Previously, the earnings reports from Google and Amazon both indicated that they will continue to invest heavily in the construction of AI data centers in the future.

Whether it's NVIDIA's cards or the ones they build themselves. To unleash the full power of the cards, they ultimately rely on key auxiliary components such as storage, liquid cooling, power management, and communication connectivity.

Yesterday, VRT's earnings report actually wasn't that strong, it opened high at a gap of 20 points.

ALAB's earnings report wasn't bad, but it corrected by 20 points. After the correction, its valuation has reached a relatively low level in ALAB's historical valuation range.

The core reason ALAB can maintain a gross margin as high as 75% lies in its absolute leadership in PCIe 6.0 technology. ALAB is the world's first and currently the only company to achieve mass production of PCIe 6.0 fabric switch chips (Scorpio X). In the AI race, "speed" is the moat. While giants like Broadcom are still in the testing phase, ALAB has already secured rack orders from giants like Amazon (AWS) for 2026.

ALAB's lead isn't just at the physical layer. The application of its CXL (Compute Express Link) technology allows CPUs and GPUs to share a memory pool. This "memory consistency" is the only way to achieve inference for ultra-large-scale models.

Therefore, its early binding with Amazon, although it comes at a small equity cost, this cost is exchanged for the giant's recognition and early capacity reservations. I believe this is not a negative, but rather a positive.

I'm considering establishing a watchlist position.

$Astera Labs(ALAB.US)$Vertiv(VRT.US)$Amazon(AMZN.US)