Intel’s stock has had a dramatic month. Finally, Intel woke up from the good old days trying to adapt new reality 1: to be a IC Manufacturing as a Service (MaaS), and 2: to sell the slow dying x86 business to Qualcomm. Why is Qualcomm even considering to buy? Will this lead to a worthy challenger to NVIDIA?
Take 1: IC MaaS
A few weeks ago, Intel gained $10B, 11% of market cap because it was considering to spin off the chipmaking business. On Sep. 16, Intel CEO Pat Gelsinger announced a plan to establish Intel Foundry as an independent subsidiary inside of Intel. In 2021, I thought 2 Intel is better than 1. Now may be a bit late, but better than never. Not bad, Intel Foundry’s 1st MaaS win is AWS. Nevertheless, the real CMaaS king maker will still be either fabricating Apple or NVIDIA chips. Only time can tell if Intel Foundry can execute well enough to win them over from TSMC.
Take 2: Fabless IC Design House
On Sep. 20, Intel gained $3B, 3% of market cap because Qualcomm might takeover Intel’s chip-design business as Wall Street Journal reported. Another interesting bet for Intel. Hoping it’s not too late. After all, X86’s beginning of the end has began since 2020. The questions are: Why does Qualcomm want to bet on it? And, how would it work for Qualcomm?
Their last quarter revenues may shade some light.
- NVIDIA had $29B revenue from Data Center & Gaming.
- Intel had $11.8B revenue from Client, Data Center & Network.
- Qualcomm has $8.1B revenue from Mobile, IoT & Automotive.
- AMD had $5B revenue from Data Center, Client & Gaming.
With Qualcomm + Intel to $20B quarterly revenue vs NVIDIA’s $29B, could them be a worthy challenger to NVIDIA?
NVIDIA's Worthy Challengers
AMD and a few more startups are rushing to sell a dream to get a cut of NVIDIA's AI lunch money. Qualcomm must be dreaming to be a big part of it too. Buying Intel's CPU design & getting the current business may just be its ticket to the main table. Or at least, it may accelerate the time to the heavy weight games: data center & personal computing. Those are new to Qualcomm. Why leave it on the table for AMD? Even if both fail, Qualcomm can still avoid other front runners to capture the land to lead even further.
But, the buy strategy is never rosy nor easy naturally. In fact, studies suggests 70-90% of Mergers & Acquisitions failed. The question is how fast Qualcomm may retool all the x86 IPs to make the new kind of ARM SoCs for the current data center and personal computing to be AI ready while serving the legacy compute workflows with the flexibility to adapt whatever AI workflows ahead of us. To be a NVIDIA’s worthy challenger, one would have had to continue to create or acquire disruptive businesses as Christensen etc. suggested.