You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Token demand is skyrocketing and the need for AI compute continues to accelerate. The improvement in model capabilities combined with the rapid emergence of agentic workflows has driven a surge in user adoption and aggregate token demand. Anthropic added a staggering $6B of ARR in the single month of February alone driven by broad adoption of agentic coding platform Claude Code, and if Anthropic had more compute they would have added more. Despite a huge AI infrastructure buildout over the past few years, available compute is scarce. On-demand GPU prices continue to go up even for Hoppers which are almost 2 generations old.
From our own experiences, we have reached out to every neocloud we know asking if they have small clusters available, but everything is already firmly locked up. This tight supply environment explains the sharp reset in hyperscaler capex plans. Consensus estimates have moved materially higher across the board, with Google standing out as the most extreme example, where 2026 capex expectations have roughly doubled versus prior expectations, primarily driven by datacenter and server spend.
I find curious the two charts on TSMC N3 wafer capacity, and N3 wafers shipped: They show N3 wafer capacity decreasing QoQ in Q4 2024 through Q3 2025, but N3 wafers shipped somehow increases quarterly each time.
..
That aside, I found it insightful that switching memory DRAM to HBM has a 3:1/4:1 impact on wafer used per bit.
the capex forecast tells it all. That is why we have a shortage.
Reminder... HBM isnt the only thing short and DDR5 prices and margins have exploaded. MOST of Memory company increase in revenue is coming from DDR5 price increases, not just HBM
newsletter.semianalysis.com
