

Gilead Sciences just walked away from Assembly Biosciences' hepatitis B drug, returning a Phase 1b asset and $330 million in potential milestones. With 80+ companies chasing an HBV cure and GSK's bepirovirsen hitting Phase 3 endpoints, the rejection says more about the crowded field than it does about one drug.
Imagine you're the biggest name in liver disease. You've already conquered hepatitis C. You've got a hepatitis B suppression drug on the market. And a small biotech partner hands you a shiny new antiviral with early clinical data.
You pass.
That's exactly what Gilead Sciences just did. The company declined its option on ABI-4334, a Phase 1b hepatitis B drug developed by Assembly Biosciences. The decision, disclosed in Assembly's full-year 2025 earnings release, sends the asset back to Assembly and strips roughly $330 million in potential milestone payments from the collaboration.
Gilead didn't explain why. But in biotech, silence speaks volumes.
To understand the rejection, you need to understand the drug. ABI-4334 is a capsid assembly modulator, or CAM. Think of it like a wrench thrown into a factory's assembly line: it targets a key viral protein that hepatitis B needs to copy itself and build a hidden reservoir called cccDNA.
That reservoir is the whole reason hepatitis B is so hard to cure. Current drugs (like Gilead's own Vemlidy) can suppress the virus, but they can't eliminate the cccDNA hiding in liver cells. Patients stay on treatment for life. A true "functional cure" means clearing the virus's surface antigen (HBsAg) and keeping it gone after stopping therapy. With today's standard drugs, that happens roughly 1% of the time.
ABI-4334 showed it could reduce HBV DNA and several markers linked to cccDNA in lab models. Phase 1a safety data looked clean: mostly mild side effects, no worrisome heart signals. Assembly advanced it into a Phase 1b trial in mid-2024 under the Gilead partnership.
So far, so good. But "so far" is doing a lot of heavy lifting in that sentence.
Gilead hasn't said a word publicly about why it passed. But the clues aren't hard to find.
First, look at what Gilead did keep. In December 2025, the company exercised options on two Assembly herpes simplex virus programs (ABI-5366 and ABI-1179), paying for the privilege. Gilead talked about the firm's focus on "novel antiviral therapeutics" when announcing those deals. The implication: ABI-4334 didn't clear the novelty bar.

Eric Topol, one of medicine's most influential voices on AI and prevention, just signed on as an advisor at Flagship Pioneering. The move signals Flagship's deepening commitment to preemptive medicine, and it tells you a lot about where biotech's smartest money is heading.


Join thousands of biotech professionals who start their day with our free, daily briefing.
Second, Gilead already has a different HBV bet on the table. Its partnership with Hookipa Pharma on HB-400, an immunotherapy approach in Phase 1a/1b testing, represents a fundamentally different strategy. Rather than blocking viral replication (what CAMs do), immunotherapy tries to wake up the patient's immune system to attack the virus directly. And in early 2026, Gilead quietly started a new clinical study on an undisclosed HBV drug. The company isn't giving up on hepatitis B; it's just being picky about how it gets there.
Third, and perhaps most importantly: the neighborhood is getting very crowded.
Over 80 companies are now developing more than 90 programs aimed at a hepatitis B functional cure. That's a staggering number for a disease that many people outside of Asia and sub-Saharan Africa barely think about.
And the frontrunner isn't a CAM. It's bepirovirsen, an antisense therapy from GSK and Ionis Pharmaceuticals. Bepirovirsen just hit primary endpoints in two Phase 3 trials (B-Well 1 and B-Well 2), enrolling over 1,800 patients across 29 countries. GSK is planning global regulatory filings for Q1 2026. If approved, it would be a finite-duration treatment, not a lifelong pill. That's a potential game-changer.
Behind GSK, the field is diverse. AusperBio's AHB-137 hit a 30% functional cure rate in a Phase 2a readout. Precision BioSciences is literally editing cccDNA out of liver cells with a gene therapy called PBGENE-HBV, and it showed the first molecular proof in patient biopsies at a November 2025 conference. China approved its first domestic functional cure drug in October 2025.
Against this backdrop, a Phase 1b CAM with preclinical efficacy data and early safety results looks, well, early. Gilead apparently decided it didn't need another horse in a race where it already has entries and the track is getting packed.
To Assembly's credit, the company is in solid financial shape despite the breakup. It pulled in $72.3 million in collaboration revenue for 2025, narrowed its net loss to just $6.1 million, and is sitting on $248.1 million in cash, enough to fund operations through 2028.
Gilead also still owns nearly 30% of Assembly's equity. The relationship isn't dead; it's just more selective now.
Assembly has already kicked off a formal process to find a new partner for ABI-4334 before pushing it further into the clinic. The company won't advance the program alone. That introduces real execution risk: finding a partner willing to invest in a CAM when the field is shifting toward antisense therapies, immunotherapies, and gene editing isn't going to be easy.
Assembly's stock has been volatile, trading around $27.52 as of March 18. The shares are up 45% year-to-date and 37% over twelve months, but that strength is driven more by the herpes and HDV pipeline than by HBV optimism. Analysts have pointed to the initiation of Phase 2 trials for ABI-5366 in mid-2026 as the next major catalyst.
Gilead passing on ABI-4334 isn't just one company making one decision. It's a signal about where the hepatitis B field is heading.
For years, CAMs were considered a promising pillar of any combination cure strategy. But the competitive landscape has shifted. Antisense oligonucleotides like bepirovirsen are further along and producing real functional cure data in large trials. Gene editing approaches are targeting cccDNA directly. Immunotherapies are trying to solve the problem from the other direction entirely.
A CAM in Phase 1b, no matter how potent in cell culture, faces a differentiation problem. It's not enough to be good; you have to be clearly better than, or complementary to, the approaches already racing ahead.
Gilead's decision sets a high bar for the entire CAM class. If the biggest liver disease company in the world doesn't want your CAM, who does?
Assembly will find out soon enough.
Genentech killed its anti-myostatin antibody emugrobart after Phase 2 trials in two muscle-wasting diseases came up empty. The failure extends a two-decade losing streak for the entire drug class, and it raises pointed questions about whether the obesity pivot investors are banking on will fare any better.