The impact of India-EU FTA for AI and semiconductor tech Premium

Explore the transformative India-EU FTA, advancing AI and semiconductor collaboration through joint R&D and regulatory alignment.

In a milestone, India and the European Union (EU) have hailed the conclusion of the India-EU Free Trade Agreement (FTA) while launching a ‘Comprehensive Strategic Agenda’ for 2030. Among other measures, the pact moves beyond supply chains to operationalise joint R&D in advanced semiconductor “heterogeneous integration” and chip design. It also formally links the European AI Office with India’s National AI Mission to jointly develop safe, human-centric AI and take advantage of what the agreement called India’s vast “multilingual datasets” and Europe’s research infrastructure to secure strategic autonomy in critical technologies.

The AI and semiconductor aspects represent the maturation of three diplomatic phases. The first phase began with the ‘India-EU Strategic Partnership: A Roadmap to 2025’. At this stage, discussions on “technology” were largely confined to general matters of cybersecurity, 5G networks, and data protection and there was no specific mechanism for a more concrete collaboration between the bloc and India on hardware such as semiconductors, or specific AI model development.

In the second phase, around 2022, Prime Minister Narendra Modi and EU President Ursula von der Leyen launched the India-EU Trade and Technology Council. This body in turn created the Working Group 1 on Strategic Technologies, which moved the relationship from diplomacy to involving technical experts. In fact the new deal has credited this group with managing its “technology and innovation” aspect.

In the third and final phase, in 2023, India and EU signed the Semiconductor Memorandum of Understanding (MoU). This MoU was primarily defensive in the sense that it focused on improving the resilience of supply chains and on providing early warnings of shortages. Between 2023 and the present deal, the MoU evolved into a more ‘offensive’ partnership, so to speak, with a focus on creating new technologies, including through designing and prototyping, in addition to monitoring existing supply chains.

Perhaps the most significant technical detail in the document is “heterogeneous integration” because it recognises that India is years away from building cutting-edge logic fabrication facilities, for example, those that manufacture chips with nodes of 2-3 nm, and thus pivots to “advanced packaging”. Heterogenous integration involves stacking different types of chips, such as logic, memory, and sensors, into a single package. Such combinations are critical for AI. Contemporary AI accelerators like Nvidia’s Graphics Processing Units (GPUs) bank more on how memory is packaged next to the processor instead of only the raw transistor size. By targetting this, the EU and India are trying to capture the value-added segment of the supply chain that is also less capital-intensive than a fabrication facility but which is just as critical for performance.

The agenda also explicitly connects semiconductor manufacturing to AI utility as the text says the deal will focus “on design and prototyping for AI applications”, thus creating a vertical market. While in 2023 the focus was on “making chips for cars”, following a crisis in their availability, the new goal is to make chips that AI models require.