[Core Information] Hedge Funds VS. Texas Instruments: The core controversy between self-operated production and outsourcing

Recently, the semiconductor industry has staged a game over its core business model: activist hedge fund Elliott invested US$2.5 billion in analog chip maker Texas Instruments (TI). The core demand is to urge TI to increase its free cash flow level. This move directly triggered a fierce market controversy between TI's self-operated production and the industry's mainstream outsourcing model.Affected by this news, TI's stock price rose 3.4% during the day, hitting a record high of $206, which also made this chip controversy the focus of the industry.

1. Hedge funds attack: declining cash flow + overcapacity, pointing out the shortcomings of TI's self-operated expansion

Elliott's core doubts focus on TI's free cash flow performance, believing that it has deviated from its long-term commitment to drive free cash flow per share growth.Judging from TI's financial report data, this doubt is not groundless: in the first quarter of 2024, TI's free cash flow was only US$940 million, a year-on-year plunge of 80% from US$4.4 billion in the same period last year.The core reason for the sharp decline in cash flow is TI's aggressive 12-inch wafer fab expansion plan-continuously rising capital expenditures, which have severely squeezed the accumulation space of cash flow.

Even more pointedly, Elliott pointed out in a letter to TI that TI’s expansion plan announced in 2022 poses a serious risk of overcapacity.The plan aims to nearly triple production capacity by 2030. The investment scale is too high and may cause TI to lose billions of dollars.According to Elliott's calculations, TI's production capacity will exceed market demand by 54% in 2026. Even if the production expansion plan is completed by 2030, the overcapacity ratio will still remain at 50%.

As a radical hedge fund whose core approach is to target corporate weaknesses and promote strategic adjustments, Elliott has already had successful cases in the IC industry - it once promoted the split of flash memory manufacturer Western Digital into two companies. This investment in TI is essentially targeting the shortcomings of its self-operated production expansion model, trying to push TI to adjust its business strategy, and then seek high investment returns.The decline in TI's free cash flow and the hidden dangers of overcapacity are just the starting point for Elliott to launch an attack. Its core goal is to change TI's asset-heavy self-operated development path.

TI's financial report is under pressure: revenue and profits have both dropped, and profitability continues to decline.

In addition to the cash flow indicators that Elliott focuses on, the overall decline in TI's recent financial reports has aroused widespread alarm in the investment community and industry.In the first quarter of 2024, TI's revenue fell by 16% year-on-year to US$3.661 billion, and its net profit fell sharply by 35% year-on-year to US$1.105 billion. What is particularly noteworthy is that its gross profit margin dropped to 57.22% year-on-year. Under TI's long-term strategy of promoting product price increases and manufacturing cost reductions, this core profit indicator has exceeded 70%.The significant decline in gross profit margin intuitively reflects the weakening of TI's profitability and provides strong support for Elliott's attack.

Compared with the overall performance of the industry, TI's growth momentum is also weak.Since the market low in October 2023, the Philadelphia Semiconductor Index (covering the world's major semiconductor companies) has increased by approximately 55%, while TI has increased by only 30% during the same period.The current growth hot spots in the semiconductor industry are concentrated on AI-related companies such as NVIDIA, whose main business is GPUs, which are in strong demand; while TI's core business focuses on analog devices, which are facing the dilemma of sluggish industry demand.More importantly, the difference in business models further amplifies this gap: NVIDIA adopts a fabless model, which is an asset-light operation and does not need to bear huge production capacity investment costs; while TI continues to expand its self-operated production scale and is burdened with high capital expenditures. The combination of short-term supply and demand pressures and long-term cost burdens has caused some investors to have doubts about its strategic initiatives.

2. The core of the industry game: the core choice between self-operated production and outsourcing models

The essence of this dispute is the collision of two core manufacturing models in the semiconductor industry - TI's self-operated production model, and the industry's mainstream wafer foundry (outsourcing) model. Which one is more suitable for the current market environment and the long-term development of the company?

The rise of the wafer foundry model stems from the cost and risk pressures brought about by the continuous upgrading of transistor processes: as process accuracy improves, the investment costs and technical thresholds for chip companies to build their own wafer factories have increased significantly. The emergence of professional foundry companies can help companies balance the contradiction between process upgrades and cost control.Even for basic components such as MCUs, when the process reaches a certain level, they will choose to outsource production. For example, original manufacturers such as Renesas rely on professional wafer foundries such as TSMC for products with 40 nanometer and more advanced processes.

Especially with the rise of emerging applications such as automobiles and AI, the performance requirements of basic components have continued to increase. Coupled with changes in the global supply chain pattern, wafer foundry has become an important support for the local layout of IC companies.Taking TSMC as an example, it has deployed wafer factories in the United States, Europe, and Japan to accurately match the needs of different regions: the European factory mainly provides production capacity for local automotive IC companies such as NXP, Infineon, and Bosch; the Japanese factory serves Renesas, Sony, and local automotive-related customers; even American IC original manufacturers such as ADI and Microchip have booked the production capacity of TSMC's Japanese factory in advance, which can not only expand production, but also greatly improve the response speed to demand in the Asia-Pacific region.

Industry Case Reference: Practice and Enlightenment of Fab-liter Model

In addition to the fabless model that relies entirely on foundry, some OEMs choose the fab-liter model, which means streamlining their own production capacity, focusing on core products, and increasing outsourcing of non-core products. ON Semiconductor is a typical representative of this model.As a strong OEM in the field of automotive chips, ON Semiconductor will continue to retain and increase investment in internal production capacity (including silicon carbide) for products with differentiated advantages such as smart power supplies and sensors. For products without patent advantages, it will reduce operating costs and improve profitability through outsourcing.

In 2022, ON Semiconductor has withdrawn four of its own wafer fabs and concentrated its remaining production capacity on the more efficient 12-inch production line - this is similar to TI's 12-inch production expansion strategy, but the core difference is: TI's 12-inch production expansion uses independent new production linesFocusing on production capacity, combined with the direct sales strategy, it promotes products to the market through continuous cost reduction. This is also an important reason for it to dare to start a price war with domestic materials last year; while ON Semiconductor achieves cost control and profitability improvement by streamlining its own production capacity and optimizing the outsourcing ratio.

TI’s persistence: strategic considerations behind self-operated production

Judging from the product characteristics of analog chips, they do not have high requirements for process accuracy, but have extremely high requirements for product quality uniformity and long-term, high-volume, low-cost supply capabilities - this is the core reason why TI insists on self-operated production.In order to consolidate its advantages in the field of analog chips, TI has set a clear goal: to achieve more than 90% of manufacturing and packaging in-house by 2030.In order to realize this plan, TI currently plans to add seven 12-inch wafer fabs, located in Sherman, Dallas, Richardson, Utah, and Lehi, Utah, etc., in an attempt to continue to support the needs of global customers through a complete self-owned supply chain.

However, TI is not completely ignoring the current market difficulties.Elliott also knows that TI's investment in production expansion has basically been put in place, so he does not require it to give up production expansion completely, but urges it to increase spending flexibility.In fact, the investment in the wafer fab itself is feasible. TI can avoid the rapid increase in depreciation costs by delaying the introduction of machines after the completion of the factory. And as the peak period of production expansion passes, TI's free cash flow is expected to gradually rebound, thus meeting the core demands of investors.