HIV Explained — How the Virus Works and Why It's Treatable Now

HIV is a retrovirus that writes its genes into the DNA of the immune cells it infects, and that single trick explains both why it nearly broke modern medicine and why it is now controllable but not curable. Daily or long-acting drug combinations can push the virus below the level a blood test can detect, and at that point it no longer passes to sexual partners. This article explains how HIV gets into a cell, why the first decades of the epidemic were so lethal, how today's drugs work, and what the 2025 numbers say about where the epidemic stands.

Educational content, not medical advice. This article describes general, well-established science. Decisions about testing, prevention, or treatment belong with a clinician who knows your situation.

What HIV actually is

HIV, the human immunodeficiency virus, belongs to the lentiviruses, a genus of retroviruses that cause slow, chronic infections. A virus particle is roughly 100 nanometers across, a layer of lipid membrane stolen from a previous host cell wrapped around a protein capsid (a cone-shaped shell) that holds two copies of an RNA genome and a small kit of enzymes. Think of it as a shipping container carrying a set of instructions plus the three tools needed to act on them.

The defining feature is the direction of information flow. Most cells copy DNA into RNA and RNA into protein. A retrovirus carries RNA and runs the first step backward: its enzyme reverse transcriptase builds a DNA copy of the viral genome, and a second enzyme, integrase, splices that copy into a chromosome of the host cell. Once spliced in, the viral DNA is called a provirus, and for the lifetime of that cell and all of its descendants it is just another stretch of the genome.

Two types infect people. HIV-1 drives the global pandemic. HIV-2 is concentrated in West Africa, is generally less transmissible, and tends to progress more slowly. Both trace back to simian immunodeficiency viruses (SIVs) in African primates, with HIV-1 derived from viruses infecting chimpanzees and gorillas. Molecular clock analyses of archived Kinshasa samples have placed the common ancestor of the main pandemic group of HIV-1 around 1908, which means the virus circulated for decades before anyone recognized a syndrome. The general mechanics of animal-to-human transfer are covered in our piece on how viruses jump from animals to humans.

The disease name needs one clarification. HIV is the virus, and AIDS (acquired immunodeficiency syndrome) is the late stage of untreated infection, when the immune system has been damaged badly enough that infections and cancers that healthy people shrug off take hold. A person can live with HIV for years without AIDS, and with modern treatment can live indefinitely without developing it. For the broader biology of viruses, including how they differ from bacteria and why they cannot reproduce without a host, see our guide to how viruses work.

How HIV infects an immune cell, step by step

HIV's main target is the CD4+ T cell, sometimes called a helper T cell. These cells are the coordinators of the adaptive immune system: when a dendritic cell shows them a fragment of a pathogen, they release chemical signals that recruit antibody-producing B cells and killer T cells. HIV also infects macrophages and other cells carrying CD4, but the loss of helper T cells is what eventually disables the response. The way T cells read those fragments is described in our explainer on T-cell receptor recognition.

The life cycle runs through seven stages, and each one is a drug target.

1. Attachment. The envelope protein gp120 binds to the CD4 molecule on the cell surface, then to a second docking molecule called a co-receptor, usually CCR5 or CXCR4. The pair works like a key that has to turn two locks in sequence.
2. Fusion. The envelope protein gp41 folds into a hairpin that pulls the viral and cell membranes together until they merge, releasing the capsid into the cytoplasm.
3. Reverse transcription. Inside the capsid, reverse transcriptase copies the RNA genome into double-stranded DNA. This enzyme has no proofreading function, so it makes errors at a high rate, and that error rate is the engine of HIV's variability.
4. Integration. The new DNA enters the nucleus, and integrase inserts it into a chromosome. After this, the cell is permanently infected.
5. Transcription and protein production. If the cell is active, host machinery reads the provirus and produces viral RNA and long precursor proteins. If the cell is resting, the provirus can stay silent.
6. Assembly and budding. New RNA genomes and proteins gather at the cell membrane and push out through it in immature particles, each carrying a piece of host membrane.
7. Maturation. The viral protease cuts the long precursor proteins into their working pieces, and the particle rearranges into an infectious virion.

An infected, actively producing cell is a short-lived factory. The more important feature for the long run is the second path at step 5, where the provirus goes quiet.

The disease in stages

Without treatment, HIV infection follows a recognizable arc. In the first weeks, many people have a flu-like illness while viral levels in blood climb steeply. This acute phase is when a person is most infectious, and also when standard antibody tests can miss the infection because the immune response has not yet built up. A chronic phase follows, often lasting years with few symptoms, during which the virus keeps replicating and CD4 counts slowly fall. A healthy adult typically has several hundred CD4 cells per microliter of blood, and clinical definitions of AIDS have long used a count below 200 or the appearance of specific opportunistic infections.

Opportunistic infections are the diseases that take advantage of an immune system with a missing coordinator: Pneumocystis pneumonia, tuberculosis, certain fungal and parasitic infections, and cancers such as Kaposi sarcoma. The disease was first recognized through exactly this pattern. On June 5, 1981, the U.S. Centers for Disease Control and Prevention's Morbidity and Mortality Weekly Report described Pneumocystis pneumonia in five young men in Los Angeles, according to the HIV.gov timeline. Researchers at the Pasteur Institute led by Françoise Barré-Sinoussi reported a retrovirus that could be the cause on May 20, 1983, and the first licensed blood test for the virus followed on March 2, 1985. The first drug, zidovudine (AZT), was approved on March 19, 1987, per the same timeline.

The timing of those events matters. For the first years, the cause was unknown, there was no test, and a person could transmit the virus long before any symptom appeared. That window, more than any single biological property, shaped the early epidemic.

Why HIV was so hard to treat

Four biological features stacked on top of one another, and each one defeated an earlier therapeutic idea.

It destroys the cells that would fight it. Most viruses meet an immune system that is at full strength. HIV goes after the coordinators, so the response weakens as the infection continues. The effect resembles removing air traffic controllers from a busy airport: the planes are still flying, but nobody is directing them.

It mutates constantly. Because reverse transcriptase does not proofread, every round of copying introduces changes. Inside one untreated person, billions of virions are produced daily, which produces a swarm of slightly different variants known as a quasispecies. A drug that blocks one target will usually leave some variants that happen to resist it, and those variants then take over. This is the same selection logic discussed in our piece on how antibiotic resistance spreads, compressed into weeks instead of decades. It is also the reason early single-drug treatments failed within months and the reason modern therapy always uses combinations.

It hides. Some infected cells, particularly long-lived memory T cells, carry a provirus that is transcriptionally silent. Silent proviruses make no viral proteins, so the immune system cannot see them and drugs that target viral enzymes have nothing to act on. Studies of this latent reservoir in treated patients have found that it decays very slowly: a 2003 paper in Nature Medicine by Siliciano and colleagues reported a half-life of about 44 months, which would mean that clearing it with drugs alone would take decades. Later work suggests the picture is more complicated, because infected cells can also multiply, which keeps the reservoir topped up.

It integrates. A drug can stop new cells from becoming infected, but it cannot find a single silent provirus inside a cell that looks normal and remove it. This is the central obstacle to a cure.

There was also a social failure layered on the biological one. Stigma, delayed political attention, and slow research funding in the 1980s cost years that the biology did not require. The WHO notes that cumulative HIV-related deaths have reached about 44.2 million, a total that includes those years.

How antiretroviral therapy works

Antiretroviral therapy (ART) combines drugs from different classes so that the virus has to acquire several resistance mutations at once to escape. That is statistically very unlikely when the drugs are taken consistently, because each drug also suppresses the replication that would generate the mutants. The principle is the same as using a lock and a deadbolt and a chain: a burglar who can beat one will usually fail at all three.

Drug class	What it blocks	Example drugs
Nucleoside reverse transcriptase inhibitors (NRTIs)	Act as faulty DNA building blocks that end the growing chain	Tenofovir, emtricitabine, abacavir
Non-nucleoside reverse transcriptase inhibitors (NNRTIs)	Bind reverse transcriptase and jam it	Doravirine, rilpivirine, efavirenz
Integrase strand transfer inhibitors (INSTIs)	Block integrase from splicing viral DNA into the host chromosome	Dolutegravir, bictegravir, cabotegravir
Protease inhibitors	Prevent maturation after budding, leaving non-infectious particles	Darunavir, atazanavir
Entry and fusion inhibitors	Stop the virus attaching to or fusing with the cell	Maraviroc, enfuvirtide
Capsid inhibitors	Interfere with capsid function at several stages	Lenacapavir

Integrase inhibitors now form the backbone of most first-line regimens, often combined with two NRTIs in a single daily tablet. Long-acting injectable options have widened the choices. Cabotegravir with rilpivirine is given by injection every two months for people who are already suppressed, and lenacapavir, a capsid inhibitor, was first approved in 2022 for people whose virus resists multiple other drugs.

The aim of treatment is viral suppression: a viral load (the number of HIV RNA copies per milliliter of blood) so low that tests cannot detect it. The CDC describes a high viral load as 200 copies per milliliter or more, and suppression is generally defined as staying below that threshold. Most people who start ART reach suppression within months. Staying suppressed depends on taking the drugs consistently, because stopping lets the latent reservoir restart replication, usually within weeks.

Modern drugs are also much easier to tolerate than the early ones. Treatment guidelines for adults are published and updated by national and international bodies, including the WHO consolidated guidelines and the U.S. Department of Health and Human Services antiretroviral guidelines, and they are the right place to look for current regimens rather than any general article.

U=U: undetectable equals untransmittable

The most consequential clinical finding in the field is that a person on ART with a durably undetectable viral load does not pass HIV to sexual partners. The WHO fact sheet states it directly: people on ART with an undetectable viral load will not transmit HIV to their sexual partners. The CDC says that having an undetectable viral load prevents transmission during sex.

The evidence is large enough to be unambiguous. The HPTN 052 trial, led by Myron Cohen, followed heterosexual couples in which one partner had HIV and found that treating the partner with HIV sharply reduced transmission, and its final 10-year results confirmed that ART is, in Cohen's words, a highly effective and durable intervention for prevention. The PARTNER2 study then enrolled 782 male couples across 14 European countries between 2010 and 2017. Those couples contributed almost 1,600 eligible couple-years of follow-up and more than 76,000 reports of condomless sex, with the partner who had HIV on ART at a viral load under 200 copies per milliliter and the other partner not taking PrEP or post-exposure prophylaxis. There were zero within-couple transmissions, according to the HIV.gov summary of the 2019 Lancet paper.

Two qualifications are worth stating plainly. U=U depends on staying suppressed, which requires continuing treatment and periodic viral load tests. And the CDC notes that a small number of people cannot reach an undetectable viral load even when they take medication as prescribed, and those people have other prevention options to use. The finding also concerns sexual transmission. Blood exposures and mother-to-child transmission are governed by other evidence, discussed below.

Preventing infection: PrEP, PEP, and prevention in pregnancy

Prevention now includes a set of drug-based options that work at the same viral targets as treatment.

Pre-exposure prophylaxis (PrEP) means HIV-negative people take antiretroviral drugs before potential exposure so that the virus cannot establish itself. The WHO now recommends oral tenofovir-based PrEP, the dapivirine vaginal ring, long-acting injectable cabotegravir, and long-acting injectable lenacapavir.

The newest option has the strongest results to date. In the PURPOSE 1 trial, published in the New England Journal of Medicine in 2024, researchers randomized 5,338 initially HIV-negative adolescent girls and young women in South Africa and Uganda to twice-yearly lenacapavir injections or to one of two daily pills. There were no HIV infections among the 2,134 women who received lenacapavir, compared with 39 among 2,136 on daily emtricitabine and tenofovir alafenamide and 16 among 1,068 on emtricitabine and tenofovir disoproxil fumarate. The background infection rate in the screened population was 2.41 per 100 person-years, and lenacapavir's rate was significantly lower than that figure. The U.S. FDA approved lenacapavir for PrEP, under the brand name Yeztugo, on June 18, 2025.

Trial figures need careful reading. "Zero infections" in a trial arm is an observed result in a specific population over a defined period, not a guarantee for every person. Adherence, testing before starting, and local circumstances all matter, and cost has been a concern raised by advocates.

Post-exposure prophylaxis (PEP) is a short course of antiretrovirals taken after a possible exposure, such as an occupational needle stick. It has to be started quickly, within 72 hours, and is taken for 28 days.

Mother-to-child transmission is the other major route. Without any intervention, a substantial share of infants born to mothers with HIV acquire it, and the risk is higher when breastfeeding continues for a long time. When mothers receive ART during pregnancy, delivery, and breastfeeding, the risk falls to very low levels, and this is one of the main reasons treatment programs prioritize pregnant women.

Transmission routes are narrow. The WHO lists blood, breast milk, semen, and vaginal secretions, and states that HIV is not spread by kissing, hugging, shaking hands, or sharing personal objects, food, or water.

Is there a cure?

No cure is available as a routine treatment, but a small number of people have been cured through bone marrow (stem cell) transplants that were given to treat blood cancer. The first, Timothy Ray Brown, known as the Berlin patient, received his transplant in 2007 from a donor with two copies of a rare variant called CCR5-delta-32. People with that variant make cells that lack a working CCR5 co-receptor, so the main entry route for HIV is closed. The transplant replaced his immune system with resistant cells, and he remained free of the virus until his death from cancer in 2020. The London patient (2019) and the Düsseldorf patient followed. A team at Düsseldorf University Hospital reported the third case in Nature Medicine in February 2023: a man transplanted in 2013 who stopped ART in 2018 and showed no viral rebound. A fourth case, the Geneva patient, was announced ahead of the 2023 International AIDS Society conference and is notable because the donor did not carry the CCR5 variant, although specialists cautioned that long-term monitoring is needed.

These cases prove that eliminating HIV from a body is biologically possible. They do not give a practical therapy, because a stem cell transplant carries a high risk of death and is only justified for people who also have life-threatening cancer. Current research on a broader cure follows four main lines:

• Latency reversal ("shock and kill"). Drugs wake up silent proviruses so the infected cells can be recognized and cleared. Early attempts have woken some proviruses without removing the cells.
• Gene editing. Tools like CRISPR could cut out proviral DNA or recreate the CCR5 variant in a patient's own cells. For background on the technology, see our explainer on what CRISPR is and how gene editing works and the update on what gene editing has delivered so far.
• Therapeutic vaccines and immune boosters. The goal is to train the immune system to hold the virus down without drugs.
• Broadly neutralizing antibodies. Engineered antibodies bind parts of the virus that vary little between strains.

None of these approaches is an approved cure. The honest summary is that the field has several credible leads and no finished product.

Why there is still no HIV vaccine

Most successful vaccines work by training the body to recognize a stable feature of a pathogen and then clear it before it establishes. HIV breaks those assumptions. Its surface protein varies enormously between strains and even within one person, so antibodies that work against one version often miss another. The envelope is also heavily coated in sugar molecules that shield vulnerable sites. And because the virus integrates early, it can seed the reservoir within days of exposure, before any vaccine-trained response could clear it.

Researchers are working on designs intended to coax the immune system into making broadly neutralizing antibodies, and messenger RNA platforms have made some experiments faster to run. The platform itself is described in our article on mRNA vaccines, and the general principles of immune memory in how vaccines work. The scientific problem is different from those of influenza or COVID-19, though, and no licensed HIV vaccine exists. The WHO page on HIV does not list one among available prevention tools.

The global picture

UNAIDS estimates that 41.0 million people were living with HIV at the end of 2025, with 1.2 million new infections and 570,000 AIDS-related deaths that year. About 32.1 million people were on antiretroviral therapy at the end of December 2025. Annual new infections peaked at 3.5 million in 1994, and AIDS-related deaths peaked at 2.1 million in 2004, per the UNAIDS global fact sheet.

Measure	2025 estimate	Peak
People living with HIV	41.0 million	Not stated
New infections	1.2 million	3.5 million (1994)
AIDS-related deaths	570,000	2.1 million (2004)
People on ART	32.1 million	Not applicable

Source: UNAIDS global HIV statistics, data year 2025.

The "95-95-95" targets aim for 95 percent of people with HIV to know their status, 95 percent of those diagnosed to be on treatment, and 95 percent of those treated to be virally suppressed. The two main sources report progress with different denominators, which can confuse comparisons. UNAIDS measures each step against all people living with HIV: 88 percent knew their status, 78 percent were on treatment, and 74 percent were virally suppressed. The WHO fact sheet uses the standard cascade, where each step is measured against the previous one: 88 percent of people with HIV knew their status, 89 percent of those diagnosed were on ART, and 95 percent of those on ART were suppressed. Both are accurate, but they answer different questions. The gap that matters most is between knowing one's status and starting treatment.

WHO data put 64 percent of people living with HIV in the WHO African Region. Infections there fall disproportionately on adolescent girls and young women. Outside Africa, transmission is concentrated among key populations, including men who have sex with men, transgender women, people who inject drugs, and sex workers, plus their partners. WHO, the Global Fund, and UNAIDS share a goal of ending the epidemic as a public health threat by 2030.

How HIV compares with other viral infections

Feature	HIV	Hepatitis C	Influenza	SARS-CoV-2
Genome	RNA (retrovirus)	RNA	RNA (segmented)	RNA
Integrates into host DNA	Yes	No	No	No
Curative treatment available	No (rare transplant cases)	Yes	Not applicable	Not applicable
Effective vaccine	No	No	Yes (seasonal)	Yes

The hepatitis C comparison is instructive. Direct-acting antivirals can cure more than 95 percent of people with hepatitis C, according to the WHO hepatitis C fact sheet, usually within 12 to 24 weeks, because hepatitis C never integrates into the host genome and there is no reservoir to defend. HIV has comparably excellent drugs and still no cure, which shows that the difference lies in the virus's architecture rather than in the quality of the medicines. Seasonal influenza, by contrast, is controlled mainly through yearly vaccination, as discussed in our influenza explainer, and avian strains raise separate questions covered in the H5N1 guide.

Common misconceptions

"HIV and AIDS are the same thing." HIV is the virus and AIDS is a late clinical stage. With effective treatment, most people never reach it.

"You can catch HIV from casual contact." The WHO states that HIV is not spread by kissing, hugging, shaking hands, or sharing food or water. The virus needs a direct route into the body through specific fluids.

"Treatment means the virus is gone." Treatment suppresses the virus, but the latent reservoir remains. Stopping therapy generally allows the virus to rebound.

"An undetectable result means the person is cured." Undetectable means below the limit of the test. It is a measure of control, and it holds only while treatment continues.

"HIV is no longer a problem." Treatment has transformed the outlook for people who are diagnosed and engaged in care. The 2025 data still show 1.2 million new infections and 570,000 deaths, and access gaps between regions remain large.

What this means

The HIV story is unusual in medicine because the virus was understood in molecular detail faster than almost any previous pathogen, yet the structural fact that defines it, integration into the host genome, is the one thing drugs cannot undo. The practical outcome is a condition that can be controlled for life with the right drugs, that cannot be passed on sexually while it is controlled, and that can increasingly be prevented with a few injections a year. The remaining problems are mostly about delivery: finding people who do not know their status, keeping people in treatment, and making new prevention tools affordable where infection rates are highest. The scientific frontier is a cure, and the three-decade lesson is that it will probably come from understanding the reservoir rather than from a better antiviral.

Frequently asked questions

Q: Can HIV be transmitted through saliva, sweat, or casual contact? A: No. The WHO lists blood, breast milk, semen, and vaginal secretions as the fluids that transmit HIV, and states that it is not spread by kissing, hugging, shaking hands, or sharing personal objects, food, or water.

Q: Is HIV still a death sentence? A: No. The WHO says treatment controls the virus and allows people to live long, healthy lives. Deaths still occur, with 570,000 AIDS-related deaths estimated for 2025, mostly among people diagnosed late or not on treatment.

Q: Why is there still no HIV vaccine? A: The virus varies enormously between strains, shields its surface proteins, and integrates into the genome early. Those features make it harder for a vaccine-trained response to clear it than for most other pathogens.

Q: What does "viral load" mean? A: It is the number of HIV RNA copies in a milliliter of blood. The CDC describes 200 copies per milliliter or more as a high viral load. Effective treatment pushes the value below the detection limit of the test.

Q: What does U=U mean? A: It means undetectable equals untransmittable. The WHO states that people on ART with an undetectable viral load will not transmit HIV to their sexual partners. The evidence includes PARTNER2, in which 782 male couples recorded zero within-couple transmissions.

Q: What was the Berlin patient case? A: Timothy Ray Brown received a stem cell transplant in 2007 from a donor with the CCR5-delta-32 variant, which blocks a main HIV entry route. He remained free of the virus until his death from cancer in 2020. The procedure is too risky for routine HIV care.

Q: How effective is PrEP? A: In PURPOSE 1, no HIV infections occurred among the 2,134 women who received twice-yearly lenacapavir. Effectiveness of any PrEP depends on use, testing, and the person's circumstances, which is why clinical guidance is individual.

Sources

• WHO. HIV and AIDS fact sheet, published July 27, 2026 (data for 2025).
• UNAIDS. Global HIV statistics fact sheet, data year 2025.
• CDC. Protecting others when living with HIV.
• HIV.gov. HIV and AIDS timeline and PARTNER2 summary.
• Bekker L-G, et al. Twice-yearly lenacapavir or daily F/TAF for HIV prevention in cisgender women. N Engl J Med 2024;391:1179-1192.
• WHO. Hepatitis C fact sheet.
• Siliciano JD, et al. Long-term follow-up studies confirm the stability of the latent reservoir for HIV-1 in resting CD4+ T cells. Nat Med 2003;9:727-728.

For related virus coverage, see our explainers on HPV and cancer and measles.

---

Read next:

• Influenza Virus Explained — How Seasonal Flu Actually Works
• Measles Virus Explained — Why It's Coming Back in 2026
• Dengue Virus Explained — Why It's Spreading and How It Works