TL;DR: The Gates Foundation announced on September 14, 2026, that it plans to spend at least $1 billion over two years to make artificial intelligence useful and reachable for poorer communities.
The commitment frames AI as a choice rather than a fixed outcome. Left to commercial incentives alone, the most powerful systems would reach paying customers first, while clinics, classrooms and small farms in lower-income settings would wait.
According to the Gates Foundation press release, the pledge was issued in Seattle alongside the foundation's 10th annual Goalkeepers Report. That report is titled Make This Matter: AI, Equity, and the Choice We Can't Delay and argues that the period to shape who gains from AI, and how quickly, is brief.
Associated Press reporting carried by the Chronicle of Philanthropy described the same announcement from New York, where foundation leaders said technology firms and political leaders should act promptly so AI reduces social disparities instead of reinforcing them.
What was announced in Seattle
The foundation said it intends to spend at least US$1 billion during the next two years to widen availability of AI and AI-enabled services, opportunities and knowledge. The statement presented the money as support for partners already working to close access gaps and to align tools with the priorities of people most in need.
The announcement came with the launch of the 2026 Goalkeepers Report. That annual publication tracks progress toward the development objectives adopted by the United Nations for improving living standards by 2030, according to the Associated Press account.
Foundation chair Bill Gates wrote in the report that the organization was established partly to address a market shortcoming: groups with the largest needs often have the smallest influence over where innovation and funding flow. He wrote that AI raises an equivalent issue at considerably greater speed.
A second passage quoted in the foundation materials calls for deliberate action by public officials and technology developers, urging that performance be judged not only by value for profitable customers but by value for people with the most to gain.
How the $1 billion would be divided
The foundation outlined an approximate allocation across its priority areas for the two-year period.
About 40% would support education, including personalized tutoring and teaching aids for classrooms in the United States and other countries. Another 40% would support health care, ranging from diagnostic assistance and decision support for frontline workers to tools for maternal and newborn care and research toward new medicines and vaccines.
About 10% would support agriculture, including customized guidance for smallholder farmers based on soil, weather and crop conditions. The remaining 10% would support the digital groundwork for fairer AI, including data collections in languages that current systems do not yet handle well.
The Associated Press account translated those shares into dollar terms: roughly $400 million for education, roughly $400 million for health, roughly $100 million for agriculture and about $100 million for language data and local context. Its examples included software that helps instructors revise lessons around ideas students found difficult, a system that reviews clinician notes for missed symptoms, and a farming application with current advice on weather, pests and fertilizer use.
Why language and local context are central
A central argument in the Goalkeepers Report is that early AI systems reflected a narrow linguistic base. The foundation said more than 90% of the information used to train early large language models came from English-language sources, leaving many communities underrepresented in the foundations of those tools.
The report identified three places for near-term action. First, make tools function in the languages people actually speak. Second, construct tools for the settings where they will be used, drawing on information and specialists from those places and populations, with countries and communities choosing how information is governed and safeguarded. Third, invest in training and availability so local doctors, farmers, teachers and software builders can assess tools and adapt successful approaches.
Chief Executive Mark Suzman told the Associated Press that AI must perform well in local languages, not only English and to some degree Mandarin, if it is to help speed development gains. The foundation said it announced about $100 million to help create data collections in the languages of underserved groups and to reflect their circumstances.
The foundation also said successful tools built with input from intended users could place practical knowledge and guidance in the hands of hundreds of millions of people who missed benefits from earlier waves of technology.
Tech-company partnerships named in the plan
The initiative is not presented as foundation funding alone. The Associated Press reported that the foundation is working with major AI firms and is in further discussions about technical contributions and philanthropic investment.
Google.org and Microsoft's AI for Good Lab helped start an open funding call for projects that strengthen the infrastructure needed for thousands of underrepresented languages across Africa, according to that reporting.
OpenAI committed $50 million alongside the foundation to prepare health workers in Rwandan clinics, in a pilot that Suzman said would soon extend to other African countries. Anthropic is collaborating with the foundation on faster vaccine research and on improving its chatbot's collection of information about local crops, among other activities.
Suzman acknowledged that the foundation totals represent a small share compared with the commercial spending by large technology companies on their own products. He said additional conversations were underway with major AI firms about how their technical capacity could advance the effort.
Timing: aid cuts, safety debate and UN meetings
The pledge arrived during a tense period for development assistance and AI policy. Suzman told the Associated Press he would keep urging Congress and the administration of President Donald Trump to maintain U.S. leadership on foreign assistance, saying, "We haven't given up on the U.S.," while acknowledging difficulties with the U.S. government.
The Associated Press noted that the foundation projected 2025 as the first year this century with a rise in child deaths in its prior report, and that interruptions to a global AIDS relief program had hindered prevention and treatment in dozens of nations. It also cited a recent U.N. assessment that the divide between wealthy and poor countries was continuing to grow.
The report also landed while some industry executives called for slower development and after an ex-Anthropic researcher's departure renewed discussion about whether powerful models might evade human oversight. Gates addressed possible misuse in an August post on his personal website, writing that AI could become either the greatest equalizer or the worst source of injustice, and listing fraud, disinformation, surveillance, cyberattacks and bioterrorism among the risks.
The Goalkeepers Report was published ahead of high-level meetings of the United Nations General Assembly the following week in New York. U.N. Secretary-General António Guterres had sought to bring developing countries into discussions about AI governance.
The foundation closed its announcement with a request for others who want AI to be more practical, reachable and affordable to contribute money, knowledge and public support.
What happens next
The document describes a future spending intention, not completed disbursement. Distribution would occur over two years through external organizations working on foundation priorities, from encouraging AI use among medical staff, growers and educators to creating data collections in additional languages.
Bloomberg noted on September 15, 2026, that Gates tied the next 12 to 18 months to decisions about development and deployment that could determine whether access broadens. The foundation's own report used similar language about a short opportunity to affect who benefits.
For consumers and public-interest observers of applied technology, the practical test will be whether tutoring aids, clinical checks, farming advice and multilingual data sets reach everyday users outside wealthy markets — and whether technology companies lower cost and connectivity barriers alongside philanthropic funding.