The HIV Cure,
in Plain Language

A report that someone has been cured of HIV can be both true and easy to misunderstand. This guide separates exceptional individual outcomes, promising experiments, and treatments people can use now.

Part oneWhat would an HIV cure mean?

Four terms that headlines often mix together

• Viral suppression: Treatment reduces HIV in the blood. An “undetectable” result means the test did not measure virus above its detection limit, not that every part of the body is free of HIV.^{1, 2}
• ART-free remission: HIV stays controlled without regular antiretroviral treatment. Virus may remain, and control may end. Researchers sometimes call this a “functional cure,” but remission is usually the clearer word when permanence is uncertain.²
• Eradication: The goal is to eliminate HIV capable of restarting infection. Establishing that this has happened is difficult because researchers cannot test every cell and tissue.^{2, 5}
• Delayed rebound: HIV takes longer to return after ART is stopped. That is a useful research outcome, but a delay of weeks or months is not automatically lasting remission or cure.⁶

“Off pills” is not the same as “off treatment.” ART can be delivered through pills or long-acting injections. Experimental antibodies can also remain active after ART is paused, so a person may be receiving a different kind of antiviral treatment rather than controlling HIV entirely on their own.^{1, 6}

Why effective treatment does not remove all HIV

HIV can insert its genetic instructions into a person's cells. Some infected immune cells become quiet and long-lived. Together, the cells carrying persistent HIV are called the reservoir. ART blocks viral replication, but does not remove all of this hidden genetic material.^{2, 7}

Think of active virus as a running program and the reservoir as stored instructions that can restart it. The analogy has limits: these are living cells, some of which can divide and carry the instructions into new cells. Reservoirs also exist in tissues, not just in the blood that is easiest to sample.^{8, 3}

Much of the HIV DNA found in cells is damaged and cannot make a complete infectious virus. The smaller fraction that can restart infection is especially important. This is why “less HIV DNA” and “no HIV capable of rebounding” are different claims.⁹

Why there is no simple “cure test”

A routine viral-load test measures HIV RNA, genetic material from the virus, in blood. Reservoir studies ask other questions: Does a cell contain HIV DNA? Does that DNA appear complete? Can the cell produce infectious virus when stimulated? Different tests answer different questions and have different blind spots.^{1, 9, 7}

One DNA test checks selected regions of the viral genome; a copy may look intact at those locations but be broken elsewhere. More extensive sequencing can help, but can also miss copies. Tests that try to grow virus from cells can fail to activate every potentially infectious copy. A negative result therefore means “not detected in these samples by this method,” not “proven absent everywhere.”^{9, 5}

Numbers describing reservoir decline also depend on the assay, population, and length of follow-up. A 2020 study followed 81 adults on suppressive ART and found that estimated decline in intact HIV DNA slowed over time. Its much-quoted “226 years” was a mathematical extrapolation for a specified reduction in DNA, not a countdown to anyone's cure.⁸

How researchers test remission, and why not to try it yourself

An analytical treatment interruption, or ATI, is a planned pause in ART within a research protocol. Teams monitor participants frequently and set rules for restarting treatment. They may measure time to rebound, time off ART, or control below a chosen viral-load threshold; those are not interchangeable outcomes.^{6, 10}

HIV can rebound during an interruption, bringing health risks and the possibility of sexual transmission. U=U should not be assumed to continue during an ATI. Research teams need a monitoring and prevention plan for participants and their partners, not simply permission to stop medication.^{11, 10}

Part twoWhat the research has actually shown

Stem-cell transplants: proof of possibility, not a general treatment

The first widely documented case was Timothy Ray Brown, known as the Berlin patient. His transplant was undertaken to treat leukemia, not as an elective HIV treatment. The 2009 report described prolonged HIV control after transplantation from a donor whose cells carried two copies of the CCR5Δ32 mutation.^{12, 13}

CCR5 is one of the entry points many HIV variants use to infect cells. The mutation can make donor immune cells resistant to those variants. The transplant also replaces much of the recipient's immune system, and donor immune activity may help remove infected cells. Researchers are still separating the contributions of these processes.³

Subsequent cases provide important, but not identical, evidence. The London report documented 30 months without ART in 2020; the Düsseldorf report documented four years in 2023. These are follow-up durations at publication, not claims that everyone has been observed for the same length of time.^{14, 15}

The Geneva case used a donor without the CCR5Δ32 mutation and reported 32 months of remission in 2024, with other treatment and immune effects complicating the interpretation. The second Berlin case reported more than six years of remission after a donor transplant with only one copy of the mutation. These cases show that the two-copy mutation is not the only route to remission; they do not establish a safe transplant strategy for people who do not need one for cancer.^{16, 17}

In the 2026 Oslo report, one participant had no rebound during 36 months of ART interruption after a transplant from a sibling with two copies of CCR5Δ32. Researchers found no intact provirus in the sampled blood and gut cells and no replication-competent virus in extensive cell testing. They described sustained remission and possible eradication, while acknowledging the limits of sampling.⁵

These procedures can cause severe complications, including graft-versus-host disease, in which donor immune cells attack the recipient's tissues. They are complex, costly, and potentially fatal. Reviews emphasize that they are not an appropriate routine cure intervention for someone whose HIV can be treated with ART and who has no transplant-requiring cancer.^{3, 16}

People who control HIV without a transplant

Some rare individuals control HIV without ART; others maintain control after stopping treatment, often after starting it early. Researchers call these groups elite controllers and post-treatment controllers. Their experiences can offer clues about protective immune responses, but control does not necessarily mean all HIV has been eliminated.^{13, 18}

A landmark in the history: France's VISCONTI cohort

VISCONTI is the French early-treatment cohort, not the separate French transplant case. A cohort is a group of people followed over time. Its original 2013 report described 14 people who began ART within ten weeks of the earliest stage of HIV infection, took treatment for about three years, and then maintained viral control for about seven and a half years after stopping. Those durations are medians, meaning the middle values for the group, not how long every person was treated or followed.¹⁹

Researchers studied these people because they had already achieved unusual control after treatment. The report does not mean that any group of 14 people starting ART early could all stop it successfully.¹⁹

The original study did not report that everyone rebounded to detectable levels. Eight of the 14 had viral loads below the detection limit in every available sample after stopping ART; six had occasional increases. The study's controller definition used a viral load below 400 copies/mL for at least two years, so “control” did not require every test to be undetectable.¹⁹

A 2015 conference follow-up described an expanded group of 20 people. Among the original 14, one had a substantial rebound after six years and restarted ART; another had persistent low-level detectable virus with declining CD4 counts, a measure of important infection-fighting immune cells; one restarted treatment after developing cancer; one was no longer in follow-up; and ten continued to be followed with undetectable viral loads. The report did not provide the same detailed follow-up for all six additional participants.²⁰

Research has continued. A 2025 VISCONTI/PRIMO analysis studied 27 post-treatment controllers recruited through VISCONTI and CODEX and reported five ART restarts during follow-up. That is a particular research sample, not a claim that the entire VISCONTI programme always had 27 participants, nor evidence that everyone ultimately rebounded.²¹

The historical lesson is important: early treatment can be followed by unusually durable control in some people, but remission is not necessarily eradication or permanent. An occasional detectable result, sustained loss of control, and restarting ART for another health reason are distinct outcomes. VISCONTI gave researchers a model of remission to understand, not a safe instruction for other people to stop treatment.^{19, 20, 21}

Follow-up matters. The child known as the Mississippi baby had no detectable virus for 27 months after treatment stopped before HIV rebounded. A Chicago transplant recipient also experienced rebound after an initial interruption, then later entered another period of remission. A long quiet period is meaningful evidence, but not a guarantee that rebound can never happen.^{13, 22}

Broadly neutralizing antibodies: a promising clinical direction

Broadly neutralizing antibodies, often shortened to bNAbs, are laboratory-made antibodies that can block a range of HIV variants. Researchers are testing whether combinations can suppress HIV and help the body's own immune responses sustain control. Not every person's virus is equally sensitive to a given antibody.^{6, 11}

This randomized result is evidence that the antibody combination delayed the trial's rebound endpoint compared with placebo. It is not proof of eradication, a promise that three quarters of everyone with HIV could stop ART, or permission to apply the trial's restart thresholds outside that protocol.⁶

Participants who maintained suppression could receive a second infusion at week 20. That matters when a headline says “without medication”: ART had stopped, but long-acting antibodies could still be exerting an antiviral effect. Later control after antibody levels have fallen is a different and particularly important question.^{6, 11}

A September 17, 2026 Rockefeller report on additional, exploratory RIO results described links between participants' own virus-blocking antibodies and longer time off ART, along with reductions in HIV DNA that appeared intact. These findings offer clues about who responds, but finding two things together does not prove that one caused the other. The original Nature Medicine paper could not be retrieved in this recheck, so this update is attributed to the institution's report rather than presented as independently reviewed full-text evidence.²³

FRESH: why population and follow-up matter

The FRESH study in South Africa tested two antibodies with the immune-stimulating drug vesatolimod in 20 women who had begun ART very early. Unlike RIO, this was not a randomized placebo comparison of the treatment-interruption outcome. Treatment Action Group's 2025 review reported four of 20 maintaining post-treatment control through week 55.^{24, 10}

That early result should not be frozen into a permanent “20% cured” claim. NPR's later reporting described one of those four experiencing rebound and two restarting ART for other reasons. These are different outcomes: restarting does not always mean biological failure, and remaining off ART does not automatically prove cure.²⁵

FRESH also illustrates why studies in women and in African settings are essential. An outcome in a small, closely monitored group treated immediately after diagnosis may not apply to someone who started ART years later. Representation, HIV variants, other health conditions, and the practical burden of frequent visits all need attention.^{10, 25}

Other approaches: what stage are they at?

• Therapeutic vaccines: These aim to strengthen immune control in people who already have HIV. In AELIX-003, a randomized trial of 50 early-treated men, vaccines plus vesatolimod generated strong T-cell responses, but all participants experienced viral rebound. The difference in the proportion remaining off ART at 24 weeks was not statistically significant. This is a result for that regimen and study, not proof that every possible therapeutic vaccine must fail.^{26, 27}
• “Shock and kill” or “kick and kill”: The idea is to make quiet infected cells reveal HIV, then eliminate those cells. Reversing latency is only the first job; showing that the remaining reservoir is reduced enough to produce durable remission is another. A laboratory signal that HIV has “woken up” is not itself a cure.^{2, 24}
• mRNA delivery: A 2025 laboratory study used lipid nanoparticles, tiny fat-like carriers, to deliver mRNA into resting T cells and expose latent HIV. The Doherty Institute described cell-based proof of concept, with animal testing still ahead, not a treatment tested for cure in people. The use of mRNA does not turn this result into an available vaccine or cure.²⁸
• Gene editing: EBT-101 used CRISPR technology to try to cut HIV genetic material. In the initial results reported in May 2024, five people had received the intervention and three had interrupted ART; all three rebounded, one after 16 weeks. That report did not establish a cure and should not be presented as the final outcome of every gene-editing approach.²⁹
• “Block and lock”: This aims to keep HIV's genetic instructions deeply switched off rather than remove them. The supplied review of didehydro-cortistatin A describes laboratory and animal evidence, including experiments with cells obtained from people. Those are not demonstrations of durable treatment-free remission in people.³⁰
• Engineered immune cells, including CAR-T: Researchers modify immune cells to recognize infected cells more effectively. A 2026 Caring Cross announcement described nine recipients in an early duoCAR-T trial, but the release reproduced by EATG did not provide quantitative remission outcomes or sufficient follow-up to assess a cure claim. Treat it as preliminary company reporting, not a demonstrated cure rate.³¹
• Immune activators and T-cell engagers: Some agents try to stimulate antiviral responses; others help direct T cells toward infected targets. The 2025 TAG pipeline review describes early studies, including IMC-M113V, with laboratory signals or limited transient control. These approaches remain investigational, and reservoir or immune changes are not substitutes for durable clinical outcomes.²⁴
• Long-acting antibody treatment: Not every antibody study is a cure study. The EMBRACE report described N6LS combined with long-acting cabotegravir to maintain suppression, with safety and tolerability findings. Continued control on that combination is a treatment result, not proof of treatment-free remission.³²

These approaches may eventually need to be combined. The useful question is not which technology sounds most futuristic, but whether a particular intervention safely produces a meaningful, lasting benefit in an appropriately designed human study.^{2, 11}

A cure must also be usable and reachable

A cure intended for broad use would need to be safe, affordable, and scalable, not just biologically possible. UNAIDS estimated that 41.0 million people were living with HIV in 2025 and 32.1 million were receiving ART. Estimated treatment coverage among children aged 0–14 was 55%. These are 2025 estimates published in the 2026 factsheet, not real-time counts.^{2, 33}

Research also depends on sustained funding and functioning clinical services. A 2026 rapid survey found research disruption at 42% of 76 responding programmes across 32 countries following US funding freezes. That was a survey of responding HIV programmes, not a measurement showing that 42% of all cure trials had stopped.³⁴

Funding statistics require dates and categories. TAG's figure that NIH supplied about 80% of global HIV-cure research funding referred to 2021. Separately, Science reported in May 2025 that two HIV vaccine consortia would not receive renewed awards: their $129 million grants were each spread over seven years, not annual cure-research budgets. Vaccine research and cure research overlap scientifically, but their funding totals are not interchangeable.^{24, 35}

Part threeHow to read the next cure headline

Common claims, translated carefully

• “Undetectable means cured.”

  It does not. An undetectable viral load is an important treatment goal, but hidden HIV may remain. U=U concerns sexual transmission when suppression is maintained; it does not claim eradication.^{1, 2}

• “A twice-yearly injection is a cure.”

  Dosing frequency does not determine whether something is a cure. Lenacapavir as Yeztugo is used for PrEP to prevent HIV in people without HIV; testing is required before starting and before subsequent doses. Lenacapavir as Sunlenca also has a separate treatment indication, in combination with other antiretrovirals, for certain heavily treatment-experienced adults with multidrug-resistant HIV. Neither use is an HIV cure.^{36, 37, 38}

• “An HIV vaccine is available.”

  NIH's fact sheets state that there is no FDA-approved preventive or therapeutic HIV vaccine. A preventive vaccine would help people avoid acquiring HIV; a therapeutic vaccine would help people already living with HIV control it. PrEP medicines and experimental antibody infusions are not interchangeable with vaccines.^{39, 27, 37, 6}

• “Herbs, supplements, or a product sold online can replace ART.”

  The evidence reviewed here does not establish a herbal or supplement cure. UNAIDS and WHO have warned that such remedies cannot replace comprehensive HIV treatment and care; TAG also describes unsupported commercial cure claims. Ask for credible human evidence and regulatory authorization rather than testimonials, and do not replace ART with a product promoted as a cure.^{40, 24}

• “A cure is five years away.”

  There is no broadly available HIV cure. NIAID describes research toward sustained remission without daily treatment and toward eliminating the virus, including work on the reservoirs that let HIV persist. These are research goals, not a promised delivery date.⁴¹

U=U is not diminished by the absence of a cure

In PARTNER2, 782 eligible couples reported 76,088 condomless anal-sex acts while the partner living with HIV had a viral load below 200 copies/mL. There were no transmissions genetically linked to that partner. The study's conclusion supports U=U for sexual transmission; it did not test whether HIV had been eradicated.⁴

HPTN 052 likewise reported no linked infections when the partner with HIV was stably suppressed on ART. These are major clinical and prevention achievements in their own right. A future cure is a research goal, not a reason to discount the health and freedom that effective treatment can provide now.^{42, 1}

Seven questions worth asking

1. What was actually studied? Cells in a dish, animals, or people? A laboratory discovery can be important without being a treatment ready for clinical use.
2. How many people, and which people? A case report is real evidence about an individual, not a population-wide success rate. Check whether participants started ART early, had cancer, or were selected for antibody-sensitive virus.
3. What counted as success? Fewer DNA copies, stronger immune responses, delayed rebound, and undetectable viral load are different endpoints. “Still off ART” may allow detectable virus under a study's rules.
4. Was there a comparison group? Without one, it is harder to distinguish an intervention's effect from the unusual biology of participants who might control HIV anyway.
5. Was other treatment still active? An antibody that remains in the body can suppress HIV after pills stop. Ask whether researchers measured control after the intervention's antiviral activity had waned.
6. How long did the result last, and what happened to everyone enrolled? Look for rebound, restarts, withdrawal, adverse events, and missing follow-up, not only the longest-running success.
7. What kind of publication is it? A peer-reviewed trial, conference report, and company announcement carry different levels of detail and review. Check corrections and retractions too; publication is not a guarantee against error.

These questions reflect the differences visible in the trials, laboratory studies, and follow-up reports above. They also explain why a reference list alone is not enough: each claim needs the right source, endpoint, and timepoint.^{6, 26, 28, 25, 31}