2015/10/15 · Half-Life · 15 min read

Don't Wait for Half-Life 3 — Valve Already Answered the Question

Half-Life 2 shipped in 2004. Episode Two ended on a cliffhanger in 2007. Two decades later, Valve has shipped neither Half-Life 3 nor Episode Three. The reading below is that "Valve will not finish the franchise the way fans want" is not a missing-sequel problem — it is a business choice, and Half-Life: Alyx (2020) was Valve's actual answer.

The Half-Life 3 question has, by now, a longer history than most game franchises have entries. Half-Life released in November 1998. Half-Life 2 followed in November 2004. Episode One in 2006 and Episode Two in 2007 were structured as the first two installments of a three-part epilogue — the sentence "the third episode is coming" was a publicly held position at Valve as late as 2008. None of the three planned formats — Half-Life 3, Episode Three, or the cancelled internal project Return to Ravenholm — ever shipped under those names.

What did ship, in March 2020, was Half-Life: Alyx — a full-length, full-budget, VR-exclusive entry set between Half-Life and Half-Life 2. Alyx reviewed at 93 on Metacritic, won Game of the Year at multiple ceremonies, and provided a canonical sequel to Episode Two in its closing scene — without using either of the sequel titles fans had been waiting for.

The 2015-era version of this argument — that Half-Life 3 simply wasn't coming — held for the next half-decade. The 2020 release of Alyx updated the argument rather than refuted it. The honest 2026 reading is that Valve has made its statement on the franchise, that the statement was Alyx, and that the question "when will Valve ship Half-Life 3?" is best understood as the wrong question.

The Cliffhanger That Held for Twenty Years

Episode Two's closing minutes are explicit. Eli Vance is killed by an advisor. Gordon Freeman and Alyx Vance are left in the snow with the unopened Borealis question — the ship containing Aperture Science technology that the Combine would weaponize and the Resistance is therefore trying to scuttle. Eli's last words ("Prepare for unforeseen consequences") repeat the G-Man's framing. The setup for Episode Three could not be more pointed: an artifact retrieval that would also be a grief plot, with the franchise's two strongest characters reckoning with the loss of the third.

What we know about the would-be Episode Three is unusually well-documented for a cancelled game. Marc Laidlaw, the original Half-Life writer who left Valve in 2016, published a "lost epilogue" titled Epistle 3 in August 2017 on his personal site. The document is a deliberately encoded version of the Episode Three plot — character names replaced (Alyx becomes "Alex", Gordon becomes "Gertrude"), but the structure transparent. The Borealis is reached. The mission goes wrong. Eli's death is avenged at structural cost. The franchise's mythology is partially resolved.

Laidlaw's publication was, in practice, a release valve. The community accepted that Episode Three in roughly that form would not be coming, that the canonical resolution would be either in Half-Life 3 or in some other product, and the original urgency around the cliffhanger softened. The "Half-Life 3 confirmed" meme, which had run on every social-media platform for nearly a decade, hit a natural inflection point.

The hl3.txt File and Other Engine Leftovers

The recurring "evidence" for an in-development Half-Life 3 through the mid-2010s was, in retrospect, almost entirely engine artifacts misread as confirmations. Valve's Source 2 engine inherited file-naming conventions from Source 1, which inherited them from Half-Life 1's Goldsource. The file hl3.txt found in Dota 2 Reborn's 2015 build was, by most plausible readings, a placeholder for engine subsystem documentation rather than a leak from an active game project.

This matters because the "hl3.txt is real evidence" reading was load-bearing for fan speculation through 2015-2017. The file contained references to procedural generation, citizen AI, VR support, and zipline mechanics. The 2016 reading was that this combination suggested a hybrid open-world / VR-enhanced Half-Life sequel. The 2020 reading is that the VR-support references were for Alyx, the citizen AI was the engine-level NPC framework being rebuilt for Source 2, the procedural-generation hooks were for Dota 2's in-game tutorial, and the zipline mechanic was for the Half-Life: Alyx environment toolkit.

In other words: hl3.txt was real, but it was not what fans read into it. It was a manifest of Valve's engine-development priorities for a Source 2 release that, when it shipped in 2020, did so as Alyx rather than as the sequel everyone expected.

Why Valve Didn't Make the Sequel

The strategic case against Valve shipping Half-Life 3 has three components, each of which the 2015 reading anticipated and the 2026 reading confirms.

1. The Financial Math

Valve's business model since the mid-2000s has been the Steam platform, not first-party development. Steam's revenue share (typically 30% on standard publishers, less for high-volume titles) on global PC software sales is, at industry estimates, in the high single-digit billions of dollars annually. The platform's monthly active user base grew from ~50 million in 2014 to ~130 million by 2024. Steam's contribution to Valve's revenue dominates any plausible game release.

A Half-Life 3 at the scale audiences expect would require a five-to-seven-year development cycle and a $100-200 million budget. Even an optimistic 15-million-unit lifetime sales projection at $60 retail nets ~$540 million gross to Valve (after platform fees that Valve would notionally not have to pay to itself), minus development cost. That's a healthy single-product return. It is also a fraction of one good Steam-platform quarter. The opportunity cost of allocating Valve's most senior engineering and design talent to a single game project for half a decade is higher than the game's own return justifies.

This is the simplest reading and the one Gabe Newell has repeatedly hinted at in interviews: Valve doesn't ship Half-Life 3 because the company's actual business is somewhere else, and the talent that would build the game has more valuable things to do.

2. The Creative Problem

The second reason is the harder one. Half-Life 2's critical success rested on a specific combination of innovations — the gravity gun, Source's physics simulation, the tightly cinematic linear pacing, the environmental storytelling that replaced cutscenes. Each of those was a generational step forward in 2004. Building a sequel that matched the step requires identifying a new technology that the franchise can be the first to articulate.

By the mid-2010s the candidates for that new technology were narrow: open-world traversal (which would have violated the franchise's linear-cinematic identity), procedural generation (same problem), and VR. Of those, only VR offered a step that would let the franchise be first into a clearly defined new space. Valve's research investments through 2013-2018 — the SteamVR platform, the Vive partnership with HTC, the Index headset, the Knuckles controller — were the technical scaffolding for the version of Half-Life that could justify the step requirement. That version was Alyx.

Building a flat-screen Half-Life 3 in 2020 would have meant shipping a game that was technically and creatively a follow-up rather than a step. Valve's apparent position was that the franchise did not survive a follow-up. The franchise survived a new statement.

3. The Organizational Structure

Valve's internal organization is structurally different from a conventional game studio. The company's "flat" management model — documented in the leaked 2012 employee handbook — places staff in self-organizing project teams that they can join or leave at will. The historical pattern is that projects without active staff interest stop progressing.

Half-Life 3, by multiple ex-Valve accounts, started and stopped at least four times between 2007 and 2017. The largest of those attempts, internally known as "Source 2 RPG", was a hybrid Half-Life/RPG project that never coalesced into a single shippable form. The flat structure made it easy for staff to drift to other projects (Dota 2, Artifact, the Steam platform itself, the SteamVR R&D) when momentum stalled. There was never an executive who could force the project to ship.

Alyx succeeded inside this structure because it had a clearly defined platform requirement (Valve Index hardware) and a relatively contained scope (a single five-to-six-hour campaign at full polish). The organizational lessons of the previous decade were applied to a smaller-bore project. The same approach was not viable for a larger Half-Life 3.

Why Alyx Is the Sequel

Half-Life: Alyx released for SteamVR on 23 March 2020. The game runs ~12 hours, is set in 2008 (between Half-Life and the start of Half-Life 2), and stars Alyx Vance in the years before Gordon Freeman's reappearance. The narrative establishes the Combine's full occupation of City 17, builds the resistance network around the Vance family, and resolves several long-running franchise threads.

The closing scene is the part that matters for the "is this really a sequel" question. Alyx ends with the G-Man returning Gordon Freeman to a moment of his choice — and explicitly overwrites the closing moments of Episode Two. Eli Vance does not die. The advisor is killed. The G-Man's framing is updated. The Borealis subplot remains unresolved but is positioned for a future entry.

This is not "Episode Three by a different name", but it is a canonical continuation that resolves Episode Two's ending and sets up whatever the next entry would be. Valve has, in effect, said: the franchise is not over, the cliffhanger has been retconned, the future of the story is in flight, and the next entry will arrive when there is another step worth taking. The 2020 statement is consistent with the 2015 reading — Valve is not going to ship a numbered Half-Life 3 — but it adds the part that 2015 could not anticipate, which is that VR was the form Valve was waiting for.

Comparable Catalog Decisions

Valve's strategy on Half-Life is the inverse of the strategies that Konami and Ubisoft have taken with their catalogs.

Konami's response to its catalog question was to monetize the IP through licensing into pachinko while letting the franchises themselves go dormant. Ubisoft's response was to retrofit a shared continuity across an enormous open-world catalog, leaning into the appearance of cohesion. Valve's response was to do neither — to refuse to extend the franchise on any schedule that did not match a creative step forward, and to invest the corporate surplus from Steam in the technology that would eventually let the franchise be extended on Valve's terms.

The Nintendo comparison is the cleanest. Nintendo, faced with the question of where Zelda could go after Skyward Sword, built the open-air systemic-traversal model and shipped Breath of the Wild. The timeline-fitting controversy that resulted is the cost of the choice — but the franchise survived and remained creatively forward. Valve made a structurally similar choice on a longer timeline. The cost is twenty years of audience anticipation. The result is the franchise still being capable of generating a Game of the Year award when it does ship.

Both decisions have the same root: the franchise's value is preserved by allocating it to step releases rather than to incremental sequels. That value-preservation logic is invisible to audiences who measure the franchise by release cadence. It is visible only when the next step actually arrives.

What "Half-Life 3" Actually Means Now

The phrase has, by 2026, three plausible meanings, and the answer to "will Valve ship Half-Life 3?" depends on which meaning is meant.

Meaning 1 — A flat-screen sequel to Half-Life 2, in the franchise's original mode, with a number 3 on the cover. No. This was not coming in 2015, was not coming in 2020, and is not coming in 2026. Every available signal about Valve's allocation of senior engineering effort, every Newell interview, every leak from ex-staff points the same direction: this product is not being built and is not on the roadmap.

Meaning 2 — A continuation of the canon that resolves the post-Episode Two Borealis arc. Almost certainly yes, eventually. Alyx set up the narrative scaffolding for this and Valve has indicated (with a typical lack of commitment) that Alyx's success is "encouraging" for future Half-Life work. Whether that work ships in 2027 or 2032 is unknown.

Meaning 3 — A return to making Half-Life flat-screen the way it was. No. Valve's signaled position is that any future Half-Life entry will exploit a new platform capability — VR continuing, or some hybrid form involving the Steam Deck, or whatever the next-generation hardware platform turns out to be. The flat-screen-first mode is not on the table.

The audience question that has actually been worth asking for the last decade is not "when will Valve ship Half-Life 3?" but "what platform does Valve need to be the first into for the next Half-Life to make sense?" The 2020 answer was VR. The 2026 answer is open. The franchise is not dead and Valve has not abandoned it — but it is going to continue to be released on Valve's preferred cadence, which is generational steps rather than annual sequels.

Closing

The 2015 reading of the Half-Life 3 question concluded with "don't wait." That conclusion has aged well. The 2020 release of Alyx updated the reading without overturning it. Audiences who internalized "don't wait" in 2015 got the better deal: they accepted that the franchise was structurally not going to ship the sequel they wanted, and they got a five-year head start on the realization that Valve was building toward something different.

The honest position in 2026 is: the franchise is alive, Valve has shown that it can ship Half-Life when the technology and the team align, and the next entry will arrive on a timeline measured in platform generations rather than years. Anyone still waiting for Half-Life 3 specifically is waiting for a product that was never on Valve's roadmap. Anyone who absorbed Alyx as the answer it actually is is, in 2026, in a less uncomfortable position than the meme-driven Half-Life 3 community of the 2010s ever managed to occupy.

Frequently Asked Questions