Anyone who was hoping that PCIe 6.0 would soon give the mainstream market a boost can safely put their optimism in the drawer next to “Windows 10 will stay forever”. According to Wallace Kou, CEO of Silicon Motion, nothing will happen in the next five years. This may sound sad, but from a technical and economic perspective, it is unfortunately frighteningly logical.
Performance is not an end in itself – at least not in the retail market
Let’s start with the harsh reality: PCIe 5.0 has almost doubled the transfer performance compared to Gen 4 – we’re talking about up to 14 GB/s at the peak. Sounds impressive, but unfortunately the average user will notice very little. The difference between a Gen 4 and Gen 5 SSD in everyday use? Measurable with a stopwatch, but not with experience. And here it comes: PCIe 6.0 is supposed to double that again – to 64 GT/s per lane. Sounds like Formula 1 on the data highway. Except that in reality, this highway on consumer mainboards can be a maximum of 3 inches long before the signal dies. Anyone who now brings retimers and complex signal reclocking chips into play can prepare themselves for board prices that look like the Apple Mac Pro: absurd.
Wallace Kou delivers the death blow
The CEO of Silicon Motion says what many think but no one wants to say out loud: “There is simply no need for PCIe 6.0 in the consumer sector.” Period. And the statement doesn’t come from some analyst on LinkedIn, but from the man whose company dominates the market for SSD controllers. Silicon Motion doesn’t have a single Gen 6 controller in the pipeline – because they simply don’t have to. There is no demand, the ROI is negative. And this is not a coincidence, but a calculation: according to Kou, the tape-out costs for a PCIe 6.0 controller are between 30 and 40 million US dollars. In comparison: PCIe 5.0 costs “only” 16 to 20 million dollars. In other words, almost double the cost for a little more throughput – a total economic loss when you consider the margin of consumer SSDs.
The PC market on valium
While the enterprise sector (thanks to Rubin, Blackwell and co.) is at least symbolically flying the Gen 6 flag, the retail market is in a kind of “digital slumber”. Manufacturers continue to install Gen 4 SSDs because they are cheap and fast enough. Gen-5 is already a luxury that many boards and PSUs can’t even use sensibly. Gen-6? That’s talk of the future that nobody wants to hear – because it costs more than it brings. Even more absurd, most users operate SSDs at the edge of their thermal load capacity – and then PCIe 6.0 is supposed to punch through the line with even more watts? The next “feature” would probably be an integrated hot air blower on the board.
Conclusion: Controlled standstill with an announcement
PCIe 6.0 will come – but not before 2030 and definitely not for normal users. Until then, Gen 5 is here to stay. Not because it’s the best, but because it’s the most economical. In times of consumer restraint, rising production costs and geopolitical uncertainties, that may not be sexy – but it’s damn sensible. And perhaps that’s the bitterest part of the whole story: that we could have been further ahead technologically long ago. But who wants to spend 40 million dollars on a feature that nobody uses?
Source: Tom’s Hardware
18 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Veteran
Mitglied
Urgestein
Urgestein
Mitglied
Urgestein
Mitglied
Urgestein
Urgestein
Urgestein
Mitglied
Urgestein
Urgestein
Urgestein
Urgestein
Veteran
Urgestein
Alle Kommentare lesen unter igor´sLAB Community →