Steam Client vs Epic Games Launcher V2: Steam's Post-Patch Startup Time Wins a Contested Benchmark

Within 24 hours of Epic Games launching a preview of its AI-assisted "vibe-coded" Epic Games Launcher V2 with a marketing pitch aimed at Steam's slow startup, Valve shipped a small Steam client update that a gaming-news report measured as faster on cold boot. The claim rests on a two-row benchmark table published by the gaming site RespawnFirst, and Hacker News commenters disputed both the numbers and the framing almost immediately.

What happened, in order

Epic Games announced and launched in preview a new launcher, Epic Games Launcher V2, whose main marketing pitch was a direct jab at Steam's loading speed. Roughly 24 hours later, Valve pushed a small Steam client update. According to RespawnFirst, that update made the Steam client boot faster than Epic's new launcher, neutralizing Epic's primary performance claim.

The report frames the outcome as a contrast between two engineering approaches: Epic "vibe coded" a faster client, while Valve applied targeted optimization to a decades-old ecosystem. The report's own conclusion is that Valve "proved that targeted optimization on a decades-old ecosystem can still beat out AI speedruns."

The benchmark numbers, and what they do and do not show

The report published this comparison table:

Client version	Startup time (cold boot)	Tray restore speed	Architecture
Epic Games Launcher V2 (Preview)	~5.47 seconds	~0.36 seconds	AI-assisted ("vibe coded")
Steam Client (post-patch)	~5.23 seconds	~0.42 seconds	Optimized modular CEF

Two details matter more than the headline. First, Steam's reported win is on cold boot only; Epic's launcher is still reported as faster at restoring from the system tray (0.36 seconds versus 0.42 seconds). Second, the report describes the test hardware only as "basic test systems" and does not state sample size, number of runs, or methodology, so the 0.24-second cold-boot gap is not independently verifiable from the article alone.

The report also notes that Epic Games Launcher V2 remains "a massive step forward compared to its notoriously sluggish legacy build," even if it lost the head-to-head.

Commenters disputed the numbers on the day of publication

Hacker News readers who ran their own measurements did not reproduce the article's figures. One commenter, tancop, reported the opposite ordering for tray restore and a much slower Steam cold boot:

I used to think vibe coding would get us slow Electron apps but now we got speed wars between game launchers. I hope this comes for all types of software. With that said, I tested it right now and Steam takes about 2 seconds to restore from tray and 10-15 for cold boot. The new Epic launcher is around 0.5 and 5 like the article mentioned. Maybe it's still in A/B testing or something.

Another commenter, himata4113, said they reverse-engineered the Epic client and concluded the speedup is partly a perception trick rather than a pure optimization:

I had suspected that they simply invoked client ready faster and let misc information fetch in background. After a little bit of reversing they did infact just make the client spawn earlier. Steam client was always fast, majority of load time is (and forever will be) network speed.

A recurring objection was that a launcher taking roughly five seconds to start is not a result worth bragging about in either direction. As one commenter put it, "App to launch apps is loading in little over 5s, on what I assume is high-end gaming pc. Truly something to be proud of.." Another compared it to a 1998 first-person shooter: "it takes longer to start the steam client than it does to load and start the first level in half life."

The feature gap between the two launchers

The RespawnFirst article argues that Epic's problems run deeper than startup time, listing features the Epic launcher lacks that Steam has had for years: discussion forums and community hubs, user-written text reviews, community-created guides and walkthroughs, screenshot/clip/artwork sharing, community groups and curator recommendations, profile customization (badges, showcases, backgrounds), a community marketplace for in-game items, direct player-to-player item trading, digital trading cards and a points shop, universal controller remapping, Big Picture mode for TVs and handheld PCs, Remote Play and Remote Play Together, integrated broadcasting tools, family library sharing, a platform-wide mod hub equivalent to Steam Workshop, personalized discovery queues, local network game file transfers, and game version rollback options.

Not every commenter treated that list as a pure indictment. One argued that several of the "missing" features are things other platforms would be better off without:

All I can think with this whole situation is how much time is wasted by people getting angry and defensive on behalf of a Valve. It's almost like people have a parasocial relationship with the application itself. That's not even getting to some of the "missing" features. No Community marketplace for in-game items No Direct player-to-player item trading No Digital trading cards, levelling, and points shop I hope more application developers "forget" to add these features.

Others read the absence of community features as a deliberate product decision rather than a resource constraint: "I can't believe they don't have the resource to bring these features to their launcher, they actively refuse to make it a tool for communities. It must be a way to buy game and launch them, that's all."

The "vibe coding" framing drew the sharpest pushback

The article's central narrative — that hand optimization beat AI-assisted development — was the most contested claim in the thread. The most-upvoted counterargument was that AI is not a meaningful differentiator for a task as narrow as startup optimization:

Optimizing a project with AI is one of the most trivial things you can do with AI. I also wouldn't buy it for a second that Valve engineers aren't using AI, especially for something as trivial as "hey claude rank the top causes of slow startup" -> "oh okay now fix". The most I'd grant is that everyone uses AI secretly as a sort of dark forest problem.

That view was echoed by several others, including a commenter who wrote simply, "I don't know why anyone's assuming Valve isn't 'vibe coding' too," and another who noted the strategic problem with Epic's marketing: "The problem with advertising 'vibe coding' is that anyone can do it."

A separate thread of the discussion focused on what aggressive AI-driven optimization actually produces. One commenter described the failure mode:

When you vibe code and tell your agent to "keep working till everything loads inside 100 milliseconds", you end up with a super hideous codebase which does insane things like having a custom memory allocator, handwritten assembly, etc etc. It is fast though!

Another raised the question of whether AI can be directed at efficiency as a first-class goal: "What are the possibilities of AI learning how to code efficient apps, e.g. low memory apps, or faster apps, or both?"

Why Steam did not fix this sooner

The most common criticism in the thread was not aimed at Epic but at Valve. The Steam client's slowness was widely described as a long-standing, fixable problem that Valve simply chose not to prioritize until a competitor made it a marketing issue.

It is incredibly funny that Steam could have fixed this at any time, but didn't prioritise doing so until Epic openly challenged them. This is why we can't have nice things™

A blunter version of the same point: "Ok but why didn't they do it sooner? The Steam client is terrible and there is no excuse for it." And a one-line summary of the whole episode that got traction: "Public shaming works."

One long comment generalized the lesson to the entire software industry, arguing that performance is a choice that teams routinely decline to make:

This is why software is slow and everything uses 2GB of RAM at idle: because nobody really cares. [...] But either of these companies, presumably with dev teams of dozens-hundreds of SWEs, could have pointed to one of them 5 years go and said "spend the next 2 months optimizing the performance of the launcher" and would probably have gotten similar results. I make my software fast and efficient mostly because I enjoy it. Absolutely nothing stopping web apps, Electron, everything else from being efficient, slim, and performant.

The competitive effect is the part commenters agreed on

Where the thread found consensus was that the rivalry itself benefits users, regardless of who wins a given benchmark. "Thank you Epic for improving Steam! Appreciated," one commenter wrote. Another framed it as the point of competition: "And that's why having some competition is useful for consumers. I would like to see this trend to continue and game launchers improving not only startup times, but RAM usage and disk size usage."

That commenter also supplied a concrete data point on disk usage: their .steam directory occupied over 10GB with only a single installed game of roughly 500MB. Others hoped the pressure would push Epic toward feature parity rather than exclusivity deals: "If this is the start of epic competing on features and quality for both devs and players instead of just buying up exclusive rights then I'm all for it."

A broader prediction was that AI-assisted development lowers the cost of catching up, which raises competitive pressure across the board: "I think vibe coding will force everyone to stay on their toes and keep being at the edge, moat defences have lowered and competition is up, as consumers we will get much better products for much cheaper."

Alternatives, side notes, and unresolved details

• A commenter recommended Rare, an open-source alternative Epic Games Launcher client, describing it as "much more lightweight and fast" after a couple of years of use.
• One commenter said Epic's store is unappealing enough that they prefer paying for a game on Steam over getting it free on Epic.
• At least one reader could not access the linked article at all, reporting a "Your country is not allowed to access this resource" message and asking what the site's issue with South Africa is.
• Several commenters criticized the tone of the source article itself, including its use of a Gabe Newell image, calling it "kind of obnoxious" and describing the whole thing as "a D-measuring contest."
• One commenter noted that Steam's reported speedup came from optimizing CEF (Chromium Embedded Framework), the browser engine underpinning the client, and linked the classic "Speedup Loop" article about repeatedly finding a zero to dock from an optimization.

What is and is not established

The verified facts are narrow: Epic launched a preview of a new launcher and marketed it on speed; Valve shipped a Steam client update within roughly a day; and one gaming-news outlet measured Steam as faster on cold boot and slower on tray restore. The benchmark methodology, hardware, and sample size behind those numbers are not documented in the source, and at least one reader's independent test produced materially different results. Whether AI assistance was decisive in either client's performance, and whether the Steam update was a direct response to Epic's launch, are interpretations rather than established facts.