Signal Over Noise
A live map of the AI build-out — where the capital is actually flowing, and where the story is getting ahead of the fundamentals.
Global Indices
Supply Chain Heatmap
Today's Focus
Overnight US → Next Session
NVIDIA and Broadcom set the pricing anchor for the whole accelerator complex. A strong session in this pair is one of the more reliable next-session tells for the chip-equipment and test names that supply into the same build-out.
Microsoft, Amazon and Alphabet's capex commentary is the single most-watched signal for the power and cooling infrastructure names that build out the data centres behind that spending.
Supermicro and Dell set the tone for AI-server build volume. Networking and interconnect suppliers into the same racks tend to trade on the same signal a session later.
Supply Chain Map
The AI Hype Trends
The AI Capex Supercycle: Where the Money Is Actually Going
Hyperscaler capex guidance keeps beating estimates, but the mix has shifted hard toward power, land and interconnect. A look at which parts of the chain are absorbing the incremental dollar.
The mix is shifting, not just the total
Two years ago, an incremental hyperscaler capex dollar went overwhelmingly to accelerators. Today a meaningfully larger share is going to power infrastructure, land, cooling and networking — the parts of the stack that don't scale as elastically as chip supply does. That's why 'capex beat' headlines increasingly need a second look at where the dollars landed.
Power is the new bottleneck, and it shows up in lead times
Gas-turbine order backlogs at major manufacturers now stretch past three years in some cases — longer than it takes to build the data centre shell itself. That single fact is doing more to explain deployment timelines than any GPU allocation story, and it's why nuclear, SMR and grid-equipment names have become part of the AI conversation rather than an adjacent one.
Where that leaves the equity picture
The clearest, least-debated read-through is in electrical equipment and liquid cooling, where backlog-to-book ratios are climbing and lead times keep extending across the group. The more contested layer is the neocloud/GPU-rental segment, where headline demand is strong but utilisation and contract-duration questions make the unit economics harder to underwrite from public disclosures alone.