The launch of the On Think Tanks (OTT) State of the Sector Report 2025 set out to do more than present charts. During this OTT Talks Live session, we asked a harder question: how are think tanks actually navigating today’s political, financial and technological headwinds—and what needs to change? I was joined by Camila Ulloa (principal author of the report), Diana Hollmann (GIZ – Knowledge for Transformation, Africa), and Christopher Chibwana (Hewlett Foundation – Gender Equity and Governance) to unpack the findings and, crucially, to hear what leading funders make of them.
Below is my readout from the conversation and the report.
Download the On Think Tanks State of the Sector Report 2025
A global snapshot—ten findings, three that matter now
The 2025 report draws on 300+ think tanks across 100+ countries, enriched by the Open Think Tank Directory and 25 regional partners. It is broad rather than statistically representative, and that is by design. Its value is in surfacing patterns sturdy enough to trigger local conversations and decisions.
For this OTT Talks Live, we focused on three cross-cutting findings:
1) Diverging optimism for growth in a tighter funding climate
Only about one in three organisations expect sector growth in the next 12 months; nearly half expect no change. Optimism is lowest in wealthier democracies and larger organisations (crowded markets; competition from consultancies, universities, advocacy groups), and highest in lower-income regions, where more than 60% foresee growth; even as ODA cuts (notably the USAID withdrawal) bite hardest.
Christopher offered four grounded hypotheses for Africa’s relative optimism:
- Exposure to USAID was uneven. Many strong African think tanks were sub- or sub-sub-recipient organisations, so the immediate shock varied; only a tiny share held direct USAID grants.
- Demand for evidence is rising. Fiscal pressure is prompting ministries to seek cost-effectiveness and efficiency.
- Philanthropic buffering. Some funders have brought forward or bridged payments to avoid job losses.
- Collaboration tailwinds. New consortia calls (e.g., AU platforms) are energising the field and nudging cooperation.
Diana underscored GIZ’s long-term investment stance: this is not a short bet, but a multi-year commitment to evidence-informed policymaking, with think tanks as change agents, by connecting African knowledge to African and global decision-making spaces.
Still, the discussion about growth challenges us to ask: what do we mean by growth? As Diana pointed out, this could be about headcount and budget, or resilience and impact. Smaller organisations can be highly consequential if they are strategically clear, well-governed, and networked.
Christopher pushed the field to consider mergers and acquisitions (yes, in think tanks): why sustain three parallel organisations doing the same thing (each with its own expensive back-office) when a combined entity could deliver better value for money and a stronger voice? Egos and politics make this hard, but the economics are compelling. I am not sure about this suggestion – think tanks are not simple service providers – but it is nonetheless worth exploring.
My take: There is also a new breed of policy entrepreneurs, local philanthropists, and funders emerging across low and middle-income countries and emerging economies. We are increasingly contacted by young think tank funders who have received backing from peers in the local private sector or universities, which have been given a new mandate by their governments to establish new think tanks.
2) Political polarisation is up; trust in evidence is the buffer
More than one-third of organisations report being significantly affected by political polarisation; up 12 percentage points from last year. Effects include harder research sharing, fewer collaborations, tighter funding, and reduced media access. In East and Southeast Asia, the “significantly affected” share rose from 7% to 35%; in Latin America & the Caribbean, from 31% to 54%.
A crucial protective factor emerged: where societal trust in evidence is high, think tanks report less disruption, regardless of regime type or income level. Conversely, where evidence is not valued, half of the think tanks are found operating “hard or very hard”.
Christopher reminded us: evidence is a means, not an end. “A riot on the street can be more powerful than a p-value.” The implication is not to abandon rigour but to re-earn legitimacy:
- Be transparent about funding and conflicts of interest.
- Safeguard intellectual autonomy from funders and politics.
- Re-centre audiences: engage communities and media, not just broadcast expertise.
- Funders should avoid product-only grants and support the relational work (meetings, trust, and political communication) that makes evidence usable.
Critical to this discussion is the think tank’s association with elite political communities, which are often detached from the general public. So, do think tanks’ agendas align with public concerns?
One lively exchange contrasted think-tank priorities with public opinion polls (Afrobarometer, Latinobarómetro, etc.). In most regions, agendas align, at least at the aggregate level; however, there are notable gaps, e.g., Latin America: polls put crime and security at the top, while surveyed think tanks prioritised elsewhere; EU Europe: publics stressed the cost of living, while think tanks reported a greater emphasis on the environment.
My take: alignment should not mean populism. But building a constituency for ideas requires knowing where publics stand—and explaining why certain issues matter to them.
In polarised settings, “communication” increasingly means political communication: negotiation, coalition-building, risk management.
3) Organisational plumbing and an AI divide will shape resilience
Across sizes and regions, fundraising remains the #1 capacity gap (70%+). But the report also shows that vulnerability is not only about size or grant type; it is about internal design:
- Organisations without finance or HR teams are far more likely to report operational gaps.
- Groups identifying more as NGOs (with looser ties to academic networks) report gaps in research methods.
- In short: structure predicts resilience.
In terms of technology, AI adoption is rising rapidly (~71% use it primarily for research and communications), but it is uneven: over 80% uptake in Europe versus approximately 50% in South Asia and the MENA region. Barriers are infrastructure, training, and resources, not interest. The risk is a new inequality line: better-resourced institutions accelerate; others are stuck as tool-takers, not tool-shapers.
Christopher’s caution is timely: if models are trained on Northern data, uncritical use can misrepresent local contexts and erode analytical capacity. Funders should support responsible use, evaluation of model performance in Majority World contexts, and—where possible—participation in model development.
My take: core administrative, project management, strategic planning and AI literacy are no longer overhead; they are mission-critical. The sector needs greater collaboration and shared services (HR/finance/MEL/IT), as well as guardrails for AI use.
What funders are (and aren’t) funding – and what should change
A theme threaded through the discussion: we love to fund the visible (the glossy report, the panel, the “deliverable”). But influence depends on what is hardest to fund: relationships, positioning, legitimacy, and field-level infrastructure.
Three practical shifts (of many others):
- Back the back-office (together). Encourage shared infrastructure across organisations to lower indirect rates, and co-fund core systems (finance, HR, MEL, etc.).
- Fund the pathway, not just the paper. Resource audience work (policy partnerships, media strategies, constituency building) alongside research.
- Invest in responsible AI capacity. Support training, contextual model testing, and open data systems so organisations can shape, not just consume, AI tools.
Three levers to pull in 2026
Drawing the conversation to a close, Camila highlighted the foundations of a stronger, more equitable sector, which requires:
- Strategic clarity – know your purpose, audiences, and route to influence.
- Strong internal systems – finance, human resources, governance and management, MEL, and now AI literacy.
- Societal trust in evidence – earned through transparency, independence, and relevance.
I would add:
- Ecosystem thinking – strengthening the entire think tank community, such as through collaborating on delivery and on back-office investments.
- Responsible technology – adopting AI with intent, context and safeguards.
- Constituency building – engaging the public and media so evidence has a home.
An invitation
This launch event inevitably spotlighted Africa — reflecting our guests’ portfolios — but the report covers every region. We’re keen to support local and regional discussions that test these findings against lived realities. If you would like to convene peers in your country or network, please get in touch. The State of the Sector is not a verdict; it’s a conversation starter.
Download the report, dissect the data, and tell us what we’ve missed.