Atos SE FAQ: 3 reasons why Atos is splitting up (and 1 why not)

Why would you want to split a company with +/- 11 bn Euro revenue and +/- 112.000 employees? It is not a small undertaking and for sure not a decision a leadership team would take lightly. What are, in my opinion, the key reasons behind this decision?

As a regular analyst and commentator on Atos’ strategy, and the IT Services industry in general, I speak often with 3rd party advisors, investment firms and other interested parties. These conversations address a wide spectrum of topics, ranging from financials, competitors, and unique selling points, all the way to how to build the right sales or delivery teams.

There are however some questions that frequently pop-up in these conversations. So, I thought I might spend some time answering them in a new series of blog posts.

Today we will look at 5 reasons why Atos is going to split the company in 2 new public companies.

On June 14, 2022 Atos announced a plan to study “a possible separation into two publicly listed companies to unlock value and implement an ambitious transformation plan“. And in the communications on the third quarter results, confirmed that “The separation project is well on track to be completed in H2 2023 as initially planned“.

1. Declining revenue and margins in IT infrastructure services

Mostly due to the rise of public cloud providers such as Amazon, Microsoft and Google, a lot of software is moving to the cloud. This results in a steep decline in the cost of running these applications. In the public cloud the infrastructure can be shared. Management can be highly automated. And locations can be transparent allowing for even more flexibility in costs. Cloud computing also takes away the need for local, privately owned, data centers. And companies are embracing cloud computing en masse. In 2018 Gartner predicted that “Around 10% of enterprise-generated data is created and processed outside a traditional centralized data center or cloud. By 2025, Gartner predicts this figure will reach 75%”.

Atos’ problem is that a significant portion of their IT services revenue and assets are still not cloud-based. Atos owns and operates still many datacenters and needs customers to fill those with hardware and software.

This “move to the cloud” is not only a problem for Atos. In recent history also IBM decided to split off their IT Infrastructure Services in a newly formed company Kyndryl. And in April 2017, HP performed a similar action when it split-merged their infrastructure services with CSC.

2. Create a clearer strategy

The one company that Atos is today is a bit of a collection of IT topics that not necessarily work together. In some cases, they might even be competitive. This makes it harder to create a single strategy across all offerings of the company. By splitting the company in two, the two components can focus on their respective strengths without (negatively) impact the other. E.g., the part that offers infrastructure services can focus on customers that have a need for local and/or physical datacenters.

3. Workforce challenges

There is a very big difference between the attributes of a workforce oriented on traditional Infrastructure Outsourcing Business, versus the more innovative oriented business of Digital and Data Analytic services. While the workforce for the services provided in infrastructure come most of the time from the customers of Atos that outsource their business to Atos, the growth business in Digital is dependent on recruiting new talents.

When the traditional outsourcing services are dwindling, and the Digital Services are growing, Atos’ needs to hire and/or educate more and more people in a market that is already short on potential candidates. At the same time a growing part of the workforce, that is most of the time local (meaning ‘not in off-shore countries’), is generally not immediately fit to move into this new technical territory.

Recruiting new staff, both off-shore and local, will be easier if the profile of the company is not about traditional services, but instead radiates innovation and ‘cool’ technologies.

… and why should Atos not split the company?

Ever since I have been made aware of the strategy of Tesla, I believe firmly that ‘verticalization’ is a very strong asset for any company. If Atos can bring in verticalization in its offerings, it could build some very strong offerings.

Atos can combine their scale, their own hardware, software, services, and consulting skills into a mix that allows customers to work with them, instead of Atos being a contractor if separate IT services. Today Atos has all these assets under one roof and that sets the company apart even from their biggest competitors. I consider this to be a huge opportunity. And it can be implemented gradually, with a focus on the most profitable business topics first. And I think that such a program would cost less that the investment that Atos is currently raising to pay for the split.

It seems to me that previous Atos leadership teams, most notably their former CEO Thierry Breton, understood this. And it is a shame that recent leadership by Elie Girard and later Rodolphe Belmer, were unable to build on his legacy and vision.

Disclaimer: Paul, who is the author of this blog post, holds at time of writing a small amount of stocks in Atos SE. All information in this blog post is believed to be public information, enriched with the authors personal opinion. No confidential information is being shared.

Tomorrow it starts.

On February 24 I will embark on a monthlong sailing trip from the Netherlands to the south of France. I will be a passenger (and supporting crew member) on Sailing Vessel “De Morgenster”. Very much looking forward to meeting the crew and fellow passengers tomorrow in Den Helder.

4 ways the Industrial Internet of Things can enable digital transformation

Disclosure: This post was previously published on Atos Ascent and was co-authored by Mr. Andrea Sorrentino (who wrote a significant part of the text below) – minor format and content edits have been applied to fit it to this website.

Today the manufacturing industry faces one of the biggest challenges in modern times: how to embrace the next industrial revolution. The technological disruption which has arisen from “Industry 4.0”, the current trend of automation and data exchange in manufacturing technologies, has drastically changed how we see the world today. Moreover, the highly competitive landscape poses urgent questions of change management that manufacturing companies need to address quickly.

Nowadays, in a more inter-connected world, companies need to adopt the right tools to get closer to the market and become more competitive. The manufacturing industry needs to adopt big data to innovate, to optimize their processes, and improve yields. But how should managers approach this technological revolution and work towards creating a smarter factory?

Machinery performance can now be measured with small sensors connected to the internet, monitoring where efficiency can be optimized. For example, workers will be able to foresee machinery malfunctioning, and intervene in a timely manner. Gartner’s latest forecast predicts 20.4 billion things will be connected by 2020, which will completely change the way we work in several sectors. In the meantime, machine-to-machine communication is fast becoming a reality, and the Internet of Things (IoT) represents only the first step in that process. In this post, I look at four ways in which managers in the manufacturing industry can exploit IoT solutions for commercial benefits:

1. Understand how your company can profit from IoT solutions

Industrial standards for IoT are still not clear, in part due to the wide range of flexible solutions that can be developed in any industry. Factory and operations managers need to define a series of objectives to understand how best they can benefit from connected devices. This is because while it is relatively easy to collect data, it is difficult to understand how to cluster it to avoid complexity in analysis. It is key to define first what type of data can be useful to increase efficiency, rather than trying to analyse huge amounts of data which is not necessarily relevant. By honing in on the aspects which will make the most difference to the business, in terms of profit, subsequent analysis becomes much simpler.

2. Clarify your security strategy

Security is a fundamental aspect to take into consideration when adopting any IIoT (The Industrial Internet of Things) solution. Consumers still hesitate when purchasing IoT products, in part due to safety and privacy concerns. Similarly, manufacturers are hesitant to adopt IoT solutions since data and app platforms can be subject to cyber-attacks as well. Operational processes and risks need to be coordinated around these safety issues, a topic Atos explored in their Journey 2020 report.

The IIoT is fundamentally changing the cyber-security landscape; the old logic of go-to-market quickly to gain market share over competitors does not apply anymore. Any device connected to the internet can become a weapon for hackers, and IT companies need to effectively secure their infrastructure before delivering to clients. Having real-time security analytics and a cyber-resilient system are essential when deploying IIoT solutions to protect against any potential attack.

3. Consider sustainability

Sooner rather than later, companies need to consider how they will ensure that their IoT solutions remains sustainable i.e. future proof. Their IoT strategy needs to define, among other things, what type and the expected amount of data they need and how to manage it effectively to minimize potential negative impacts. It is also crucial to understand how to manage potentially millions of connected devices, and how to build a scalable and reliable, distributed computing environment around the production factory.

Sustainability in the sense of Corporate Social Responsibility (CSR) also offers opportunities. It is currently one of the hottest topics in the manufacturing industry. The adoption of sustainable strategies is something many manufacturers are beginning to take into consideration, since brand reputation depends on companies working to reduce the level of emissions and waste they generate.

IIoT will enable the creation of what researchers have termed a ‘circular economy’; the concept that puts re-usability and recyclability at the center of any type of process. Through IIoT solutions, managers will be able to extend the lifespan of machinery, thus cutting energy costs. Therefore, the development of smart-factories would likely result in a more ecological manufacturing industry, thus drastically reducing the impact of industrial processes in the environment.

4. Get out of your comfort zone

The industry is still in an early stage when it comes to IIoT adoption, but some pioneers are taking steps to ensure they are future leaders of the industry 4.0 era. The opportunity is out there, and decision-makers need to act rapidly to advance in this next wave of technology change.

It is fundamental to assess your own capabilities and role within the IoT ecosystem: will you push data or will you pull data? In a push model, you need to look at the smartness of your devices and data platform. If you pull data you need to look at your data analytics capabilities so you know when to ask for data and what data you need. Companies cannot bear the risk connected to data management for the entire production chain, therefore, it is necessary to build a partner ecosystem of buyers and vendors that co-operate for creating secured, efficient and scalable end-to-end solutions, leading to real added value in the production chain.

Internet of Things : My disappointment(s) with the Fibaro HomeCenter 2

Recently I started investigating and investing in Home Automation. I feel the market is readying itself for mainstream and some interesting products are now available. It started with me buying a set of Philips Hue light-bulbs, a motion sensor and a Raspberry Pi. Connecting these devices with the help of existing apps and some tinkering with open source ‘Home Assistant’ and ‘Home Bridge’ allowed me to create simple automation scenes and use Siri to set the lights in the living room and hallway.

This first success led me to think I could take the possibilities inside my house to the next level; I researched extensively possibilities of various Home Automation platforms. In Europe there are various systems available. In the end I decided for the Fibaro Home Center 2 system. Based on Z-Wave technology, which allows 2-way communications between central hub and device, with a mesh network to combat low connectivity situations it seemed the right choice. Fibaro also has a ‘Light’ version of this hub, but I wanted to have some serious strength running my IoT platform so opted for the strongest system. It also promised interconnection with my Hue, SONOS and (upcoming) Alexa systems, so I was really excited to get started.


Fibaro's compatibility claims
Source: https://www.fibaro.com/en/compatibility/


It turned out to be a terrible disappointment.

Before I describe my activities that led me to the final conclusion, there are a few things to note:

  1. There are 4 ways to access devices and sensors in the Fibaro system:
  • A simple panel that shows an on/off or volume/brightness slider.
  • Through a magic scene creator that works with simple colored boxes that allow you to do an IF…THEN scenario
  • A more complex colored boxes interface that allows more variables
  • The LUA programming language

Fibaro Colored Block 'programming'
Source: http://www.5smart.ru/page/kak-rabotaet-umnyy-dom


  1. I researched everything, I have seen more YouTube videos than I care to remember and have seen more programming examples that I wished for. I took a quick course on LUA programming.
  2. I am not stupid with DIY projects in house; electric cabling, switches, dimmers and wiring. Most of this stuff I have done in the past on my own. I understand the color coding of wires, the difference between Ground and Neutral and have had my fair share of accidentally touching a Live wire (at least once in a person’s live…).

Let me first give you an overview of my purchases from the local Domotica Shop in the Netherlands:

  • 1 Fibaro Controller Home Center 2
  • 3 x Fibaro Module Dimmer 2
  • 1 x Fibaro Module Switch Single
  • 1 x Fibaro Module Switch Double

btw: these things are not cheap so I really researched (at least so I thought)

Secondly I will go into the various activities I undertook to get this box running the things I wanted.

Connectivity with Philips Hue lights and sensors

As far as I have researched it, there is no out of the box support in Fibaro for Philips Hue. There is the possibility to download a virtual device component that will connect to the individual Hue lights. Unfortunately this means I was unable to address the Hue lights in the visual programming interface of the Fibaro (a colored block system) and can only address the lights through the programming language LUA. This was a major set-back as I was let to believe integration between Hue and Fibaro would work out-of-the-box.

The Hue sensors such as the motion sensors and dimmers are not accessible through the Fibaro; maybe they could be accessible through the LUA programming interface, but I found no evidence for that.

Instead of being able to address the Hue lights in the ‘Magic Scenes’ or the ‘colored boxes’ interface, I had to install virtual devices using the Philips CLIP API Debugger. This would give me the API Access code that could then be used to activate the virtual devices. After that the virtual devices would become accessible in other code, that I had to write for the scenes. Are you still there?

Due to the lack of access to the Hue sensors, I would have to rely on the Fibaro sensors, which I hoped not to (to be honest I knew upfront that the ‘rely-on-Hue-sensors-part was going to be tricky as they do not support the Z Wave protocol in use by Fibaro).

Conclusion: Could work, but would need a serious amount of manual programming

Connectivity with SONOS

I have 4 SONOS loudspeakers. 2 are connected in a stereo pair. One other is in the kitchen and one is upstairs in my study.

Fibaro shown as connected to SONOS
Source: https://www.fibaro.com/en/why-fibaro/

There was no out-of-the-box connectivity with the SONOS system. It required the download of a virtual device (remember those?). As a result I managed to connect to 1 SONOS loudspeaker. It became clear that I had to download a virtual device for every loudspeaker separately. After doing this I was able to start/stop the music and volume of 1 speaker. The others were not responsive. And the stereo-pair was nowhere to be found in the system. Moreover, if the one that worked had no active playlist, it would not do anything.

By now you will have guessed that no devices were easily accessible; all the work had to be done through the LUA programming language.

Conclusion: Could work, but would need a serious amount of manual programming.

Installation of Fibaro modules

So, as it turns out Fibaro on/off switches need power; makes sense right? As a result installing switches behind a light switch can only be done of the box behind the switch carries not only a live wire (brown), but also a neutral wire (blue). It turns out that that is not the case in my house. Maybe more modern houses have this or maybe in other countries, but my house is 28 years old and only has a brown (Live) and black (Switch) cable.

Luckily, Fibaro has a module called Dimmer 2, which can be installed if the neutral (blue) wire is absent. So all is well, no unfortunately not. The switch-boxes in the wall of my house are not deep enough for the combo of a light-switch and a Fibaro module. So I hunted the local DIY market for light switches with a very small footprint and with some clever wire maneuvering was able to fit everything, very snugly, into the switch-box. Following that I connected the module to the Fibaro controller and, voila, I could remotely operate the lights. Alas, the physical switch was not working anymore, which would really not make my wife very happy. I checked the wiring, looked at another 4 YouTube videos (warning, look at installation for “Dimmer 2”, otherwise you will get really confused).

Conclusion: Could work if your house is built to fit Fibaro or you do not mind some reconstruction of wall-mounted switch-boxes (including drilling new holes and putting in new wiring).

Conclusion and words of warning

I have no doubt that the Fibaro system will work, once you get the switches, dimmers and specific other Fibaro modules installed; but this fairyland only exists on the various Fibaro websites. No real life scenario worked for me. Existing wall-switches, existing Hue system and existing SONOS speakers could only be installed with great cost, either in time or with additional investments and serious physical changes in my house. I do not believe that house automation should go hand in hand with yielding a jackhammer to replace switch-boxes.

in conclusion, device-by-device:

  • SONOS – no out of the box support of a SONOS system – no visibility in the graphical scene creator
  • Hue Lights – no out of the box support of a SONOS system – no visibility in the graphical scene creator
  • Hue Sensors – no support
  • Fibaro Modules – not working in my electrical system

And a word of warning: the Fibaro website mentions interoperability with Apple HomeKit. This is only true for a very limited set of Fibaro modules/devices; a smart power-plug, a motion detector, a flood sensor and a door/window sensor.

So at the end of a week with not so much sleep and a lot of learning of new things, I concluded that this (rather expensive) set up was not for me and as a consequence I have requested the seller to take it all back; I hope they agree that this was a bust.

PS. I still have my Raspberry Pi running with Home Assistant and Home Bridge. I am very proud to show the Siri integration with the Philips HUE lights, the SONOS Speakers, the Honeywell EVO home heating system and the NEST smoke & CO2 detectors. At least some systems do work in my house, and at a fraction of the cost.

3 times news: Website – Hosting – WordPress

cropped-j0390061.jpgLast couple of days I moved my website from TypePad to a self hosted WordPress site. I have chosen “GeneratePress” as my theme and have purchased their Premier Plugin pack to do some cool stuff in the design.

I am still experimenting so you can expect some changes, none of which I can guarantee to be permanent. It is great to have the ability to experiment and see how things work out.

I must say it feels good to be much more in control. TypePad became a bit frustrating.

Anyway, let me know if you have comments.