After more than a decade of massive growth in cloud spending, it's clear that data centers have become a modern computing unit staple, powering all sorts of basic services and future tech developments. With public and private cloud spending expected to grow on average by double-digit percentages every year in the decade ahead -- in which time it could become a $1 trillion-a-year industry -- businesses are scrambling to capitalize on the opportunity.

To that end, International Business Machines (IBM 1.72%) recently announced a divestiture to help it focus on this key market, and reports indicate Advanced Micro Devices (AMD 2.12%) is close to announcing a deal to acquire fellow chip designer Xilinx (XLNX). Both moves will help the companies focus on data centers and cloud computing, and will also greatly improve profitability. 

A cloud surrounded by computers, illustrating a data center.

Image source: Getty Images.

A leaner, meaner legacy tech giant

IBM completed its acquisition of cloud-based application development and management outfit Red Hat in 2019. The takeover was billed as transformational for IBM as it signaled its shift to emphasizing data center and cloud computing over its legacy hardware lineup. While Red Hat most certainly updated IBM's cloud capabilities, it thus far hasn't had the same desired effect on the financial front. 

In Q2 2020, IBM's revenue fell 1.9% year over year (when adjusting for one-time items affecting comparability) to $18.1 billion, and free cash flow (revenue less cash operating and capital expenses) was down 4.2% to $2.23 billion. The company has also released preliminary third-quarter results, with an expected $17.6 billion in revenue implying another 2% decline from last year.

To IBM's credit, its cloud business, driven by Red Hat results, remains in double-digit percentage growth mode, but it is weighed down by declining sales in its legacy businesses. To fix the problem, IBM will be spinning off its managed infrastructure segment into a new company (currently billed as "NewCo") by the end of 2021. Management expects NewCo will haul in about $19 billion a year and the new and improved IBM focused solely on the growing hybrid cloud opportunity about $59 billion a year.  

NewCo will be focused on helping customers update their network hardware and infrastructure -- a big market estimated at some $500 billion a year in global spending by IBM. But although the infrastructure hardware and management segment remains a necessary component of the tech industry, this is a mature business with low profit margins and limited growth potential. NewCo will focus on generating operating efficiency. For reference, IBM's Global Technology Services segment (home to the bulk of the managed infrastructure assets to be spun into NewCo) generated a gross profit margin of just 34% on revenue of $6.32 billion in Q2. By comparison, the cloud and cognitive software segment generated a gross profit margin of 77% on sales of $5.75 billion.

In a nutshell, by the end of next year, the new IBM will be a far leaner machine focused on a high-flying tech market and will generate far higher profit margins than it is now to help it continue on its growth journey. This could be a top cloud stock to own ahead of the divestiture. 

AMD goes big to capture more data center pie

On the other end of the spectrum is AMD, which is riding sales momentum higher with its next-gen chip designs for PCs and data centers. Long the underdog to larger rivals Intel (INTC -0.43%) and NVIDIA (NVDA 1.24%), AMD has made great strides in the last few years with its semiconductors and it has especially been capturing market share from "chipzilla" Intel. 

In Q2 2020, AMD's revenue grew 26% year-over-year to $1.93 billion, building on the 40% gain notched in the first quarter and the 4% increase in full-year 2019. For all of its advances, though, AMD remains a far less profitable concern than its larger peers. Its Q2 gross profit margin was up to 44% compared to 41% last year, but over the last 12-month stretch, gross margin of 44% lags far behind margins of 63% at NVIDIA and 58% at Intel. Likewise, free cash flow of $611 million was good for a free cash flow margin of 8% compared to 38% at NVIDIA and 28% at Intel.

Increasing its size and scale -- especially in the increasingly important data center and cloud industry -- would go a long way toward helping AMD improve its bottom-line. Xilinx could help in a big way. If the acquisition of Xilinx comes to fruition, it would introduce a new chip type, FPGAs (field-programmable gate arrays), to AMD. FPGAs can be programmed post-manufacture, making them a flexible and affordable semiconductor for a wide range of uses like data centers, networking equipment, and consumer electronics. With $3.04 billion in sales over the last 12 months to AMD's $7.65 billion, adding the FPGA leader to the mix would be a game-changer.  

But as with IBM's divestiture, AMD's potential tie-up with Xilinx is about more than generating more growth from the cloud market. At 67%, Xilinx's gross profit margin on its hardware sales and related software platform would be a big boost to AMD. Plus, over the last year, Xilinx generated $1.02 billion in free cash flow -- far more profit than AMD made on more than double the revenue.  

If AMD can secure a deal to add Xilinx to its fold, its ability to chase down other semiconductor companies in the data center arms race would be significantly improved. More chip options mean the ability to cross-sell more, and a jolt of fresh cash each quarter could help AMD on the research and development front, too.

With a divestiture to refocus on its most promising market and a potential acquisition to double down on recent success, IBM and AMD are both worth keeping an eye on. The two respective deals head in opposite directions, but each should significantly improve the companies' strengths in the cloud industry and boost profitability.