Tech in big picture: Emerging trends in 2025

Frontier technologies have come to occupy centre stage in geopolitical discussions and national strategies in recent years. In 2025, the trepidations associated with establishing dominance in sectors like Artificial Intelligence are apparent through various high-level projects and initiatives announced by governments globally. Major economies have expressed their ambitions to become the next AI superpower or the most favoured destination for AI development and deployment. Moreover, AI scaling laws are being held up with increasingly powerful multimodal frontier AI models released by companies like OpenAI, Anthropic and Meta. Technological innovations and the release of more powerful AI systems are occurring amidst rising geopolitical tensions between the United States and China. In light of the rising interpenetration of geopolitics and frontier technologies, this paper will highlight emerging policy and technology trends to look out for in the coming year.

AI and the rising geopolitical tensions

The year started with AI quickly becoming implicated in geopolitics, with the new Trump administration identifying it as a key driver for an American “golden age”. Shortly after entering the White House, President Trump announced Project Stargate with a budget of US$ 500 billion committed to developing AI infrastructure in the country through collaboration with OpenAI. The announcement for Stargate was accompanied by a stark shift in official rhetoric regarding the US government’s stance on AI development. Scrapping the 2023  Executive Order 14110 on ‘Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence’ highlighted this shift and the President’s insistence on removing regulatory obstacles to US dominance in global AI development. However, the Chinese DeepSeek R1 AI model quickly challenged the aspiration for technological unipolarity, pushing tech stocks off a cliff and causing the biggest stock market drop in US history. The US policy shift and the intensifying AI arms race have prompted the emergence of a divergent policy landscape globally, with the US and the UK abstaining from signing the ‘Pledge for a Trustworthy AI’ in a ‘World of Work’ at the 2025 Paris AI Action Summit co-hosted by France and India. The Summit also witnessed the European Commission President Ursula von der Leyen committing to reducing regulatory red tape and launching a US$206 billion initiative to promote AI development in Europe, diverging from Europe’s historically pro-regulation approach.

However, pro-safety events on the sidelines of the Summit, like Canada and Japan signing the European Commission Framework Convention on AI and China not abstaining from signing the Pledge, signalled that global cooperation on AI safety may not be a lost cause yet. However, the unpredictable trajectory of AI development still presents the risk of frontrunners like the US and China moving away from global dialogues and commitments on AI safety principles. The policy trend towards deregulation will push jurisdictions like the European Union to interrogate its risk-based regulations like the EU AI Act to stay competitive with an increasingly galvanised US and a rapidly advancing Chinese AI sector. Furthermore, the need for scaling national investments in AI will present yet another challenge to the safety principles outlined in existing frameworks like the 2021 United Nations Educational, Scientific and Cultural Organization (UNESCO) Recommendation on the Ethics of AI, the 2023 Bletchley Park Declaration, the 2023 Global Parternship on Artifical Intelligence (GPAI) New Delhi Declaration and the 2024 Seoul Declaration.

AI agents

Ever since the ‘ChatGPT moment’ in November 2022, AI discourse and investment have largely been driven by increasingly powerful large language models (LLMs) with added modalities (video, audio, image generation, data analysis and so forth). However, as NVIDIA Chief Executive Officer (CEO) Jensen Huang highlighted in a recent conference, the next phase of AI development will focus on agentic AI or AI Agents. In addition to having LLM capabilities, AI agents can perform a range of tasks using computers and networks. OpenAI released its agent named ‘Operator’ in January 2025 and its ‘Deep Research’ tool in February, which exceeded the expectations of critics and analysts in autonomously performing operational tasks and conducting complex research. Industry leaders like Sam Altman (CEO, OpenAI) and Marc Benioff (CEO, Salesforce) have predicted AI Agents joining the workforce to “materially change the output of [companies].” As the AI hype cycle stabilises following the crash caused by DeepSeek R1 in January, major AI developers will continue to invest in AI agents with increasingly autonomous operational capabilities. Speaking on the potential of AI agents, Altman suggested that the coming years will require serious discussions on potentially reworking the ‘social contract’ in global economies. If AI agents enter the workforce within the projected timeframe, policymakers and society may face the reality of a dynamic labour market readjustment, necessitating comprehensive international discussions on future jobs.

Humanoid robotics

On the hardware front, 2025 is also shaping up to be a milestone year for quadruped and humanoid robotics due to advances in AI and engineering, with industry experts suggesting that humanoids will be in homes doing chores by 2026. A 2024 report by Goldman Sachs stated that the total addressable market for humanoids is projected to grow by 70 percent annually through 2035. China, the biggest market in the world for industrial robots, is embracing the ‘robot revolution’ by offering sector tax breaks and subsidies to robotics manufacturers. In the West, global AI frontrunners like OpenAI, Meta and Tesla have begun investing in research and development strategies to leverage their AI capabilities for training robots. Breakthroughs in AI and LLMs over recent years have contributed to increased interest in robotics, and 2025 may become an inflection point. In February 2025, Figure AI announced that it would be terminating its year-old partnership with OpenAI due to a  ‘major breakthrough’ that will lead to advancements that “no one has ever seen on a humanoid.” Although no further information has been released regarding the breakthrough, the statement contributed to a surge in investor interest for the three-year-old company, which is now in talks to raise US$1.5 billion in funding at a US$39.5 billion valuation.

Currently, the top 15 market leaders and notable entrants to the robotics sector are exclusively located in North America and China. Existing geo-economic tensions between the two nations will likely expand beyond the software and supply-chain side of tech to encompass robotics. Despite being in a relatively nascent stage of development and relevance compared to consequential economic drivers like advanced chip-making, integrating robots into sectors like manufacturing and the service industry is poised to accelerate international competitiveness in the tech sector. It is still too early to estimate the rise in demand for robotics in the short to medium term. However, government and private-sector interest in robotics is unlikely to fade in the absence of an innovation slump, given the lure of using robots to drive down labour costs and adjust to public spending pressures caused by demographic shifts across labour markets.

Looking ahead

The opening salvo to the year marked a reenergised scramble for global technological superiority and competitiveness. The relentless pace of innovation and growth opportunities in the AI sector with more capable frontier models, AI Agents and tools demonstrating PhD-level competence in certain use cases is set to challenge the adaptability of regulatory regimes. Moreover, robotics is emerging as an attractor of private and public investment, promising solutions for labour shortages and longitudinal demographic shifts. As the year progresses, policymakers globally will need to take a renewed look at public-private partnerships and inter-governmental relationships to maintain a degree of cooperation on matters like responsible innovation and cross-border regulation of AI. Existing frameworks like the 2023 Bletchley Declaration on AI Safety and the 2024 Seoul Declaration, among others, may be stress-tested by US big tech companies emboldened by the deregulatory and America-first policies of the Trump administration. Nevertheless, the intensifying technological rivalry between the US and China in the shadows of a brewing trade war is simultaneously creating new opportunities for international partnerships. The treaties signed on the sidelines of the Paris AI Action Summit, alongside the joint co-chairing of the Summit by France and India, indicate that geopolitical groupings at odds with the US and Chinese spheres of influence can be “not just participants but architects” of the emerging techno-polar world.

Siddharth Yadav is a Fellow in Technology at the Observer Research Foundation, Middle East.