Description

Most people associate TI with the handheld calculators they used at school.  However, over its 90-year history, TI has been active in many areas.  The company can trace its origins back to a small oil services business established in the 1930s which helped oil companies test the viability of new oil fields using sound waves.  From this work the company accumulated expertise in radar technology which it used to develop equipment for the U.S. military in World War II.  As part of its R&D it established an electronics equipment lab out of which it commercialised the first silicon transistor in the 1950s.  Later that decade employee Jack Kilby invented the first integrated circuit.  For this work he would later be jointly awarded the Nobel Prize in Physics.

Today, TI is the global leader (c.1.5x larger than its nearest competitor) in the design and manufacture of ‘analog’ chips, a business it started pivoting towards in the 1990s.  TI has been involved in analog electronics since the 1940s (radar is an analog technology).  But it was not until the company began concentrating on electronics for communications in the 1990s that the role of analog electronics became central to its business strategy.  Over the course of that decade it sold its original oil services unit to Halliburton, its missiles and defence electronics business to Raytheon, its computer operations to Hewlett-Packard and Acer and the memory-chip business, which had at one time been its biggest unit, to Micron Technology.

Digital semiconductors work with discrete on and off binary one and zero inputs.  In contrast, Analog semiconductors work with real-world signals, such as sound, temperature or pressure which are continuous and variable.  Analog chips condition, amplify and then often convert these analog signals to a stream of digital ones and zeros which can then be processed by digital semiconductors.  

This fundamental difference goes a long way towards explaining why digital and analog companies have different underlying economics.  Digital functions can be expressed in a branch of mathematics called Boolean algebra which can be applied to program software.  Most digital circuits are highly regular and repeated, with predictable performance.  Digital semi-conductor companies typically produce a small number of general-purpose chips that can be programmed to do multiple tasks.  The name of the game in this business is to drive down costs through volume and scale.

In analog circuits, there is no corollary branch of mathematics. The IP in analog is more tacit in nature.  Analog chips tend to be highly customised for specific customers and applications.  Companies like TI therefore tend to have very broad product catalogues containing tens of thousands of different chips.  Customer switching costs in analog are also much higher.  The regularity and predictability of digital chips makes them highly interchangeable with little or no re-qualification by customers required.  However, if an end customer wants to source an analog chip from a new supplier they must recheck multiple performance parameters to confirm the chip behaves the same way.  Given the average analog chip costs 40 cents the risk of switching out of chips that are working well far outweighs potential cost savings.  As such, once an analog chip is designed into an end product it is typically reordered for the lifecycle of that product, which for TI is on average ten years.  All these factors make it much harder for new entrants, no matter how well funded they are, to gain a foothold in the market.

In order to be successful as a catalogue business, you need a very large range.  TI has the largest catalogue in the industry with over 80,000 SKUs.  For a new entrant to compete it would need to create tens of thousands of products, most of which would have very little revenue associated with them even after they had been established for decades

The demand for analog chips does tend to be more volatile than end market demand due to long order lead times and inventory builds.  In periods when supply is tight, customers tend to place more orders than they need to ensure that they receive enough supply.  The decision to build inventory needs to be made well in advance given production lead times can be as long as 8-10 weeks.  When end customers see a downturn in demand the effects are amplified for analog companies as customers work through their excess inventories and hold off placing new orders.  The result is that analog companies are occasionally affected by short-term inventory cycles which can last several quarters.  However, over a long time period analog semiconductors’ average growth rates tend to be in line with end market growth.

We believe these inventory cycles provide occasional attractive buying opportunities. TI is currently going through what we believe is one of these short-term inventory down cycles.  This has resulted in recent earnings and guidance disappointing the market which in turn has put pressure on the share price.  We view this weakness as a potential buying opportunity.  However, the company still doesn’t look particularly cheap trading on a forward p/e of 21x.  For there to be a sufficient margin of safety in the valuation, the company has to be able to grow mid-single digits in real terms over the next decade while continuing to earn very attractive returns on capital.  Having put a lot of work into understanding and modelling underlying end market growth drivers we believe that TI should be able to grow at a multiple of 2x to 3x GDP over time. 

Underlying this growth are significant tailwinds in two of its largest end markets, automotive and industrial, which combined account for >60% revenue.  In industrial applications the demand for analog chips should grow as automation of manufacturing continues and the complexity and capability of new generations of industrial machinery increases.  In automotive, the analog industry will benefit from the increase in vehicle automation and electrification.  For example, there is on average $375 of semi-conductor content in an internal combustion car whereas a typical electrical vehicle has $1,200 worth of content. 

Further potential upside for the company comes from the fact that analog engineers are a scarce ‘commodity’ and becoming scarcer.  The majority of university electronic engineering curricula focus on digital design, not analog.  Furthermore, most analog engineers learn their skill through on-the-job training, and consequently may take two to three times as long to become as productive as their digital engineer counterparts.  Most of the best analog design engineers have 20 to 30 years of experience in a specific analog field – they learn what works from working on designs then constantly improving them. 

A decade ago, the engineer at the customer that TI was working with was usually an analog savvy engineer.  Today, the engineer at the customer is largely a digital engineer who needs to buy in analog technology.  This has resulted in an asymmetry in the skills and capabilities that TI possess versus what the customers are capable of doing.  One consequence of this is that analog companies are seeing increasing pricing power.  Whereas historically customers were able to negotiate price reductions of up to 5% per annum on reorders, we have heard from several sources that analog companies are no longer giving such price reductions.  If this trend continues, it will lead to significant further margin expansion.

Finally, we think TI has the best management team in the industry.  Current management have been with the business as far back as the 1990s and were instrumental in defining and executing the company’s strategy to focus on areas where it could achieve durable competitive advantage.  In addition to the high marks that we assign management for strategic and operational ability, we also rate them highly with regards to stewardship of shareholder capital.  Management make the point that they only buy back stock ‘when discounted cash flow value exceeds stock price’.  We think this is logically obvious, but it is a concept most management teams seem to struggle with.  Furthermore, the company has a track record of walking away from acquisitions if it does not believe it can meet its >10% internal rate of return hurdle.

Given the company’s high barriers to entry, tailwinds to growth, improving pricing power and value-orientated management, we don’t think it takes one of the company’s calculators (which they do still make!) to work out that TI is an attractive investment. 

However we would suggest TI is only suitable for the long-term orientated investor given most analyst and investor concerns seem to be focused on short term issues.  On the contrary we view these ‘issues’ as either intrinsic characteristics of the industry or sound strategic management decisions aimed at increasing both the company’s moat and valuation. 

TI and the broader Analog/MCU industry are multiple quarters into a downturn and are currently tracking below the long-term trend-line, which is leading to analysts maintaining sell ratings on the stock given their expectation for margin degradation through 2024.  Q4 sales outlook of $4.1bn was below consensus at $4.5bn, and implies the 5th consecutive quarter of year on year sales declines, worse than prior cycles that have lasted 2-4 quarters.

Many investors and analysts are also concerned about the company’s commitment to spending ~$5bn per annum on capex through to 2026 as it looks to strengthen its competitive advantage and respond to customers who are in search of geopolitically dependable capacity.  For reference 2019-2022 annual capex averaged $1.7bn.  This $5bn capex number is gross meaning it does do not include ITC or grants from the CHIPS Act which could be worth up to $600m per annum.  These investments are reducing free cash flow margins to a 17 year low of just 9% versus a target of 25-35% of sales.  

Capital return, on the other hand, remains muted in the near-term, particularly relative to history, with the company repurchasing only $46m of stock in the quarter (or ~$1.1bn on a twelve month  basis which is down 63% year on year. The company recently announced a 5% increase in its dividend which compares with the past 5 year average dividend increase of ~15%.

We believe however these concerns obscure from what is really important – increasing long-term free cash flow generation. 

“The best measure to judge a company’s performance over time is growth of free cash flow per share, and we believe that’s what drives long-term value for our owners.” - Rich Templeton, chairman of the board

We believe the strategy being employed by management is the best way to achieve this.

Catalyst

FCF generation once company exits current capex cycle.