High rate of mutation at the env gene segment of the virus might hinder this approach. 

Evidence is available from experimental animal models that neutralizing antibodies can prevent HIV acquisition, even though no similar data is available on human infection, just as there is also no certainty that these antibodies are effective against fully established HIV infection. However, it is known that HIV-1 has some features that make antibody response rather ineffective, which include high antigenic diversity and ability to evade host's immune responses.    

It is established that neutralizing antibodies could bind to the Gp120 and Gp40 envelop glycoproteins to block interaction with CD4 and chemokine receptors. A second mechanism of antibody response to HIV is through "Antibody-dependent cell cytotoxicity" (ADCC) that leads to killing of infected host cells. Thirdly, phagocytosis and viral inhibition by cytokines, termed as "Antibody-dependent cell-mediated viral inhibition" (ADCVI), leads to reduced viral replication.

The relative contribution of these mechanisms of antibody response are not precisely quantified, and they remain major topics of intense on-going research towards an effective HIV vaccine development. 

Ref:

1. Fauci AS, et al. 2008. HIV Vaccine research: The way forward. Science 321:530-532.

2. Lifson JD, et al. 2011. Lessons in non-human primate models for AIDS research: from minefields to milestones. Cold Spring Harb Prospect Med 10.1101/cshperspect.a007310.

3. Overbought J, et al. 2012. The antibody response against HIV-1. Cold Spring Harb Prospect Med. 10.1101/cshperspect.a007039. 

Lab evidence often cannot be confirmed in living people, especially whether HIV is concerned. 

True, HIV is highly diverse. HIV is one of the most diverse viruses that infect human population. There are 2 types (HIV-1, HIV-2), 4 HIV-1 groups (M, N, O, P), and several group M sub-types (A, B, C, D, F, G, H, J and K.). Although not quite new, more often we are speaking now about circulating recombinant form (CRFs), recombinant forms achieving epidemiological relevance. Recombinant forms are described since more than two decades, and there is a quite large database with published CRFs. CRFs are more frequent but not limited to IVDUs and other HIV infected individuals.

Fact. The HIV diversity develops extremely fast. HIV is a Lentivirus which aims to stay as long as possible in a living host, thus the numerous and diverse evading mechanisms developed by HIV.  Plus the fact that HIV not only escape the immune response, but also controls it in various ways.

Let's remember that HIV is the only "diploid" virus, packaging 2 copies of the same RNA genome in a single virion. Let's also remember that its life cycle involves a reverse-transcription into a cDNA which ultimately randomly integrates into the human genome with the aid of newly synthesized viraly encoded LTRs. RT is a most unspecific known enzyme (as opposed to CK, the most specific one). It introduces lots of errors during readings. A decent percentage of errors leads to diversity. Too many errors lead to non-viable virions. To prevent non-viability, HIV-RT is using a replication mechanism known as "copy choice", i.e. can jump from 1 RNA copy to the other in order to keep the balance in favor of a more accurate reverse-transcription.

Being so diverse already and counting on, a pan-vaccine against HIV is a dream. That's why the latest efforts a focusing on specific geographical areas where some strains are more prevalent (see HVTN 701 project for Sub-Saharan Africa, which is expected to finalize in 2022). Testing safety and efficacy takes long time in clinical trials, whilst the HIV does not sleep.

This is the vaccine situation, in which HIV infection prevention is the goal. It is also known that people (about 1% of Caucasian people) with two copies of the CCR5 delta32 gene (inherited from both parents) are virtually immune to HIV infection. Maybe working with this could lead to a HIV vaccine. 

Speaking about HIV neutralization after the infection occurred is another utopia, having in mind all those unimaginable interactions of this particular virus with its host: concealing the CD4 receptors after TCD4+ cell infection, the ability to infect T-cell precursors (and thus CD8+ cells too), the co-receptors which not only direct but also extend the virus tropism, the virus sanctuaries, the super-antigenic mechanism that freezes viable T cells, syncitia formation which depletes several healthy lymphocytes starting from a single infected one, etc. Last but not least, the molecular mimicry which really allow HIV to control lots of immune processes in the host, opening the "door" for opportunistic infections and other disease conditions...

In my opinion, limiting HIV does not rely on a vaccine. It relies on a clean life from uncontrolled sex anywhere everywhere with everybody, drug use, promiscuity, etc. EDUCATION. HEALTHCARE.