Skip to content

Leadership1 publisher3 min readPublished

Gates Foundation will spend $1bn in two years scaling AI across health, education and agriculture

Two model-vendor partnerships worth $250m came first. Now the world's largest global health philanthropy is putting $1bn over two years into AI for health, education and farming in systems that lost thousands of clinics last year.

The Board Room · Leadership desk

Photograph accompanying Gates Foundation will spend $1bn in two years scaling AI across health, education and agriculture
Photo: abcnews.com

What happened

  • The Gates Foundation said on Tuesday it will give out $1bn over the next two years to scale up AI interventions across healthcare, education and agriculture around the world.
  • A month before Davos, the foundation published a report on collapsing foreign aid infrastructure that documented a rise in child deaths for the first time this century.
  • The foundation says one tool improved the accuracy of medical diagnoses by 16 percent across several Kenyan clinics.

Compiled by The Board RoomSomething wrong?How this is made

Why it matters

  • constraint A software layer improves what a nurse or a teacher decides, and it does not reopen the clinics that closed, so the money buys better decisions inside a thinner system.
  • exposure A diagnostic aid built on one vendor's models inherits that vendor's pricing and availability, and the grantee carries the cost of rebuilding it if the partnership ends.
  • precedent With the State Department already sending $150m to Zipline, other funders now have a large philanthropy's example for routing aid budgets to US technology suppliers.

The pace is about $500 million a year. The foundation has already committed $250 million to two model vendors this year. Bill Gates announced $50 million with OpenAI at Davos, for healthcare in Rwanda; $200 million with Anthropic followed a few months later [2][3][17]. Vox describes those as earlier commitments and leaves open whether they sit inside the new billion or beside it [16]. If they are inside it, a quarter of the two-year programme is already allocated to two American companies [19].

The trade-off is between layers. The foundation's own report, published a month before Davos, documented collapsing foreign aid infrastructure and a rise in child deaths for the first time this century [4]. Vox reports that last year's cuts shuttered thousands of clinics and perhaps hundreds of thousands of classrooms [9]. Software reaches the nurse's decision and stops there; someone still has to administer the vaccine [10]. Mark Suzman, the foundation's chief executive, told Vox there are "very real risks and challenges" around AI, and also said, "This is the most transformative technology of our lifetime" [8][7].

The strongest number in the announcement comes from the funder. One tool improved the accuracy of medical diagnoses by 16 percent across several Kenyan clinics, according to the foundation [5]. That is a sponsor reporting on its own programme, and the foundation published no baseline, sample size or count of clinics. The nearest independent check in the record covers a different technology. The State Department put $150 million behind Zipline drone deliveries last year [12]. An NPR reporter in Rwanda last month found the drones saving lives in places where the drive to a hospital takes hours and a blood transfusion arrives in minutes [13].

For an organisation that depends on these grants, the practical question this quarter is which model family to build on. The foundation is the world's largest global health philanthropy, and the two vendor names in the record are OpenAI and Anthropic [11][2][3]. Proposals get written toward the tools the largest funder has bought. A grantee that builds a diagnostic aid on one vendor's models then owns the integration, the data handling and the cost of rebuilding it if that partnership ends.

The decade case depends on adoption that is still to come. In some countries only a sliver of the population has ever used ChatGPT [14]. Gates wrote on his blog that in poorer countries with enormous health worker shortages and no health systems infrastructure, AI can be a game-changer in expanding access to quality care [6]. Vox reports that the Trump administration shares the do-more-with-less premise, so long as American technology companies benefit [15]. The near-term gain the record supports is narrower than the blog post. The foundation hopes the tools will ease the burden on nurses and teachers now serving far more people than one person should [21], across projects covering crop analysis, diagnosis and student learning [20].

What to watch

  • Whether the foundation states publicly if the OpenAI and Anthropic money is inside the $1bn or additional to it.
  • An independent evaluation of the Kenyan diagnostic result, with a baseline, a sample size and a count of clinics.
  • Whether other global health funders follow the State Department in routing aid budgets to US technology vendors.
Loading claim ledger
Loading source directory links
Loading share composer
Loading topic controls
Loading related stories