AI Race: Nanometers or Gigawatts? Indonesia's Strategic Imperative

The race for leadership in artificial intelligence (AI) development, long framed as a contest over advanced semiconductor manufacturing, is increasingly shifting towards the vast computing power and the energy infrastructure required to support it. This evolving landscape offers significant implications for Indonesia as it aims to adopt AI technologies. For years, the AI competition was viewed as a battle over "nanometers" – the miniaturization of transistors. However, the development and widespread adoption of AI necessitate substantial computing power, which in turn demands the construction of energy-intensive data centers and a stable power supply grid. China's strategic investment in a nationwide network of AI data centers and expanded energy infrastructure exemplifies this strategic shift. Projections indicate that the United States' data center power capacity could exceed 90 GW by 2030, while China's is estimated to reach around 60 GW. AI-related usage is expected to constitute a significant portion of this capacity, with the US and China combined projected to account for roughly 77 percent of global AI-related power capacity, highlighting the immense scale of this undertaking. While the US and its allies traditionally lead in semiconductor manufacturing, China is enhancing its domestic capabilities to reduce reliance on foreign technology. China's strategy of focusing on building the infrastructure to deploy AI widely, rather than solely waiting to close every technological gap in fabrication, can be seen as an attempt to "industrialize" AI ahead of its perfection. This aligns with the historical observation that technological invention and large-scale economic adoption are often separate achievements. In the US, AI infrastructure is largely concentrated among a few hyperscale cloud providers like Amazon and Google. In contrast, China's computing ecosystem involves a broader mix, with state-owned telecom operators and numerous technology firms owning data centers, blurring the lines between state-led and private-led AI infrastructure development. While the US also has initiatives to promote AI data center development, projects face delays due to power availability and grid interconnection challenges. The International Energy Agency projects that global data center electricity consumption could nearly double by 2030, largely driven by AI workloads. While semiconductor superiority remains crucial, the ability to integrate chips, energy, networks, memory, software, cloud platforms, and industrial adoption – the "deployment capacity" – is likely to define the next phase of the AI race. For Indonesia, navigating this new phase of AI development requires not only acquiring cutting-edge semiconductors but also urgently addressing domestic power infrastructure enhancement, data center development, and strategies for effectively deploying AI technologies across its economy and society. Source: The Diplomat Indonesia