As Moore’s law begins to slow down and chip development costs skyrocketing, and customers demanding highly complex products in record low lead times, with chips delivering AI and IoT functionalities in extremely high demand, the Semiconductor industry needs to reassess the current business model. XaaS or AaS, stands for Everything as a service and this concept has been around since the 1960s, Accenture recently published a report exploring the possibilities for the Semiconductor industry incumbents to foray into this highly successful business model. Many industry segments especially software, hardware, networking, and even energy segments have benefited from the AaS model, with SaaS or Software as a service model gaining global recognition and success.

For the Semiconductor Industry however, there is a possibility for all members of the existing value chain to adopt various AaS models and disrupt the marketplace while gaining both a captive customer base and recurring revenues, but, as with all business model transitions, there are many challenges to be addressed and considerations to be made. For starters, large industry players which enjoy a virtual monopoly in their market might be apprehensive to adopt the model, for the obvious reasons related to inertia and because of the perceived risks such a transition might bring to their core business. For smaller players, however, adopting an AaS model might mean freeing up capital and disrupting the market to gain higher share through flexible models, delivering high value to customers and changing their upfront CapEx to OpEx.

Before we understand what possible models may be used by various industry players and how the MES they use might just be the biggest enabler of this transition, let’s take a step back and understand why a company should even consider the AaS model and what considerations must be made before a commitment is made to move forward, further what needs to happen the go-ahead call is made by the top management.

AaS model becomes appealing for various reasons to the Semiconductor industry, and those reasons can be broadly classified as: 

  • Operational efficiency – Service customers benefit greatly from the AaS model as their expenditure shifts from risky and upfront CapEx to a more predictable and periodical OpEx, which allows for freed funds to be diverted towards other imperatives, while reducing costs related to overhead and maintenance.
  • Flexibility and agility – In an uncertain marketplace with COVID-19, trade wars and other unforeseen global events, AaS allows to scale capacity in either direction, based on demand. AaS brings with it a ‘do more with less’ mentality and with it comes tailored solutions which allow for flexibility and agility much needed in uncertain times. Without complex development and approval cycles AI and IoT related deployments can happen faster, spurring innovation and giving companies the cutting edge needed to perform better, financially.
  • Breaking down barriers to innovation – AaS can alleviate innovation challenges for new and smaller players by allowing them access to latest technology without having to invest heavily in equipment and long R&D cycles. This disruptive approach can create multiple and rapid innovations allowing market diversification and challenge status quo.
  • Greater share of wallet – With third party vendors out of the equation, AaS increases revenue on maintenance, spares and services provided directly from supplier to end-customer.
  • Greater access to data enhances cross sell/upsell of services – With AaS it is easier to see how a particular equipment is being used, by who and what sort of issues occur with it at what frequency, this data translates directly to chances of generating enhanced revenue and increasing sales in terms of other products or spares or service contracts.
  • Recurring revenue stream – The AaS provider has a risk-free top line in most cases as revenue streams are regularized with a scheduled monthly delivery of services.

While AaS models are extremely lucrative, assessing the risks associated with sharing IP and the current lack of demand serve as major considerations to be made before deciding to move ahead with the pursuit of a particular AaS model. Questions need to be asked pertaining to the retention of the core IP, creation of a model disruptive enough to generate demand with reluctance from other peers and competitors, adoption of the AaS model which doesn’t disrupt internal operations but still creates value, risk analysis in terms of set and new SLAs, and finally and most importantly, the changes to be made within the business model.

De-risking the AaS

De-risking the AaS offering from an IP perspective is an extremely important consideration, a change in any business model brings with it many implications and this is where the modern industry specific MES platform comes into the picture.

Irrespective of the eventual AaS model chosen, whether it is a pilot/service addition, or a business inside the business, or even the transformation of the entire business, Accenture stipulates that the endeavor would require the design and implementation of new operational models which will need hyper-responsive teams, which are integrated and interconnected with shorter planning cycles and connected data, systems and workflows. Further the report highlights that force-fitting legacy processes for new AaS models will not translate to customer success.

The MES thereby becomes a key contributor here, let’s understand why.

The MES application at its core is responsible not just for the automation of process execution but for integration of the enterprise with the shop floor. AaS models per Accenture require extremely tight collaboration and immense flexibility to deliver faster service and increase revenues.

Along with integration across the board the MES brings IoT enabled edge computing and AI fueled advanced analytics to the semiconductor manufacturing value chain, which in turn allows for tighter collaboration based on deep insight and faster decision making, this directly impacts the efficiency of the core process while allowing for a base to be built for advanced process modifications and business process reengineering.

With the right MES application data collected is converted into actionable intel and can be used over time to form predictive patterns, this can be of massive value for everyone in the semiconductor value chain, right from design houses to OEMs and contract manufacturers. Data which predicts consumption patterns for example correlated with market movements can help companies offering manufacturing as a service or MaaS, to better configure their lines and either restrict or diversify their portfolio and customer base to become more aligned with developing trends and consumption patterns. 

Similarly, the MES allows for equipment manufacturers or EaaS suppliers to access not just raw data from the process but details of how a particular equipment is being used, when it breaks down, and what spares are being used on it, with this insight the service provider can not only boost own revenue, but offer better more proactive support to the end customer. XaaS models are built on hyper-connectivity and hyper-integration, but also need de-risking when it comes to IP sharing, the MES can be the foundation which allows this to happen and helps value chain members check all the boxes of pre-requisites defined in the Accenture report.

Accenture also warns against the use of legacy processes in building a new business model, this is perhaps truest for IT applications and the MES in particular. Legacy tools and processes are representative of inflexibility and lack of scalability, the modern MES platform addresses both these concerns emphatically, thereby making it the obvious solution for an AaS transition.

Roadmap for establishing the AaS model

When building an AaS offering, the C-suite of the organization must recognize this shift as a fundamental change in the way things will be done, which means their involvement has to be more than approving a plan. This has to be a top driven endeavor. Building a cross-functional and even a cross-organizational team should be the first logical step. 

Risk assessment comes next, pursuing any version of the AaS business model should be driven by a solid business case and should not be pursued only because it is being perceived as the next frontier. Understanding the financial risk and overall risk assessment should be performed before, ROI concerns should be addressed based on the AaS model chosen by the project team. A key factor in assessing risks should be the evaluation of current processes and support infrastructure in being able to manage the new model, this analysis might reveal gaps such as the need for a new MES or new personnel/vendor, which in turn may not just help establish the new business model but effectively contribute towards the improvement of the core business too.

The next step after a detailed risk assessment would be to create a 12-24 months tactical plan which would focus on avoiding financial risks while targeting revenue generation in the fastest possible time. Once the plan is made the team gets cracking with the execution, alignment and organization of teams, suppliers and operational reconfiguration to deliver value as planned and realigns based on progress. 

Depending on where a company is located in the Semiconductor value chain, everything from manufacturing to yield and silicon to design can be offered as a service, even Quantum as a Service or QaaS is an ambitious business model being considered. Irrespective of the choice you make, an extremely important consideration is the ability of not just your personnel and OT to support the new business model, but your IT and specifically your MES application to deliver the flexibility, integration and collaboration which form the basis of the AaS business!