Trump’s AI Gambit: A New Dawn For India’s Tech Ambitions?

On May 14, 2025, the Trump administration stunned the global technology community by repealing the US AI Diffusion Framework just a day before it was to take effect. This move, framed as a rejection of regulatory overreach, has upended the trajectory of global AI governance and sent ripples through the intricate web of geopolitics, innovation, and economic ambition.

For India, the fallout is both a rare opportunity and a complex challenge. One that could accelerate its AI ambitions but also expose it to the volatility of a deepening US-China tech rivalry.

The AI Diffusion Framework, a hallmark of the Biden administration, was designed to regulate the global flow of advanced AI chips and model weights.

It split countries into a three-tier system: Tier 1 (close allies like the UK and Japan) enjoyed near-unrestricted access; Tier 2 (including India) faced significant limitations; and Tier 3 (China, Russia) was effectively embargoed. The intent was clear: maintain US dominance in AI, prevent adversaries from leapfrogging, and balance national security with economic interests.

However, the framework quickly became a lightning rod for criticism. US industry leaders, notably Nvidia, warned that the rules would stifle American innovation and drive customers toward alternative suppliers. Diplomatically, the system risked alienating key partners by relegating them to “second-tier” status, potentially pushing them closer to China or other rivals.

The compliance burdens—filings, authorisations, and waivers—were seen as excessive and impractical for both regulators and businesses.

The Trump administration’s repeal is not a wholesale retreat from export controls but a recalibration. The US is now doubling down on targeted restrictions, especially against China, while seeking greater flexibility with trusted partners. New, sharper controls are expected, but the era of a rigid, tiered system is over.

The Chip War: Hardware, Software, and the Limits of Control

The timing of the repeal is highly indicative. Reports have emerged detailing Chinese firms circumventing US chip restrictions by rerouting imports through intermediary countries like Malaysia. Furthermore, a recent US congressional investigation accused the Chinese AI company DeepSeek of utilising restricted Nvidia chips to train its advanced R1 model.

DeepSeek’s R1, now trailing US frontier models by mere months, serves as a stark wake-up call: AI advancement is not solely about access to the latest chips, but equally about ingenuity in algorithms and efficiency.

This is the crux of the current chip war. Chinese companies are innovating in software and model architecture, achieving breakthroughs with less compute. The rise of “agentic” AI—models capable of real-time operation on less powerful chips—further undermines the effectiveness of hardware-centric export controls.

The global AI race is thus shifting from pure hardware supremacy to a contest defined by software, architecture, and application.

The Chip Security Act: New Risks, New Realities

As the US pivots from the AI Diffusion Framework, new legislative efforts like the proposed Chip Security Act seek to plug loopholes. This Act would mandate advanced AI chips to incorporate location-tracking and reporting features, with the objective of preventing their diversion to adversaries.

However, these measures introduce their own set of risks: increased costs for chipmakers, potential privacy infringements, and concerns among even friendly nations about “kill switches” and surveillance functionalities embedded within critical infrastructure.

For India, these developments are particularly salient. The idea of relying on chips that could be remotely disabled or monitored by a foreign power sits uneasily with the country’s push for data sovereignty and technological independence. The risk is that, in trying to secure its own interests, the US could inadvertently undermine the autonomy of its partners.

India’s AI Infrastructure Ambitions: Navigating the GPU Crunch

India’s National AI Mission, launched in 2024, is a bold declaration of intent. With plans to develop infrastructure featuring over 10,000 GPUs over the next five years and a targeted 3 GW data centre capacity, India is positioning itself as a serious contender in the global AI landscape.

Reliance Industries’ announcement of a 3 GW mega data centre at Jamnagar, Gujarat, dedicated to AI workloads, further underscores the scale of India’s ambition.

Yet, the US framework—despite its repeal—casts a long shadow. It sets a precedent for how Washington may continue to scrutinize and restrict the flow of advanced AI technology to countries outside its closest circle of allies.

Navigating the New Order: India’s Strategic Choices

A. Securing Non-Volatile Export Use (NVEU) Authorisation: Indian companies seeking to import advanced AI chips will increasingly need to obtain NVEU authorisation from the US.

This will require strict compliance with US security requirements, including clamping down on re-exports and decisively cutting supply chain ties with China. Addressing issues such as illegal chip exports to Russia will also be critical to maintaining credibility and access.

B. Diversifying Compute Infrastructure Beyond Top-End GPUs: Rather than focusing exclusively on top-end GPUs, India can build intermediate-level AI infrastructure using Application-Specific Integrated Circuits (ASICs) and Field-Programmable Gate Arrays (FPGAs).

These alternatives can be tailored for specific, high-priority workloads, reducing dependency on scarce and tightly controlled hardware. This approach not only hedges against supply constraints but also aligns with India’s unique application needs in sectors like healthcare, agriculture, and language processing.

C. Prioritising National Allocation and Strategic Access to GPUs: Given the anticipated limitations on AI chip imports, India must prioritize GPU allocation according to national requirements rather than the demands of individual firms.

One innovative solution could be the creation of regional AI corridors in partnership with allied nations, pooling resources to overcome individual capacity limits and foster cross-border innovation.

D. Learning from DeepSeek: The emergence of China’s DeepSeek model has demonstrated that it is possible to build world-class AI systems with less compute by prioritizing architectural efficiency and algorithmic innovation.

DeepSeek’s ability to achieve near-frontier performance despite hardware constraints is a testament to the potential of creative engineering, and offers valuable lessons for countries like India that face their own limitations in accessing advanced chips.

However, DeepSeek’s rapid rise has also been accompanied by significant controversy. The model has been criticized for its lack of transparency in decision-making—a “black box” problem that complicates accountability and trust.

For India, the DeepSeek experience is instructive but not a blueprint to be copied wholesale. India should indeed invest in developing indigenous, open-source AI models that are optimized for efficiency and can operate with limited hardware resources. But it must also prioritize transparency, ethical safeguards, and robust privacy protections from the outset.

E. Diplomacy as a Strategic Asset: Ultimately, continued access to advanced AI infrastructure for Tier 2 countries like India will depend on maintaining strong diplomatic ties with the US.

The framework—regardless of its current status—is unlikely to change fundamentally under the Trump administration, given its alignment with “America-first” and anti-China policies. Tier 2 applicants are “strongly encouraged” to secure government-to-government assurances to obtain NVEU status, making diplomatic engagement as crucial as technical capability.

The Broader Context: Strategic Autonomy and Global AI Governance

India’s placement in Tier 2 under the now-defunct framework was both a practical hurdle and a symbolic slight. While the immediate cap on AI chip imports was not a binding constraint for India’s current demand, the second-order effects were more troubling.

The framework risked undermining India’s strategic autonomy by making its access to critical technology contingent on external approval. This could have led to a scenario where Tier 1 countries became the preferred destination for cutting-edge AI development, sidelining India’s ambitions.

The repeal, therefore, is a positive development from the perspective of strategic autonomy. It removes an immediate barrier to technological advancement and signals that India’s push for indigenous innovation and self-reliant infrastructure is both necessary and timely. The IndiaAI Mission’s focus on building a high-end computing facility and developing indigenous GPUs by 2029 is a clear step in this direction.

India’s approach to AI development is also distinctive in its emphasis on inclusivity and open standards. The Bhashini initiative, which develops open-source natural language processing models for 22 Indian languages, exemplifies how India is tailoring AI solutions to local needs and democratizing access. This strategy not only serves India’s diverse population but also positions it as a model for other developing nations seeking to harness AI for broad-based development.

If India can rise to this challenge, it will not only secure its own digital future, but also help steer the world toward a more open, cooperative, and equitable AI era. The opportunity is real, but so is the responsibility.