers such as NVIDIA and AMD, overseas foundries such as TSMC, and foreign semiconductor equipment.
- Export restrictions, supply disruptions or geopolitical conflicts could affect India’s access to essential computing resources.
- Thus, India’s AI autonomy may ultimately be constrained by decisions taken in Washington, Taipei or other technology centres.
- Inadequate Frontier-Scale Computing Capacity:
- Indian start-ups, universities and public research institutions possess considerably less high-performance computing capacity than major American and Chinese corporations.
- The World Bank reported that high-income countries hosted 77% of global co-location data-centre capacity in June 2025, compared with only 5% in lower-middle-income countries.
- While India is said to account for 20% of global data generation, it has only 3% of global data centre capacity.
- India could consequently remain dependent on externally controlled computing infrastructure even while its domestic use of AI expands.
- Dependence on Foreign Cloud Providers:
- In the pure infrastructure space, Microsoft Azure, AWS and Google Cloud have a combined 87% market share.
- Even when their data centres are physically located in India, the underlying cloud architecture, software and intellectual property remain controlled by foreign companies.
- Limited Ownership of Frontier Foundation Models:
- India has promising initiatives such as Sarvam, BharatGen, AI4Bharat and Krutrim, but globally dominant general-purpose models remain concentrated in the United States and China.
- According to the World Bank, high-income countries account for 87% of notable AI models and 86% of AI start-ups, despite comprising only around 17% of the global population.
- India may therefore become a developer of downstream applications without controlling the foundational intelligence on which they operate.
- Weak Financial Position of Domestic Start-ups:
- Indian AI companies compete with foreign firms possessing vastly greater access to venture capital and corporate investment. The World Bank found that high-income countries received approximately 91% of global AI venture-capital funding.
- Capital shortages may compel Indian firms to rely on foreign investors, cloud providers and licensed models or to transfer intellectual property abroad.
- Exposure of Public Services to Proprietary AI:
- Indian authorities are increasingly exploring AI for welfare delivery, policing, healthcare, taxation, agriculture and judicial administration.
- If proprietary foreign models (owned exclusively by a specific company) are used, authorities may be unable to fully examine their training data or decision-making processes.
- Vendor lock-in could make essential public services dependent on the commercial and geopolitical decisions of private corporations.
- Limited Role in Global AI Rule-Making:
- AI standards concerning safety, copyright, model testing, privacy and interoperability are largely shaped by advanced economies and powerful corporations. While India has joined initiatives such as Pax Silica, its limited role in shaping the underlying standards and rules remains a concern.
- Despite high-profile international gatherings like the Bletchley Park Summit 2023, the Paris AI Action Summit 2025, and the recent India AI Impact Summit 2026, the world still lacks a single, independent global regulatory body with binding enforcement powers.
- Cultural and Epistemic Dependence:
- Foreign models trained primarily on Western data may reproduce the cultural assumptions, historical interpretations and ethical priorities of their developers.
- Indian traditions, indigenous knowledge and local sources may be ignored or treated as less authoritative.
- AI hegemony may consequently weaken India’s epistemic sovereignty, or its capacity to interpret social realities through locally grounded knowledge.
- IndiaAI Mission:
- IndiaAI Mission is India’s principal programme for building sovereign and inclusive AI capabilities.
- Its seven pillars cover compute, indigenous models, datasets, applications, skills, start-up finance, and safe and trusted AI.
- National AI Compute Infrastructure:
- India is creating affordable computing capacity for start-ups, researchers, universities and public institutions.
- The IndiaAI Compute Portal provides access to over 38,000 GPUs and 1,050 TPUs at subsidised rates.
- A secure cluster of 3,000 next-generation GPUs is being developed for sovereign and strategic applications.
- Indigenous Foundation Models:
- India is financing models trained on domestic datasets and adapted to Indian conditions.
- For instance, Sarvam AI has developed open-source large language models, including models like Sarvam 30B, Sarvam 105B and Soket AI is developing open-source 120-billion-parameter models.
- Also, India is building multilingual AI infrastructure to protect linguistic and cultural sovereignty.
- BharatGen, launched in 2025, is a government-funded multimodal model covering 22 Indian languages.
- BHASHINI supports many Indian languages and integrates key AI models for translation and speech services.
- AIKosh- India-Centric Data Platform:
- AIKosh provides domestic datasets, models and tools for researchers and start-ups.
- It supports locally relevant AI applications in agriculture, healthcare, governance, language and climate reducing dependence on foreign datasets.
- India Semiconductor Mission:
- By December 2025, 10 projects involving about ₹1.60 lakh crore across 6 States had been approved to strengthen domestic chip design, fabrication, assembly and packaging.
- Semiconductor Mission 2.0, announced in Budget 2026–27, focuses on research, skills and advanced chip technologies.
- AI Skills and Research Ecosystem: IndiaAI FutureSkills supports specialised AI education, fellowships and laboratories. A network of around 570 AI Data Labs is being developed, particularly in Tier-2 and Tier-3 cities.
- AI Diplomacy and Global Rule-Shaping: India is using platforms such as the GPAI, G20 and India–AI Impact Summit 2026 to promote inclusive global AI governance. Such engagement seeks to make India an AI rule-shaper, rather than a passive rule-taker.
- Full-Stack Ecosystem & Strategic Procurement:
- India must develop end-to-end capabilities spanning chip design, fabrication, cloud infrastructure, datasets, and foundation models, avoiding mere application-layer dependence.
- Public procurement must mandate rigorous domestic-value criteria, prioritizing indigenous research and local IP to prevent the superficial repackaging of foreign models. Government contracts must enforce open APIs, modular architectures, data portability, and secure exit clauses.
- Compute Infrastructure, Reserves & The Compute-UPI:
- An interoperable, vendor-neutral network (backed by public-private compute pooling) will allow universities, startups, and MSMEs to buy and shift workloads in real-time, supported by public-interest compute vouchers.
- Analogous to petroleum reserves, India must maintain dedicated GPU capacity for defense, health emergencies, and disaster response, verified through periodic AI resilience stress tests.
- Accelerated Indigenous Hardware & Supply-Chain Security:
- Long-term financing must back domestic AI accelerators, edge chips, and open architectures like RISC-V, with state departments acting as assured first-buyers.
- High-risk systems must disclose detailed component, chip, and data-source manifests, while critical foreign or domestic vendors must deposit model documentation and source components into secure escrow institutions to protect against sudden withdrawal or sanctions.
- Collaborative Foundation-Model Consortiums & Algorithmic Diversity:
- IITs, IISc, public laboratories, and industry should pool research talent and computing power to build open-weight, low-connectivity-optimized models.
- Independent, highly secure models must govern sensitive fields like defense and law. Furthermore, an Algorithmic Diversity Mandate should prevent critical sectors (banking, healthcare, welfare) from relying on a single model, utilizing cross-verification to stop systemic failures.
- Sovereign Data Commons, Data Cooperatives & Reciprocal Access:
- Agricultural, judicial, and administrative data should be curated through legal cooperatives and an Indic Intelligence Vault that preserves civilisational knowledge while granting communities transparent benefit-sharing and control.
- Foreign firms utilizing Indian data must provide reciprocal benefits (technology transfer, local compute investment). Additionally, sensitive data and state models must operate exclusively within Sovereign Inference Zones on certified domestic infrastructure.
- Patient Capital, Talent Retention & Independent Auditing:
- A professionally managed fund combining public and private capital should provide long-term, milestone-based financing for deep-tech startups while preventing hostile foreign acquisitions of local IP.
- Dedicated laboratories must conduct mandatory algorithmic impact assessments for high-risk applications in policing, healthcare, and credit, ensuring citizen rights to explanations and human review.
- AI Non-Alignment Doctrine: Cooperating with multiple global partners (US, EU, Japan, South Korea) while depending on none, India should build coalitions with the Global South (via the UN, G20, and GPAI) to export affordable, multilingual AI public goods and advocate for equitable global standards.
- As AI becomes a foundational pillar of national power, India must avoid dependence and isolation to achieve true AI sovereignty building upon M.A.N.A.V. vision for artificial intelligence. Countering multi-dimensional AI hegemony requires more than just raw compute or GPUs; India needs a resilient full-stack ecosystem spanning indigenous silicon, diverse data, and trustworthy models tailored to its democratic values and languages. By pursuing strategic non-alignment, diversified partnerships, and proactive rule-shaping, India can transition from a passive consumer market into an autonomous global rule-shaper.