#MarvellJumps4.5%

Marvell +4.5%, Lumentum +5%, Corning +4%, Arm and Coherent +3%, while SK Hynix, Micron and Broadcom also moved higher. To me, this rally is telling a bigger story than just another AI trade.

The interesting part is that the strength is appearing across different parts of the AI hardware ecosystem at the same time.

For a long time, the AI infrastructure conversation was dominated by the biggest accelerator and compute names. But as AI data centers become larger and models require more computing power, the bottleneck doesn't stop at the processor. You also need high-speed optical connections, networking, memory, data transmission and enormous amounts of supporting infrastructure.

That is why I'm paying attention to this move in names like Marvell, Lumentum, Corning and Coherent.

Marvell, for example, is positioned around data-center connectivity and custom silicon, while its optical networking exposure is becoming increasingly relevant as the industry moves toward faster interconnect speeds. Bank of America recently highlighted multi-year demand for optical networking and tight supply conditions, pointing specifically to Marvell, Lumentum and Coherent as beneficiaries of that trend.

Corning is another interesting piece of the puzzle. Its latest rally was supported by a new agreement with AT&T worth more than $3 billion to supply fiber-optic cable, with the companies pointing to rapidly increasing bandwidth requirements. AI is one of the forces increasing data demand, which means the infrastructure carrying that data becomes increasingly important.

Then you have the memory side.

SK Hynix and Micron are directly exposed to the huge amount of memory required by modern AI systems. Recent market analysis has highlighted how AI infrastructure spending has tightened memory supply and pushed pricing higher, making memory companies an increasingly important part of the AI investment story.

So my take is that this rally makes sense fundamentally, but I wouldn't automatically chase every AI hardware stock just because the whole group is moving together.

That's the important distinction.

A broad rally can mean investors are gaining confidence that AI capital expenditure will remain strong. But it can also mean positioning is becoming crowded. When several companies across the same theme start moving together, expectations can rise very quickly.

And we've already seen how quickly the AI trade can rotate.

Earlier this year, AI-related infrastructure names experienced sharp changes in investor sentiment, with the market becoming more selective about which companies can actually convert AI demand into revenue and margins.

That's why I'm looking beyond the percentage gain.

The question I want answered is:

Is AI spending actually expanding across the entire infrastructure chain, or are investors simply rotating money back into AI hardware?

If hyperscalers continue increasing data-center investment, demand for networking, optics, memory and custom silicon could remain strong. That would give this rally a much stronger fundamental foundation.

But if spending expectations become too aggressive while borrowing costs remain high, the market could become much less forgiving toward companies whose valuations already assume years of AI growth.

This matters even more right now because the broader macro environment isn't exactly easy.

The 30-year Treasury yield has recently moved above 5.6%, while the 10-year yield is around 5.3%, keeping pressure on long-duration growth valuations.

So we're effectively watching two forces at the same time.

AI infrastructure demand is pushing earnings expectations higher.

Higher long-term yields are pushing valuation pressure higher.

That's why I think the next phase of this AI hardware rally will need to be supported by actual results.

For me, the names worth watching aren't necessarily the ones with the biggest one-day percentage move. I'm more interested in companies showing evidence of real AI-related revenue growth, strong orders, improving margins and capacity that is actually being absorbed by customers.

The other thing I'm watching is the shift from compute to connectivity.

As AI clusters become larger, moving data between GPUs, servers and storage becomes a massive engineering problem. Faster optical links and networking aren't optional if the infrastructure is going to scale efficiently.

That means companies supplying the “roads” of the AI economy could become just as important as the companies supplying the “engines.”

And that's probably the biggest takeaway I have from today's rally.

AI is no longer just a chip story.

It is becoming a complete infrastructure cycle:

Compute → Memory → Networking → Optical → Fiber → Power → Data Centers.

If investment keeps spreading across those layers, today's rally could represent something much broader than a one-day bounce.

But I would still keep one rule in mind:

Strong AI demand doesn't automatically make every AI hardware valuation attractive.

The market eventually comes back to revenue, margins, cash flow and expectations.

So my take is constructive on the underlying AI infrastructure demand, but cautious about chasing a broad hardware rally after a sharp move.

I want to see whether the next earnings reports confirm what today's prices are already telling us.

Because if the numbers confirm continued AI spending, then the hardware story becomes much more interesting.

If the numbers disappoint, today's enthusiasm can unwind just as quickly.

The rally is interesting.
The earnings confirmation is what matters next.