Betting 8 months ahead, the "timeline game" in the semiconductor track has already begun.



Serenity shared a very interesting perspective—markets typically price in themes 8 months in advance, except for things too futuristic like space, robotics, and quantum.

Following this timeline, what should we focus on right now?

Memory is the strongest narrative at the moment. HBM3e is already in mass production at Micron, Samsung, and SK Hynix, and HBM4 is nearing production. These companies are still in a "middle bottleneck" phase—the supply-demand gap won't close in the short term, and orders extending into next year are highly likely.

Photonics is the next story. 1.6T optical modules are already known, CPO Scale Out won't arrive until the second half of 2026, NPO from the second half of 2026 to 2027, and Scale Up even later—the second half of 2027. The timeline is long enough, the expectations fresh enough—the market loves this.

Simply put, this semiconductor cycle is not a broad rally but a structural divergence. Whoever has tight capacity and high barriers to entry will enjoy a valuation premium. With an 8-month lead time, by the time market consensus forms, prices may already be at the halfway point.

Memory, Neoclouds, Photonics, Robotics—if you don't do your homework on these tracks now, you'll only be chasing highs by year-end.
#BTC下探60000美元关键关口 #以太坊基金会重组降本 #美光财报超预期盘后大涨
BTC-1.88%
ETH-1.60%
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 3
  • Repost
  • Share
Comment
Add a comment
Add a comment
0xSideQuest
· 31m ago
8 months, this number is quite interesting. Serenity previously mentioned a similar framework. This wave of storage hasn't been fully priced in yet. The HBM production capacity bottleneck is harder than imagined.
View OriginalReply0
On-ChainNightSecurityGuard
· 2h ago
Structural differentiation is well said; the broad rally ended long ago. Now it's about who can accurately calculate production capacity and barriers. I still need to brush up on the Neoclouds side.
View OriginalReply0
LatencyMonk
· 2h ago
The photonics timeline is stretched long enough—CPO Scale Out is expected to reach the second half of 2026. Staying put now is early, but not too early; once it becomes clear to everyone, there won’t be any alpha left.
View OriginalReply0
  • Pinned