A new UN report says AI capacity-building finance is fragmented and uneven. It says nothing about procurement, which Vadillo argues is where the money actually gets stuck.

On September 11, the UN Secretary-General sent the General Assembly a report warning that financing for AI capacity-building in developing countries is fragmented and geographically uneven. It calls for a new Global Fund for AI Capacity-Building, with an initial target of roughly $3 billion over two to four years. The fund would target what the report calls foundational capacity: compute, the energy and connectivity behind it, and the institutions needed to deploy AI, not just debate how to regulate it.

The gap is wide. UNCTAD has found that fewer than a third of developing countries have a national AI strategy, and that 118 countries, most of them developing, are entirely absent from global AI governance discussions. The core threats come from the rapid growth of tech companies. The report states Tech giant values rival entire continential GDPs, and AI automation could impact 40% of jobs, globally. Without clear strategies and infrastructure many countries are at risk of remaining behind in the AI race.

None of this surprises Adriana Vadillo . For 27 years she has worked the fault line the UN just formalized. She has done it as the founder of Global Computing, a Mexico City company that imports and distributes the scientific computing software Mexican universities and government agencies run on, rather than as a policy analyst.

“My mission was to bridge the gap between more developed countries and my own,” Vadillo says, “ensuring that institutions could access leading technology to develop key projects profoundly impacting productivity and research efficiency.”

Vadillo also suggests that in order for the country to be more efficient in terms of accessing and implementing technology, a change in the procurement process is necessary in which the purchase of technology shouldn’t prioritize the least pricing but the best qualified suppliers. I would tell the Secretary-General’s office to add the following perspective to the report.

“In addition to regulating AI, countries in the Global South, like Mexico, need to prioritize their investment in infrastructure to reach efficiency for further application as well as to invest in education to broaden skills and innovation”

Vadillo’s prescription is to award technology contracts to the best-qualified supplier rather than the cheapest, and to pair regulation with investment.

The bottleneck nobody writes policy around

At UN forums in Vienna last July and Bangkok this February, Vadillo sat on panels about bias audits, safety protocols and regulatory frameworks. All of those conversations assume a country already has working AI capacity. The Bangkok session was built around exactly that tension: can developing nations achieve AI sovereignty without owning the infrastructure beneath it? Vadillo is clear, without well-thought out infrastructure, AI cannot thrive.

The UN report makes that case at the level of global finance. Vadillo pushes it one layer down, to procurement. In Mexico, she says, government and university purchasing runs on lowest-price bidding. Institutions set the price and demand extensive legal warranties. Then they take 30 to 90 days to pay once a contract is signed.

The effect is quiet but it compounds. A system built to minimize upfront cost rewards vendors who can survive thin margins and slow payment. Over time, it filters out the vendors best equipped to deliver quality. Small and mid-sized technical distributors either finance the gap themselves or go under trying. That is a market failure, not a supplier’s grievance, and a $3 billion fund won’t fix it on its own. Money routed into institutions that buy on price will still be spent on price.

From one distributorship to nationwide growth

Global Computing was born of two layoffs in eight years. First, Sybase closed its Mexico operations. Then Vadillo built a Mexican subsidiary from scratch for Visual Numerics, a U.S. company. When Visual Numerics was sold to Rogue Wave Software, her role disappeared.

She started over with a single distributorship agreement, for Wolfram Research’s mathematical computation software. She built a market for it across Mexico’s major public and private universities. Other software vendors noticed and asked her to do the same for their products.

What they were buying was her read on the market. She knew what researchers actually needed and how institutions actually purchase. She also had the financial discipline to carry the payment terms that drive other distributors out. In a market built to punish small suppliers, she learned to survive and then to scale. Global Computing now represents 65 vendors .

From the procurement office to the Global stage.

Her fluency in both the technology and the economics is what brought Vadillo into AI policy. She holds a lifetime membership in the AI Forum for Developing Nations ( AIFOD ), an independent non-profit organization that bridges the gap between AI access and sovereignty across developed and developing nations. In this arena Vadillo works to build awareness in Mexico of AI governance and to broker the public-private collaboration the UN report says is missing. That means connecting academic R&D with industry, helping small and medium-sized enterprises adopt AI, and getting the people who will use the technology trained.

The Secretary-General’s report is a map of where the money should go. Vadillo has spent 27 years learning where it gets stuck. If the new fund is meant to build capacity rather than subsidize the status quo, someone will have to explain how Mexico’s institutions actually buy technology. She has been explaining it for decades. The UN is only now catching up.