Beyond the State Department’s Bad Map: What Mattered at Global AIDS Conference

Now let’s focus on the map that matters: the one on which HIV finally disappears.

I’ll be honest. As a former State Department official, I was embarrassed by the inaccurate map shown during a U.S. government presentation at AIDS 2026. It should never have made it onto the screen, especially during a presentation about partnerships with African governments.

I was also disappointed that this mistake became one of the conference’s biggest media stories. Because AIDS 2026 produced important science, serious policy analysis and real progress on issues ranging from PrEP delivery to global aid transitions.

Rather than add to the gotcha coverage, I will highlight the six stories I think that matter more.

First, a caveat. Two people can attend the same conference and have completely different journeys, conversations, partnerships and takeaways. That is one of the beauties of conference week. Yet for that one week, we come together as a family around the goal that brought us here in the first place: erasing HIV from the map. Now that is the cartographic correction that matters.

1. The HIV prevention revolution is here

The HIV prevention revolution is no longer coming. It has started.

Gilead’s lenacapavir, which showed near-complete protection in clinical trials, is now reaching people through national programs. South Africa accounts for just over half of all lenacapavir users in African national rollouts, with 31,297 of roughly 60,000 users. Among people offered a choice, 98.5% selected twice-yearly lenacapavir over daily oral PrEP. About one-third had never used PrEP before.

So Lenacapavir is not simply giving existing PrEP users a more convenient option. It is bringing new people into HIV prevention. Women represent 72% of users so far, and pregnant and breastfeeding women account for more than one-quarter. This is particularly encouraging given the disproportionate HIV risk faced by women and girls across much of sub-Saharan Africa. Read the great analysis here from Bhekisisa.

But, and this a huge but. Uptake remains strikingly low among sex workers, gay and bisexual men, transgender people and people who inject drugs. The closure of community clinics following U.S. funding cuts may be one reason. A breakthrough product cannot reach people if the trusted organizations serving them have closed their doors.

ViiV Healthcare also showed that long-acting prevention will not stop with the options available today. Its AIDS 2026 program included real-world findings on Apretude, its injectable cabotegravir product, as well as early work on a new formulation that could reduce dosing from six injections a year to three. ViiV also presented findings comparing people’s experiences with cabotegravir and lenacapavir injections. That work matters because comfort, convenience and personal preference will help determine which products people choose and continue using.

Therese three slides from Raphael Landovitz’s plenary sum it up:

Merck added another important element to the to the future. The company announced voluntary licenses with seven generic manufacturers for alimatravir, its investigational once-monthly oral PrEP pill, covering 129 countries. Three licensees are based in Kenya, Uganda and South Africa. Merck made the announcement while Phase 3 trials are still enrolling, giving manufacturers, governments and donors a chance to prepare for access before approval rather than years afterward. That represents real progress.

Still, the licensing agreement itself leaves out most of Latin America, including countries participating in the drug’s clinical trials. That is a familiar and unacceptable pattern: communities help generate the evidence but may not share promptly in its benefits. Early licensing deserves credit, but access must follow epidemiology, not geography or World Bank classifications.

The International Treatment Preparedness Coalition (ITPC) captured this contradiction brilliantly with my favorite conference display showing how companies can treat PrEP like a luxury product: innovative, desirable and available mainly to those in the right markets. See picture below. The point landed. HIV prevention is not a designer handbag.

At the Global Village, ITPC’s “Pharma Privé” makes the point brilliantly: lifesaving medicines should never be luxury goods kept behind a velvet rope.

Together, lenacapavir, cabotegravir and alimatravir could move us beyond a one-size-fits-all approach. People may soon be able to choose daily, monthly, multi-monthly or twice-yearly protection based on what works for their lives.

The science has delivered the beginnings of a prevention revolution. Now governments, companies, donors and communities must ensure that it moves at the speed of need rather than the speed of markets.

2. We need to make PrEP much, much easier to get

A prevention revolution means little if people still need to navigate a maze to participate in it.

Last year, I argued that getting PrEP should be as easy as getting Ozempic has become for some people. Rio only reinforced that view. When demand exists, health systems and markets can move quickly. HIV prevention deserves the same convenience.

“Demedicalizing PrEP” emerged as a major theme in Rio, and for good reason. PrEP is prevention, not a medical emergency. Yet we often deliver it through systems designed to manage illness: repeated clinic visits, long waits, unnecessary laboratory requirements and limited dispensing hours.

The conference offered a better blueprint. In Kenya, trained pharmacists are initiating and refilling PrEP without requiring physician visits. Brazil is testing dispensing machines in transit hubs (see picture below), allowing people to collect prevention privately and conveniently. In the United States, MISTR’s telehealth model allows people to complete consultations online and receive PrEP at home. Community-led programs in South Africa and Thailand are bringing services directly into spaces where key populations already gather and feel safe.

São Paulo is making HIV prevention fit into people’s lives: PrEP and PEP through a dispensing machine—private, convenient and easy to access. This is what demedicalizing prevention looks like.

These models do not lower clinical standards. They remove barriers that add little medical value but drive people away. And the same principle must shape the rollout of long-acting products. A twice-yearly injection will not transform prevention if people must travel hours to an overcrowded clinic to receive it. Countries should plan now to deliver new options through pharmacies, community organizations, mobile services and other trusted settings.

The remaining barriers are no longer primarily scientific. They are operational and political: who can test, prescribe, administer and dispense PrEP; how providers get paid; how public, private and community data connect; and how programs protect quality, privacy and continued use at scale.

Countries now need a common multi-channel blueprint connecting pharmacies, community organizations, public facilities, telehealth and home delivery. That means creating adaptable regulatory pathways, building interoperable data systems, paying community partners for their work and measuring success through equitable access and continued use, not the number of pilots launched or products distributed.

The message from Rio was simple: prevention must fit into people’s lives. We cannot keep asking people to reorganize their lives around the health system.

If we can withdraw cash from a machine, order dinner from a phone and receive almost anything at our door, we can surely find easier ways to deliver PrEP.

3. Communities built the HIV response. They came with receipts.

You did not need to travel to Rio to learn that communities are the backbone of the HIV response. If you have worked in HIV for more than a few months, you should already know that. Still, some institutions apparently needed the refresher.

Slide from Micheal Ighodaro’s presentation

The conference program was rich with examples of community-led monitoring and implementation. Community leaders showed how they identify service gaps, track medicine stockouts, reach people traditional health systems miss and turn policy promises into real services. They also presented practical solutions, including peer-led testing, community delivery of HIV treatment and prevention, and digital outreach designed around the realities of young people.

At the same time, it was clear that this part of the HIV response has taken one of the greatest hits over the past year. Funding cuts have weakened community organizations (read amfAR report), disrupted monitoring and removed many community voices from decisions that directly affect them.

These organizations form part of the HIV response’s essential infrastructure. They often see problems before governments or donors do. They know when stigma drives people away from clinics, when services disappear and when a policy that looks good in Washington fails in practice.

The State Department now has some catching up to do. It must bring civil society and NGO partners back to the table, although the truth is that communities built much of that table in the first place.

Doing so would restore the transparency, accountability and visibility that have long defined PEPFAR but have been lost under the current approach. Community involvement makes HIV programs better. It identifies problems earlier, strengthens accountability and keeps policies connected to what people actually experience.

Post-Rio offers a good moment for a reset. The State Department should take stock of the first 18 months of its new approach and pause the MOU implementation process long enough to assess what is working, what is not and where communities remain absent. A course correction now could prevent much larger failures later.

As financing tightens and countries integrate HIV services into broader health systems, community leadership will become even more important. We must fund communities, share decision-making power and include their knowledge from the beginning. Integration cannot mean invisibility, especially for the people most affected by HIV.

4. AI is already part of the HIV response. Now we must make it work at scale

Izukanji Sikazwe gave one of the best presentations I heard all week. It was a showstopper, but not because she promised that AI would solve everything. Quite the opposite. She cut through the hype, named the risks and showed what responsible use could look like.

She reminded us that HIV programs have used AI-related tools for decades. UNAIDS Spectrum models support national planning in more than 170 countries. Supply-chain systems use consumption data, patient volumes and procurement timelines to forecast where medicines should go.

We did not always call this AI, but the basic idea is not new.

Her main concern was “pilotitis,” the endless cycle of promising pilots that never become real programs. A tool may work for 200 patients with a dedicated research team. That tells us very little about whether an overstretched clinic can use it every day.

“Scale is where real impact lives” was probably the most important line of her presentation. She also pointed to the Global Fund’s investments in digital health as an example of what building for scale can look like. The Global Fund now invests roughly $150 million a year in digital tools across more than 90 countries.

But Izukanji was equally clear about the risks. Scaling AI without strong governance could deepen the same inequities the HIV response has spent decades fighting. These systems can reproduce bias, exclude marginalized communities, misuse sensitive health information and destroy trust.

Presentations from Audere Africa brought some of these ideas down to earth. One analysis of Aimee, a WhatsApp-based AI companion in South Africa, included more than 9,300 adolescent girls and young women. Among those who engaged with the AI, 45% took up HIV testing, PrEP or both. That suggests AI can become a new way into care, not simply another place to find health information.

A related study found that structured HIV-risk prompts produced more complete assessments than conversations directed entirely by users, 67% compared with 6%. The lesson was not that we need a fancier model. We need to design better conversations.

Another analysis exmined more than 300,000 interactions across Aimee, Coach mPilo and VimbAI. In the interactions reviewed, researchers found no harmful misinformation and no false negatives in detecting self-harm risk. Every identified high-risk interaction was escalated according to protocol. Automated triage also reduced clinicians’ workload while keeping people involved in the decisions that mattered.

A fourth presentation showed that AI conversations can capture more than 120 indicators covering HIV, sexual and reproductive health, mental health, behavior and social conditions. That information can help programs understand what communities are experiencing, target outreach and reduce the reporting burden on health workers.

This could become especially useful as funding cuts weaken community-led monitoring. But AI cannot become an excuse to replace or defund communities. It should help health systems hear them more clearly and respond more quickly.

The question is no longer whether AI belongs in the HIV response. It already does. The real question is whether we can use it responsibly and at scale.

The State Department can help move this work beyond pilots. Its new Advancing Global Health Annual Program Statement is designed to fund projects that complement bilateral agreements and fill implementation gaps. GHSD should issue an addendum focused on HIV and responsible AI, with funding for country-led and community-governed tools that can become part of real programs.

Izukanji ended with an image that stayed with me. We should treat AI more like a power grid than a product. It should support human interaction, not replace it. And we should judge it by outcomes that matter: infections prevented, people reached and lives kept healthy and engaged in care.

5. The tension we cannot ignore: Countries are finding solutions, but funding cuts are causing real harm

One of the most important tensions in Rio ran through several sessions.

On one side, amfAR and others presented evidence of what funding cuts have already caused: disrupted services, prevention gaps, workforce losses and rising risks for vulnerable populations.

On the other side, CHAI and country partners showed how governments are responding. Every country presented plans for reduced external funding, greater domestic financing and stronger government ownership.

Integration emerged as the dominant strategy. Countries are bringing HIV, TB and malaria programs together while consolidating laboratories, supply chains, workforce training and digital systems. They are trying to reduce fragmentation and make limited resources go further.

These reforms were not simply cuts dressed up as efficiency. Malawi described how it is simplifying treatment regimens, testing and training while protecting the quality of care.

Cambodia presented one of the clearest quantified transition plans. Faced with funding reductions and low PrEP uptake, the country selected a scenario for scaling up long-acting cabotegravir that could save an estimated $19.1 million over five years. (see below)

Zimbabwe reported nearly $967,000 in savings through integrated quantification and stronger coordination of supply chains across disease programs.

These are serious examples of countries adapting under pressure. But we should not mistake adaptation for evidence that funding cuts carry no consequences. Countries can improve efficiency, but efficiency alone cannot replace lost financing, health workers or community infrastructure.

The shared lesson was that transition takes time. Countries need phased implementation, interoperable systems, sustained technical support and realistic plans for absorbing health workers into government structures. None of that happens because an MOU sets a deadline.

This is where the State Department must focus. It should support what countries are trying to build, identify where transitions are already creating gaps and align technical assistance with country plans. It must also recognize that a government’s ability to improvise should not become an excuse to withdraw support faster.

Just a side note: The Rio Conference also drew many foundation leaders and representatives from donor governments. I was especially pleased to see the French Ministry for Europe and Foreign Affairs with such a prominent presence, a timely reminder that global HIV leadership does not begin and end with the United States.

6. We need to be more honest about “resilience”

We use the word “resilience” constantly. It appears in reports, speeches and descriptions of systems under pressure.

But in Rio, that word felt insufficient.

What I saw was struggle, improvisation, exhaustion, creativity and refusal. I saw people holding together systems that were never designed to support them in the first place.

I also saw activists turning on one another as resources disappeared. Scarcity does that, and it serves the people responsible for the cuts. We cannot allow a fight over shrinking resources to fracture the movement. We have to resist that.

Too often, we celebrate resilience when what we really mean is that people have been left to absorb the damage.

Much of the real work happened outside the official program. It happened in hallways, side events, late-night meetings and informal gatherings where people often build the true architecture of the response.

New alliances are also emerging outside traditional global health structures. I saw community organizations, private companies and nontraditional partners working together in ways that felt practical, not performative. These are not partnerships of convenience. They reflect a system under pressure. But they are also smart approaches.

“Resilience” may be too clean a word for this reality. It can hide the friction, inequity and constant negotiation required to keep services running. Worse, it can let institutions praise communities for surviving problems those institutions helped create.

The determination I saw in Rio was real. But communities should not have to prove their resilience by being asked to survive one crisis after another.

What I hope we remember from Rio

The global HIV response is not over.

In 2025, an estimated 1.2 million people acquired HIV and 570,000 died from HIV-related causes. The latest available breakdown shows that key populations and their sexual partners accounted for 49% of new infections globally in 2024. They must be the centerpiece of an evidence-informed HIV response.

At the same time, we have better prevention and treatment tools than at any previous point in the epidemic. The science keeps moving forward, but our policies, financing and delivery systems are struggling to keep pace.

Rio also reminded me that no institution or sector can close this gap alone. We need new partnerships that connect communities, governments, researchers, implementers, advocates, technology companies and the private sector. We also need people willing to move across these worlds, challenge the silos and keep everyone focused on results.

Personally, I left Rio seeing more clearly the silos within our field, the tension between policy and people’s lived experiences, and the disconnect between country realities and decisions made globally. That strengthened my commitment to being a bridge builder. I want to continue to bring programmatic realities into policy discussions in Washington and around the world. I left more motivated than ever to deepen my current partnerships and build new ones around the work ahead.

The map was wrong. Now let’s focus on getting the HIV response right!

---

Recent Stacks