The shares of Microsoft (MSFT) and Alphabet (PORTL) are incredibly similar stocks that collapsed this year, with similar growth profiles, identical market capitalisations and valuations.
Which one do you choose?
Both companies are excellent companies, and a credible reason can be found for both. By eliminating the differences in their revenue streams – such as operating system software versus advertising – it is important to dominate the software ecosystem. In this regard, both Microsoft and Alphabet have powerful cloud computing operations, each leading the way in the most important category of computer software, machine learning.
But Alphabet's Google unit undoubtedly gives Microsoft the upper hand.
With market capitalization of $ 751
The difference between the two depends on what's going on in these cloud operations. While scalable, Microsoft may not have the more prominent relative position with developers.
Machine learning, where computers can develop some of the code they automatically use by extracting patterns from data, is now the most important programming trend. 19659002] Together with Amazon (AMZN), both companies have used machine learning tools in their cloud computing operations. Microsoft's Azure cloud service is generally considered to be the second largest revenue from Amazon Web Services, with Google Cloud being a third. However, since neither Microsoft nor Google disclose the revenue for the units, the comparison is unclear.
However, it's safe to say that their cloud operations are important resources that both companies use to attract and retain developers of neural machine learning networks.
However, Google has taken control of the most important "framework" for creating machine learning applications, TensorFlow. A framework is a collection of pre-made code that can be used by a software developer to assemble programs without reinventing the wheel. At the moment, TensorFlow is the predominant machine learning framework.
It's not the only framework, and Microsoft has some important offerings for itself. This includes the so-called "infer.NET" framework and the other "Microsoft Cognitive Toolkit". Both are used by developers in a number of interesting projects.
But over and over again, the vast majority of machine learning work comes through TensorFlow; either that or a few other frameworks from companies other than Microsoft, such as "PyTorch" from Facebook; a frame called "Caffe," originally developed at the University of California at Berkley; and recent initiatives based on these programs, such as "fast.ai", by launching this name in San Francisco.
Nothing will change in the $ 7.5 billion acquisition of Microsoft's GitHub developer community in October. Machine learning frameworks gain the strengths or weaknesses that they possess, as well as the support they generally find. GitHub has the most popular programming repository and does not change the decisions of frameworks that individual developers encounter.
Why framework control matters is that Google builds on TensorFlow. They have their own custom chips in their cloud operations called Tensor Processing Units (TPUs). These TPUs ("TPUs") have become a leading piece of silicon and are provided by Google in the cloud computing service. In combination with the TensorFlow software, TPUs Google offer a combination of chips and software that offers an advantage over current offerings from Microsoft.
And TensorFlow has quickly expanded into mobile computing operations, where Google has the dominant presence worldwide through the control of the Android operating system and Microsoft has no presence. A study by researchers published in November, including Peking University, as well as Microsoft's own research group in Beijing, found that among the best mobile applications for Android, those using machine learning in some form, either half used TensorFlow becomes; TFLite, a smaller subset of TensorFlow created by Google for mobile devices only; or "ncnn," another framework created by the Chinese social network Tencent (TCEHY). Microsoft's code is not yet as advanced as machine learning on mobile devices.
This whole area of machine learning is still very young and very new. There are many more innovations, and Google's relative factor will shrink and evaporate over time. An interesting initiative from Microsoft is advocating something called the "Open Neural Network Exchange," or ONNX, which makes developers less dependent on individual frameworks over time.
Time will tell: Google now has Google of the Microsoft in the main tools that are used for the most advanced types of application development. That's worth a bid.
Alphabet, Microsoft and Amazon hold shares in Jim Cramer's Action Alerts PLUS member club . Would you like to be notified before Jim Cramer buys or sells Bulk, MSFT or AMZN? Learn more now.