In the past, semiconductor manufacturers could derive multiple benefits from using a fabless or fab-lite business model. By outsourcing some or all of their production, firms could focus their resources on designing new chips. Companies could also use that operational foundation to keep their overhead and labor costs at a minimum.
However, in the last half-decade geopolitical strife, unforeseeable crises, and paradigmatic end-market shifts have highlighted that methodology’s flaws.
Following so much radical change, the chip sector needs to reevaluate its practices and reimagine the fab-lite model.
Why Fabless/Fab-Lite Models Used to Make Sense
In recent decades, fabless/fab-lite business models have become increasingly popular because they make a lot of sense from a financial perspective.
For firms looking to break into the industry, commencing operations as an integrated design manufacturer (IDM) is not financially viable. It takes staggering amounts of capital to build new, large-scale factories, staff them, and equip them with fabrication tools. A new company would also need to establish a supplier network to keep their fab fully stocked with raw materials.
In contrast, startups have gained traction with modest facilities and scaled up by partnering with pure-play foundries.
Along similar lines, some large chipmakers have found remarkable success by becoming fabless component vendors.
Advanced Micro Devices (AMD) is a prime example of a corporation finding success after making the transition. It started as an IDM but spun off its production capacity as an independent firm called GlobalFoundries in 2009. AMD’s focus on design enabled it to become a more innovative and financially successful business over time.
The company’s agile nature has given it a competitive edge against less agile institutional IDMs.
Other major players within the semiconductor field have greatly bolstered their bottom lines by moving to a fab-lite structure. Renesas Electronics’ decision to outsource its automotive microcontroller production to TSMC in 2018 allowed it to slash its manufacturing costs. And ON Semiconductor greatly improved its balance sheet by divesting some of its manufacturing facilities.
However, recent developments within the sector have thrown the drawbacks of the fabless/fab-lite models into sharp relief.
Recent Developments Have Upended Industry Practices
Since 2018, the global semiconductor industry has encountered unprecedented disruption because of a sequence of world-changing events.
The U.S.-China trade war led to new tariffs being added to components moving between the two countries. Consequently, products designed by fabless chipmakers have become harder to find and more expensive at the end-market level. Many manufacturers have moved to diversify their production capacity to avoid future geopolitical headwinds.
But the massive financial outlay involved in enacting that strategy means it is not available to all firms.
In 2019, the coronavirus pandemic began and prompted widespread factory shutdowns and travel restrictions. The scale of the health crisis meant it affected all semiconductor vendors. But some IDMs mitigated the impact COVID-19 had on their operations because they could coordinate responses across the supply chain.
Vendors that rely partially or entirely on pure-play foundries to fabricate their offerings could not react as effectively.
Most recently, a massive spike in demand for automotive semiconductors late last year grew into a multisector global chip shortage. At present, industry leaders and analysts believe the supply crunch could extend into 2022 due to robust end-market interest, a scarcity of raw material, and constrained foundry space.
Market watchers pointed out the component shortfall has been exacerbated by the industry being overly reliant on a few contract manufacturers.
In addition, the sector experienced a massive consolidation wave in 2020, in part because of the compound impact of the trade war and COVID-19. The largest of those combinations occurred between fabless chipmakers looking to expand their market share. However, some of these tie-ups have raised antitrust concerns that could evolve into protracted legal actions.
The last five years have highlighted the problems inherent to outsourcing component production. However, the field has an opportunity to address that vulnerability by reevaluating long-standing business methodologies.
Collaboration is the Best Way Forward
Since the U.S.-China trade war began, several countries have launched initiatives to develop their local semiconductor ecosystems, including the two superpowers. Several national governments want chipmakers to establish new fabs in their regions to create new jobs and bolstered their local economies.
That development presents an opportunity for fabless and fab-lite manufacturers, one that can be best realized through collaboration.
Even with federal support, launching new state-of-the-art large component factories is a multibillion-dollar endeavor. But by pooling their resources, manufacturers can better manage the financial burden of opening and maintaining a production site. That way, industry players can redefine the fab-lite model to characterize providers that share production facilities.
Rival component makers working together for their mutual benefit is not unheard of, even in today’s competitive landscape. However, those alliances are commonly forged to serve public clients or mainstream unproven technologies. Going forward, firms can capitalize on the national semiconductor ecosystem trend to fund the construction of new wafer fabrication plants.
By working together to reinforce and localize their production capacity, chipmakers can protect themselves from future disruption. They will also remove a weakness identified by analyst group Future Horizons over a decade ago: fabless/fab-lite manufacturers’ inability to take advantage of sudden end-market changes.
It will also make non-IDMs more competitive without going through lengthy and legally contentious mergers and acquisitions.
The need for systemic changes in the global semiconductor industry is undeniable. Bilateral interest in making the sector more collaborative is clear. The onus is now on individual providers to take the next step in building a brighter future.