NVIDIA + Corning. 10x US fiber manufacturing expansion. 3 new factories. 3,000 jobs. The optical interconnect layer that hyperscale treats as commodity infrastructure is about to become the critical s
The 1999 comparison is the right one to stress-test. The key difference I'd argue: that cycle was built on projected demand. This one is being built on signed contracts and committed capex. The fiber laying isn't ahead of the revenue — it's chasing infrastructure that's already funded. The bust risk now lives higher in the stack, at the application layer, not in the physical substrate. But you're right that 'this time is different' is always the most expensive sentence in investing.
Data centers gave Corning its next life — though the last time fiber demand looked this certain, in 1999, the picture unfolded poorly.
The 1999 comparison is the right one to stress-test. The key difference I'd argue: that cycle was built on projected demand. This one is being built on signed contracts and committed capex. The fiber laying isn't ahead of the revenue — it's chasing infrastructure that's already funded. The bust risk now lives higher in the stack, at the application layer, not in the physical substrate. But you're right that 'this time is different' is always the most expensive sentence in investing.