The battle of AI has started shifting from chip design to interconnects and networking.
The excessive focus on Nvidia and AMD is overlooking the other side of the equation: massive GPU computing power is worthless if data center networks (Data Center Networking) are throttling (bottlenecking) the data transfer between thousands of chips.
Today’s HPE (Hewlett Packard Enterprise) story proves that the bet is expanding to infrastructure companies:
What lifted the stock and changed the outlook? 💡
Raising growth expectations: HPE increased the annual compound growth rate (CAGR) for the Networking segment to high-teens through 2029
The data center networking boom: the company expects its Data Center Networking revenue segment to grow by between 50% and 55%+ per year (CAGR) through 2029.
Billion-dollar deals: the announcement comes with a booking of a massive AI infrastructure order worth $1.2 billion

A broader view 👈The first wave of AI was the phase of buying chips (GPUs)
The current and upcoming wave is the phase of installing plumbing and networks (Networking, Switches, Cooling)
Enterprises and Hyperscalers are no longer buying standalone servers; they’re investing in a fully integrated network infrastructure that can handle high AI workloads with near-zero latency
I’m watching HPE and its peers to see whether the flow of cash toward solid infrastructure will ensure sustainable growth, away from volatility in chip prices

Your opinions matter 👇
$HPEB
$HPE