An Arkus Innovation Studios publicationPeople & possibilities
The Arkus Signal
Grants and funding opportunities worth pursuing
FUNDING EDITIONMONDAY, 14 SEPTEMBER 2026ISSUE 01 / PRINT EDITION
Grants & research funding
A seat at the table. A foot in the door.
Humanity AI wants communities to shape the technology changing their lives. Citi Foundation wants low-income young people ready for work. Two AI grants, with different ideas of what progress requires.
A say in the decisions, not just a seat beside the technology. AI-generated illustration; not a photograph of this project.
Humanity AI puts community decisions at the center. Citi Foundation backs nonprofits preparing young people for work. Both are new calls, but a proposal that fits one may miss the other.
Inside: eight funding stories, including research for small companies in São Paulo, support for small patient organizations, and a challenge opening later this month.
Follow the money
Two very different bets on AI.
Choose a call to see what it is trying to change.
Humanity AI
Let people shape what happens next.
Communities help guide how AI is built, governed or challenged. A project mainly teaching people to use AI tools will miss this call’s purpose.
Our illustration of each call’s purpose. The applicant rules differ too: Humanity AI permits an eligible US nonprofit to handle a grant for another project as its fiscal sponsor; Citi does not. Read the full applicant rules before applying.Explore the opportunities ↓
THE ARKUS SIGNAL / FUNDINGTHE OPPORTUNITIES / 02
NEW / US + collaborators
A seat at the table is only the start
1 min read ·
Humanity AI offers $75,000 to $1 million for US communities to help shape, govern or challenge AI. Teaching people to use the tools is not enough for this call.
Humanity AI · Communities Leading on AI
A say in the decisions, not just a seat beside the technology. AI-generated illustration; not a photograph of this project.
A community can be taught to use an AI tool without ever being asked what it wants from it. Humanity AI is funding the missing conversation, and the work that follows. Its call asks communities in education, labor, and arts and culture to help direct projects that shape AI’s place in their lives.
Projects mainly about learning to use AI tools, adopting them or doing research alone are excluded. Reviewers will examine the idea, the community’s leadership, whether the work is achievable, its potential impact and the team’s experience.
The work must center US communities. Eligible US nonprofits and qualifying Tribal governments can lead. A US fiscal sponsor can handle a grant for another project; foreign and for-profit organizations may collaborate but cannot lead. Applications close October 21 at 4:59 p.m. PT. An applicant webinar is scheduled for September 14.
Before you start
US nonprofits, qualifying Tribal governments, or US fiscal sponsors that handle grants for other projects. Work must center US communities. Foreign and for-profit groups can collaborate, but cannot lead. The call covers three ways for communities to shape AI. Work mainly about learning to use tools, adopting them or doing research alone is excluded.
The finer details
A Tribal government can lead if it is treated as a State under section 7871 of the US tax code. Reviewers consider the idea, community leadership, whether the plan is achievable, its possible impact and the team’s experience. Existing 2026 direct grantees cannot apply alone.
Citi Foundation offers registered nonprofits $500,000 over two years to prepare low-income young people for work. Brazil is among the eligible markets.
Citi Foundation · Preparing Youth for AI
Learning matters when it connects to a life beyond the classroom. AI-generated illustration; not a photograph of this project.
An AI skill has more value when a young person can put it to work. That is the practical concern behind Citi Foundation’s fourth Global Innovation Challenge. It focuses mainly on low-income people aged 15 to 24, covering AI and human skills, digital access, and stronger nonprofit programs that help people into employment.
Brazil is included. US applicants must serve one of the listed locations. The organization applying must be a registered nonprofit in its own right; another nonprofit cannot apply on its behalf as a fiscal sponsor.
The first short application, called a letter of inquiry, is due October 6 at noon ET and must be in English. Later, applicants need an independent financial audit that meets Citi’s rules. First-round reviewers will ask whether the project fits the call, can be carried out and could make a meaningful difference.
Before you start
Registered nonprofits working with low-income young people, mainly ages 15 to 24. Brazil is included. US applicants must serve one of the listed locations. The nonprofit must apply itself; fiscal sponsors cannot apply for another group. The first short application, called a letter of inquiry, must be in English. A qualifying independent financial audit is required later.
The finer details
The first round asks how well the project matches the call, whether it is achievable and what difference it could make. Non-US grants are handled through CAF America. The official FAQ explains the audit requirement and the application stages.
FAPESP offers up to R$500,000 for as much as 12 months of small-company research. The question must be technical; the research base must be in São Paulo.
FAPESP · PIPE Indústria do Futuro
The useful moment in research: a question the team cannot yet answer. AI-generated illustration; not a photograph of this project.
A promising idea still leaves a company with work to do: find out whether it can succeed. FAPESP’s PIPE Indústria do Futuro call offers up to R$500,000 for as much as 12 months of innovative research. It asks for a genuine technical challenge and the team and facilities to investigate it.
That work needs a home in São Paulo. The company must be Brazilian, have no more than 250 employees, and maintain a real headquarters or research unit in the state. A virtual address does not qualify. The researcher in charge must be a company partner and live in São Paulo. A company still being formed can apply under specific conditions.
The first proposal is due October 8, with no cutoff time stated in the call. Applicants invited to continue then have until December 9 for the full proposal.
Before you start
Brazilian for-profit companies with no more than 250 employees and a real headquarters or research unit in São Paulo. The lead researcher must be a company partner and live in the state. The project needs a genuine technical research challenge, a suitable team and facilities. Companies still being formed can apply under specific conditions.
The finer details
Virtual addresses do not qualify. Applicants whose first proposal is approved move on to a full proposal due December 9. The call’s company-formation rules explain the additional conditions for a business that does not yet exist.
Norcliffe accepts applications year-round from eligible Puget Sound nonprofits. Its grants can support the everyday work of running an organization as well as a specific project.
The Norcliffe Foundation · Standard Grants
The everyday work of keeping a community organization running. AI-generated illustration; not a photograph of this project.
Norcliffe gives Puget Sound nonprofits room to explain the work that needs doing. Applications are open year-round, and support can cover day-to-day running costs, programs and capital projects. The foundation’s interests stretch from arts and education to health, research and basic needs. A useful organization does not have to invent a new program every time it needs support.
The amounts call for some perspective. Norcliffe says nearly 60% of its grants are $15,000 or less. The standard application handles requests up to $50,000; larger requests use a separate form that requires a conversation with the foundation. Every application needs a fundraising plan. Norcliffe generally wants to join other funders, and says it is rarely a project’s largest donor.
The route is open to registered US charitable nonprofits working in the foundation’s eleven-county area, including King County, and eligible projects with a nonprofit handling their grant as fiscal sponsor. Start with the eligibility quiz, then the full online application. There is no preliminary letter. Allow up to six months for a decision, and check the waiting period if the organization has applied before.
Before you start
US 501(c)(3) charitable nonprofits, or projects with a formal fiscal sponsor agreement. Work must primarily benefit Norcliffe’s eleven Puget Sound counties. Complete the eligibility quiz. After a denial, wait at least 12 months; after a standard grant, 24 months from final payment; after a special grant, four years. Requests above $50,000 require a separate form and foundation contact.
The finer details
Counties: Island, Jefferson, King, Kitsap, Mason, Pierce, San Juan, Skagit, Snohomish, Thurston and Whatcom. No debt reduction or deficit financing. A fundraising plan is required. The guidelines advise allowing up to six months for review; the FAQ says a decision typically takes four. Confirm which waiting rule applies to older grants.
Cambia offers $30,000 to $50,000 for school health programs shaped by students and families in four Western states and Tribal communities. The school connection is essential.
Cambia Health Foundation · Resilient Children and Families 2026
School health involves the people students already turn to. AI-generated illustration; not a photograph of this project.
For a student, getting help can mean finding a ride, missing class and fitting an adult’s calendar. Cambia wants more care available where young people already spend their day. Its 2026 Resilient Children and Families call backs school health services, with particular attention to mental, social and emotional well-being.
The foundation wants students and parents to help shape the work. Peer support, mentoring between older and younger students, counseling and links to community care are among its examples. Established programs get priority. A new idea can qualify, but it needs strong school support and a believable plan to keep going. A school’s name on a proposal is only the beginning of that relationship.
Applicants must be tuition-free public schools or eligible nonprofit partners serving Oregon, Washington, Idaho, Utah or Tribal Nations in those states. The call permits eligible organizations to apply on another group’s behalf. Grants are expected to run for up to two years. One application is allowed per organization or collaboration, and overhead charged as indirect costs is capped at 10%. The full online application is due September 21 at 5 p.m. Pacific.
Before you start
Tuition-free public elementary through high schools, or eligible US charitable nonprofit partners, serving Oregon, Washington, Idaho, Utah or Tribal Nations within those states. The school connection is mandatory. One application per organization or collaboration. Established services receive priority; new work needs strong school support and a viable way to continue. Indirect costs are capped at 10%.
The finer details
The eligible-entity list includes active US 501(c)(3) charities, accredited schools, public agencies and federally recognized Tribal entities; the school connection still applies. Eligible fiscal agents may apply for charitable projects. Planned awards: 7-10. Decisions are expected the week of November 2; work is expected to start December 2026 or January 2027. Capital campaigns, sponsorships, existing deficits and retroactive expenses are excluded.
Gates and Temasek are backing simpler ways to read disease-causing organisms’ genetic material. Grants reach $800,000, with a target of roughly $1 to $10 per sample.
Grand Challenges · Low-Cost Pathogen Sequencing
A practical goal: bring useful sequencing tools within reach of more laboratories. AI-generated illustration; not a photograph of this project.
A laboratory can own a sequencing machine and still find routine use too expensive. This Grand Challenges call asks researchers to lower the cost of reading a disease-causing organism’s genetic material, while keeping the results useful for public health. The target is roughly $1 to $10 per sample. The ambition is easier access to reliable information in low- and middle-income countries.
The Gates and Temasek foundations are funding the call, managed by the United Nations Foundation. Three grant levels match the work’s maturity: up to $300,000 for early testing, $600,000 for further development and $800,000 for advanced work. The first two allow up to 24 months; the third allows up to 36. About 15 awards are expected. Better methods are the point: buying supplies more cheaply, on its own, will not qualify.
Eligible organizations worldwide can apply, including companies, nonprofits and research institutions in the US and Brazil. Individuals cannot. The initial submission allows three pages for the proposal and one for the budget. It closes September 29 at 11:30 a.m. PDT. Read the access and disclosure rules first: the funder explicitly says to leave proprietary information out.
Before you start
Research institutes, nonprofits, companies, international organizations, government agencies and academic institutions worldwide. US and Brazilian entities may apply, subject to the published eligibility and sanctions rules. The work must improve sequencing methods for public-health use. Individuals and organizations treated as individuals for US tax purposes are ineligible. Only one lead application per investigator. Apply the funder’s access rules and exclude proprietary information.
The finer details
The funder requires outputs to be accessible to intended beneficiaries. Partnerships are encouraged, not mandatory. Grant ceilings include permitted indirect costs. The main call asks for an 18-month work plan and first evidence within 12 months, while award terms allow up to 24 or 36 months by tier; clarify the planned milestones. No requirement for a particular degree is published, but the lead must have relevant expertise and organizational support. The rules also exclude entities with revenue from making tobacco or specified controversial weapons.
MIT Solve’s Economic Prosperity challenge opens September 22. The preview names jobs, housing, financial security and fair access to digital life; full terms still need checking.
MIT Solve · Economic Prosperity 2027
Technology meets an everyday need: access to work, housing and financial stability. AI-generated illustration; not a photograph of this project.
Work, a home and financial security sit at the center of MIT Solve’s next Economic Prosperity challenge. Its preview points to technology-based solutions, including fair access to digital life. The 2027 challenge opens September 22 at 9 a.m. EDT.
The published closing date is November 2 at noon EST. The fuller picture comes when applications open: who can apply and what an award would provide remain unconfirmed here. Until those rules appear, this belongs on the calendar for a closer look.
Before you start
Technology-based ideas addressing jobs, housing, financial security or fair access to digital life. Applications have not opened. The published closing date is November 2 at noon EST. Eligibility and award terms need checking when the full call opens.
The finer details
This is a date to watch. The preview names the challenge’s themes, but it does not yet establish who can apply or what an award would provide.
PCORI offers up to $300,000 over two years to help small nonprofits take part in research comparing care options. Existing patient and researcher relationships matter.
PCORI · Small-organization research capacity
Research begins to change when people with lived experience join the conversation. AI-generated illustration; not a photograph of this project.
A small nonprofit may know the people affected by a health question long before it has the means to take part in researching it. PCORI’s award helps smaller groups build that ability. Applicants must already have relationships with researchers, patients and other relevant people.
PCORI draws a firm line around “small.” The average annual operating budget across the last two financial years must be no more than $1.5 million. Previous PCORI Engagement Awards must total no more than $125,000, excluding payments for work on another organization’s award.
US nonprofits with 501(c)(3) tax-exempt status can apply, as can a fiscal sponsor handling the grant. The first short proposal, called a letter of intent, is due October 1 at 5 p.m. ET. Applicants cannot also apply to other Engagement Award calls in the same cycle. Full applications are by invitation, due January 11, 2027.
Before you start
US nonprofits with 501(c)(3) tax-exempt status, or a fiscal sponsor handling the grant. Average annual operating budget must be $1.5 million or less across the last two financial years. Past PCORI Engagement Awards must total no more than $125,000, excluding payments for work on another group’s award. Existing relationships with researchers, patients and other relevant groups are required.
The finer details
This grant helps small organizations take part in research comparing care options, with patients’ needs at its center. Applying here rules out other PCORI Engagement Award calls in the same cycle. The first step is a letter of intent, a short account of the proposed project. Full applications are by invitation and due January 11, 2027.
New additions come first. Still-open and approaching deadlines remain in view. Amounts stay in the funder’s currency. Dates are in 2026 unless stated. Checked September 13. MIT Solve is a forthcoming opening, not an application you can submit yet.
THE ARKUS SIGNAL / IDEAS & EXPERIMENTSBEYOND THE OPPORTUNITIES
Two heads need a method
Human judgment and AI can help each other. The useful question is how, and compared with what.
A person plus an AI system sounds like an upgrade. Sometimes it is. The plus sign is doing rather a lot of unpaid work.
In a 2024 review in Nature Human Behaviour, Michelle Vaccaro and colleagues examined 106 experiments that compared people working alone, AI working alone and the two working together. On average, the combined teams beat people working alone. They performed worse than whichever solo performer was better. Helping a person improve and producing the best result available are different achievements.
The kind of work mattered. When choosing among fixed options, the combined teams did worse than the stronger solo performer. Creating content looked more promising, but the evidence did not establish a clear win over the better solo result. The authors see scope for better collaboration, rather than a reason to abandon it.
This is a study of published experiments from January 2020 through June 2023, not a league table of today’s models. Differences between experiments, and the possibility that some results went unpublished, limit how widely its averages can be applied. A promising pattern still needs testing in the work you actually do.
Imagine a museum rewriting its visitor instructions. In this hypothetical test, a staff member drafts one version, an AI drafts another, and a person using AI makes a third. Visitors try the instructions without being told who wrote them. Compare whether people find their way, what errors remain and how much staff time each version takes. Agree on those measures before admiring the prose. If the combination wins, examine what each contributor did well. If it loses, changing the division of work may be more useful than adding another tool.
COMPARE THE SAME TASKPerson aloneUse the same measures for all three.
COMPARE THE SAME TASKAI aloneUse the same measures for all three.
COMPARE THE SAME TASKWorking togetherUse the same measures for all three.
Three choices. One useful nudge. There is no prize for sounding clever.
Mostly first answers? Run another test. Mostly second answers? Invite another voice. Mostly third answers? Plan the morning after. One of each? Put all three jobs in the plan.
A playful editorial prompt. Your answers stay in this browser.
THE ARKUS SIGNALTHE STRIP / THREE ORIGINAL VIEWPOINTS
Between the lines.
Three scenes, each sparked by an opportunity in this week’s editions.
Big ideas are welcome. So are the awkward questions.
Chandler LewisFounder · Managing DirectorWalter BenadofPartner · Global Markets and InvestmentsNadia LerougeDeputy Managing DirectorCloud (Yiyun) QiuPartner · Design and Innovation StrategyAarav GhaiResearch Analyst
Arkus Innovation Studios helps founders, investors and business teams decide what deserves their time and money, then put promising ideas to work. We examine ventures and products, compare competing opportunities and build practical AI systems. Curiosity gets an idea into the conversation. Evidence decides what happens next.
Our Startup Diagnostic takes a close look at one venture or product. The Portfolio Diagnostic compares several. An AI Workflow Roadmap starts with how a team actually works, identifies a task worth improving and produces a clear roadmap with a 30-60-90 day plan. Before setting that path, we check what data is available, where a person needs to review the result and what could prevent the tool from being useful.
The useful output is a clear recommendation: keep going, change course or stop. Each comes with reasoning the team can examine. An attractive idea deserves serious attention. It does not automatically deserve a larger budget.
If a funding call here connects with your work, send its public link, your idea and the part your organization can lead. We can discuss Arkus leading as prime contractor or contributing an agreed part as a subcontractor, where the rules allow and the partners agree. A grant reserved for nonprofits still needs an eligible lead. Start with the question you want to investigate; we can explore the right working arrangement from there.