As prepared for delivery
Excellencies, distinguished colleagues, ladies and gentlemen,
Welcome to the World Investment Forum’s track on investing in AI.
We are living through what may become the largest CAPEX cycle in modern history. The world will spend close to $800 billion on AI infrastructure this year. Cumulative spending will approach $8 trillion by the end of the decade. Announced greenfield investment in AI infrastructure grew at 47% a year between 2020 and 2025 – the fastest of any sector we track. AI infrastructure alone now accounts for three fifths of all greenfield investment in strategic sectors worldwide.
The geography of this boom is narrow. A handful of economies host the vast majority of AI infrastructure investment. 100 companies account for 40% of global R&D. And the barriers compound: a data centre needs reliable power, water for cooling, high-bandwidth connectivity, technical skills and predictable data governance – all at once, in the same place. Miss one, and capital goes elsewhere.
But from an investment, and even from a development point of view, it’d be a mistake to focus only on the concentration of AI production.
Every major technology was produced in a few places long before it changed lives everywhere. The steam engine took 80 years to be widely adopted in lower-income economies. Electricity took 40. The internet, 20. But within 6 months of ChatGPT’s launch, middle-income countries already accounted for half its global traffic. Each generation of technology diffuses faster. AI is diffusing faster than any of them.
And it is in the diffusion of use that most value will be generated. The digital economy offers a precedent. Most of the wealth the internet created came from how it was used, far more than from who built the cables or hosted the servers. AI is likely to follow a similar pattern: most economic value will come from use – in agriculture, in health systems, in logistics, in public services. Getting AI use right requires its own investments – in skills, connectivity and the capacity to adopt.
The deciding variable is trust. Countries and firms invest in building their own AI capacity when they do not trust the alternatives – when they worry about the treatment of their languages, their data, their sovereignty. When they trust the process, they use what exists and focus their investment on adoption, adaptation and the skills to make AI productive. This track will explore how partnerships can strengthen that trust.
UN Trade and Development is proud to be working with the International Telecommunication Union on this track. Our strengths are complementary: UNCTAD’s work on investment and development, ITU’s leadership in digital standards and infrastructure. The Digital Infrastructure Investment Catalyser we launched together in Sevilla last year is already connecting public and private partners with investment solutions on the ground.
I also warmly welcome the World Intellectual Property Organization, represented by Assistant Secretary-General Marco Alemán, who will deliver a spotlight address this morning. Intellectual property, innovation and the responsible use of data are central to the AI economy we are discussing.
And I welcome the investors, policymakers and technology leaders joining us across this track. The sessions ahead – on financing, skills, governance and innovation – will benefit from the range of experience in this room.
Excellencies,
Investing in the future is the theme of this Forum. There is no future industry, no future service, no future enterprise that AI will not touch. This track is about making sure the investment matches the ambition.
I thank you.