TXN - Semiconductors Winners And Losers At The Start Of 2024

2024-01-04 10:41:16 ET

Summary

• Semis had an outstanding year in 2023, and what happened in 2023 offers important lessons as to what to expect in 2024.
• AI and the outlook for an industry rebound in 2024 played important roles, but it was the Fed that continues to play the leading role.
• The semiconductor industry faces a number of challenges in 2024 and beyond, China being foremost in the long run.
• Semis are unlikely to add an encore in 2024 after a strong 2023, but the prospect of Fed easing should prevent too much damage, although with lots of volatility.

The year 2023 was a year of strong gains for most semiconductor stocks, a year after incurring substantial losses in 2022. This happened even though the semiconductor industry itself contracted after years of expansion due to a downturn that has turned out to be deeper and longer lasting than expected heading into 2023. An anticipated return to growth in the industry this year helped, as did the promise of artificial intelligence or AI. However, it was the Federal Reserve that deserved most of the credit for shaping 2023 the way it turned out, which offers important lessons with 2024 upon us. Why will be covered next.

Why the Fed is key to where semis are headed

A past article from a year ago remarked how what happened in 2022 could serve as a guideline for 2023. Specifically, the article noted how semis performed poorly even though most companies and the industry as a whole did very well with sales and profits hitting record highs in 2022. Yet all this was trounced by the Fed and its monetary tightening, which drove down growth stocks, particularly tech with semis a subset.

The result was a down year for semis in 2022. The iShares PHLX Semiconductor ETF (SOXX), for instance, lost 35-36% in 2022 and that's with a late rally reducing losses for the year. This rally was in turn triggered by the Fed signaling in October 2022 that its policy of tightening was past its peak and monetary policy was set to become less restrictive.

The ensuing rally has yet to end. For instance, the chart below shows how SOXX bottomed in October 2022, followed by a rally that continues to this day. The surge in semis took place even though the semiconductor industry entered a downturn that caused earnings to drop for most involved in the industry. SOXX gained 65-66% in 2023 to put it above where it was in late 2021, the downturn notwithstanding.

This showed the power Fed policy has on semis. It was Fed policy that drove down semis in 2022 despite record earnings and it was Fed policy that drove up semis in 2023 even though earnings shrank due to a downturn. In fact, 2023 was a worse-than-expected year for the industry with the downturn more severe than anticipated heading into 2023.

The general expectation was that the downturn would end in H1 2023, followed by a strong rebound in H2 2023 to make up for lost ground in H1. Losses would narrow down, but semiconductor sales were still expected to decline in the low-to-mid single digits in 2023, depending on the source used. This would pave the way for a record-setting 2024.

However, the prior article was skeptical of these forward projections for a number of reasons, specifically the impact of COVID-19 stimulus and the pulling in of future orders in China, both of which inflated real demand in the semiconductor market in recent years. This skepticism turned out to be warranted because the downturn has been significantly worse and longer lasting than expected.

The rally was increasingly being driven by fewer and fewer semis, specifically those stocks related to high-performance computing or HPC and AI in particular. So while semis managed to gain in Q2 after a very strong Q1, they might not have if not for the huge boost from AI, especially in May. Yet AI could only carry semis so far, something a number of bellwether companies acknowledged when they pointed out that the downturn was more severe than they expected. An earlier article at the start of H2 2023 foreshadowed struggles for semis in Q3.

The uptrend that had lasted three quarters by then came to an end. Semis turned south in Q3, weighed down by various headwinds, including disappointing quarterly results and guidance due to a downturn that was proving to be more severe than anticipated. Forecasts were adjusted by calling for the semiconductor market to decline in the low teens in 2023.

In addition, China showed it was making more progress in the field of semiconductors than anticipated, despite the presence of export controls. This is likely to lead to more competition for established players and could even result in a fracturing of the market for semiconductors along geographic lines, especially in the long run. A previous article covers this aspect in greater detail.

It was thus up to the Fed to come to the rescue once again. Note how in the chart above semis soared higher on November 1 after declining for months, which happened to be the day the Fed essentially signaled that if the Fed Funds Rate were to change from that point on, it would be with interest rate cuts and not with hikes. The most recent dot plots from the Fed sees the current Fed Funds Rate of 5.25-5.50% dropping by 75 basis points by the end of 2024.

Market expectations are for an even more dovish Fed. Fed Funds futures believe interest rates are most likely to hit 3.75-4.00% in 2024 with a probability of 39-40%, which implies a cut of 150 basis points with the first coming as soon as March. All this send stocks soaring. SOXX, for instance, gained 30.2% in November and December after the Fed came in more dovish than expected, completely reversing the decline of the preceding three months.

So if not for the Fed, semis would not have done quite as well in 2023. SOXX gained 65-66% in 2023, way ahead of the SPDR S&P 500 ETF (SPY) or the Invesco QQQ Trust (QQQ), but most of it came in H1. Semis were going down and it took the Fed to reverse the slide. Compared to current Fed policy, an industry downturn, disappointing quarterly results/guidance and AI are not irrelevant, but of secondary importance, something worth keeping in the back of the mind as we head into 2024.

Who was hot and who was not

Still, not all semis had a good year. A look at SOXX, or the companies present within in, shows why some, particularly those exposed to HPC, did much better than others. The 30 companies in SOXX are Advanced Micro Devices (AMD), Broadcom (AVGO), Nvidia (NVDA), Qualcomm (QCOM), Intel (INTC), NXP Semiconductors (NXPI), Micron (MU), ON Semiconductor (ON), Texas Instruments (TXN), Lam Research (LRCX), Marvell (MRVL), Analog Devices (ADI), Applied Materials (AMAT), Microchip (MCHP), KLA Corp (KLAC), ASML (ASML), TSMC (TSM), Monolithic Power Systems (MPWR), Skyworks (SWKS), Entegris (ENTG), Teradyne (TER), STMicroelectronics (STM), Qorvo (QRVO), Lattice Semiconductor (LSCC), Rambus (RMBS), MKS Instruments (MKSI), United Microelectronics Corporation (UMC), ASE Technology (ASX), Wolfspeed (WOLF) and Axcelis (ACLS). The table below shows their gains or losses.

Source: iShares

NVDA was the winner by a landslide with a gain of 238-239% in 2023, more than 100 points ahead of second place AMD. The two, the former in particular, are the leading suppliers of server GPUs, suitable for use in AI applications, which drew stock buyers as a way to play the AI story and HPC in general. On the other hand, NVDA's gains in H2 were rather modest with most 2023 gains coming in H1.

This is true for most semis, although some like INTC bucked the trend by doing better in H2 than H1. INTC also vied with AMD as the best performers in Q4. At the other end of the spectrum is WOLF, the only out of the 30 stocks above to finish 2023 in negative territory. Yet it was WOLF who, with the exception of MKSI, was the one with the most momentum heading into 2024 with a gain of 20% in December.

Could the semiconductor market come up short in 2024?

Fed policy helped power semis higher. AI did as well, but to a lesser extent, although AI may play a greater role when it comes to individual names. Another factor is the outlook for a better 2024 than 2023 for the industry. Keep in mind that this was already the case heading into 2023. As mentioned earlier, the semiconductor downturn in demand was forecast to be a relatively short and shallow one with the bottom in H1, followed by a strong rebound in H2, which would pave the way for a record-setting year for the industry in 2024, eclipsing the old high set in 2022. In other words, the outlook is for a V-shaped recovery.

Yet the downturn has been worse than anticipated by most. Forecasts have therefore been scaled back accordingly, but most still call for the semiconductor market to grow in the teens in 2024. For instance, WSTS predicts the market will grow by 13.1% YoY to $588B in 2024, which would surpass the current record of $574B set in 2022 and negate the down year in 2023. Some market segments like memory are expected to do ever better with growth rates as high as 40+%.

In contrast, forecasts prior to 2023 were calling for a market well into the $600B range. There is also a long-term forecast calling for the semiconductor market to reach one trillion in worldwide sales by 2030. Keep in mind the semiconductor market would have to accomplish this without the same level of global stimulus and inventory building in China that helped drive up demand as mentioned earlier.

Still, semiconductor demand in general has been below forecasts. There are signs demand is starting to improve in some segments like memory, although one could argue that the memory market has fallen so low the only direction for it to go next is up. On the other hand, other market segments have the potential to decline further. For instance, demand for automotive chips did not contract in 2023, unlike the semiconductor market as a whole.

Automotive grew in 2023, but there are increasing signs the downturn that it has eluded is starting to catch up. The market for electrical vehicles, which has driven up demand for automotive chips, has started slowing down after several years of explosive growth. This has trickled down into individual stocks like ON, which sold off in October after issuing a weak outlook. China is also increasingly implementing its own version of "derisking/decoupling" by shifting orders to local suppliers, including for compound semiconductors like silicon carbide.

Expectations are running high for 2024, but one should not dismiss the possibility the semiconductor market may not grow as fast as expected. If semiconductor demand comes up short in 2024, as it did in 2023, then that could extent the downturn, especially with semiconductor supply set to ramp up in 2024 and the coming years. A number of suppliers of semiconductor manufacturing equipment like LRCX have been able to outperform because many chipmakers have not reduced capex spending to the same extent as in previous downturns.

Some, especially in China, have actually increased spending. As a result, a total of 42 new fabs are expected to come online in 2024, including 18 in China, with more expected in the following years. A key factor enabling this fab building spree is government support, which has incentivized the building of more and more fabrication capacity in various countries around the world.

Source: SEMI

SEMI, for instance, predicts global semiconductor fabrication capacity will surpass 30M 8-inch equivalent wafers per month in 2024 as shown in the table above. In comparison, capacity was around 20M or 10M less as recently as 2020. The risk here is that supply growth could outpace demand growth, which would create an imbalance between supply and demand and put downward pressure on prices. It may not happen in 2024, but with all the fabs in the planning/building stage, it becomes increasingly difficult to see how this issue will not come to a head a few years down the road.

A swing factor in determining whether the market stays in balance or falls into a semi-permanent state of glut is what happens in China. Many long-term projections for the industry assume current restrictions on the supply of semiconductor equipment, especially at the leading edge, will hobble China's ability to increase local chip production in the coming years.

However, as shown with the release of recent products from the likes of Huawei, China may find ways to progress. This could affect the shape of the recovery for the industry. The current expectation is for a V-shaped recovery, but China may turn it into an L-shaped one or even worse, depending on its ability to increase production and advance technologically.

Could the Fed be knocked off course?

As stated before, the Fed is arguably the single most important swing factor out there for semis. Its fortunes rests in large part on what the Fed does in 2024. However, it needs pointing out that the Fed can be influenced by outside factors, which may force it to alter Fed policy in ways that the market does not currently anticipate.

While the latest dot plots were very much on the dovish side, it's worth pointing out that dot plots can be hit and miss in terms of forward projections. The Fed has deviated before from past dots plots, so if it happens again in 2024, then it would not be the first time. It's prudent to think twice about attaching too much weight to forward projections of interest rates since there are many factors involved that can force them up or down depending on the situation.

The chief wildcard here is sticky inflation because of its ability to influence decision making at the Fed. Inflation is a somewhat controversial topic, particularly as to how it is measured in a way that does not seem inconsistent with the cost of living. Still, most would probably agree inflation has come down, which gives the Fed the leeway to adjust policy in a way that favors the stock market.

At the same time, this leeway is subject to change. Inflation could rise again, which may force the Fed to come up short with its projections of a 75 bps cut in interest rates, let alone market expectations of a 150 bps cut. If inflation were to rise again, it will most likely be due to geopolitical issues, especially in the Middle East.

If say the situation in the Middle East deteriorates more than it has, then that could create the necessary conditions for inflation to get worse. A resulting increase in the cost of energy would put upwards pressure on inflation. So too would supply chain disruptions because shipping routes become inaccessible for long periods of time.

Even so, the Fed can still ease policy, provided it is willing to accept the consequences that doing so could further inflame inflation, but probably not as much as the market seems to be hoping for with a 150 bps cut. Expectations may be too high heading into 2024. If expectations are too high, then they need to be adjusted downwards. This argues against being a buyer of semis, especially after soaring as high as they did in the last two months of 2023.

Investor takeaways

The year 2023 was a very strong year for semis. This alone puts the odds against a repeat in 2024. If anything, semis are probably due for a correction in the short term after soaring higher by as much as they did in the closing months of 2023. Semis have essentially rallied higher for the last five quarters, with the exception of three months in the late summer of 2023 and early fall of 2023.

Not only is the rally long in the tooth, but the outsized role of the fed in pushing semis higher is unlikely to be healthy in the long run. The Fed enabled a stock market rally with SOXX, for instance, gaining 30+% in November and December, but it bears asking why the Fed decided to turn so dovish late in the year.

Some could see the lowering of interest rates as proof inflation is under control, but others could interpret it as the Fed sensing the economy is weaker than believed and cuts are needed to provide relief. While the current Fed Funds Rate of 5.25-5.50% is much higher compared to the last 15 or so years following 2008/2009, it is not that high on a historical basis. It's similar to what is was in the nineties and that was with inflation much less an issue like it is right now. The average is 4.50-4.75%, which is where the dot pots want to go in 2024, even though the Fed itself has stated that the battle with inflation is not finished.

Further clouding the outlook is that there is a significant risk inflation may be pushed upwards by geopolitical events that have yet to be resolved and for which a solution may be difficult to achieve, particularly in the Middle East. If the situation there stays as is, which is for a gradual escalation in the region, then it may be too much to ask for no increase in inflationary pressure in 2024. While the Fed can still proceed with rate cuts, it does raise the bar for delivering on what the market is hoping for.

For the semiconductor market to put an end to the downturn on the way to a record-setting year in 2024 also looks like a high hurdle. It depends on which forecast you use, but most are predicting worldwide semiconductor sales will set a new record high in 2024. This after a year which came in lower than predicted due to a downturn that has shown to be deeper and longer lasting than anticipated.

Forecasts for 2024 may be making the same error as in 2023 by once again overestimating the level of demand for semiconductor chips. The year 2023 saw the rise of AI, which greatly helped the stocks of selected names, particularly those exposed to HPC. Yet AI cannot move the needle in a major way for the industry simply because it lacks the size to do it, at least for now.

Quite a few of the supposed AI plays are borderline AI plays at best. Adding to the confusion is that any company can refer to AI, even though very little of what they do has anything to do with actual AI. The year 2024 could see the fall of those AI plays that rode the AI wave higher, but whose actual quarterly results don't show any meaningful impact from AI.

The industry may also be underestimating the level of supply coming to the market, especially in China. There are a number of industry reports whose forecasts assume trade restrictions on the sales from the likes of ASML, AMAT, LRCX and KLAC will stymie certain Chinese chipmakers, if not put them out of business. For instance, at least one report assumes China's YMTC will be forced to leave the market due to lack of access to advanced production equipment.

This would reduce the amount of supply of NAND chips, which on paper could have a positive impact on NAND prices by bringing supply more in line with actual demand. It's also part of the reason why the report calls for the memory market to boom in 2024. Other reports assume China's SMIC will not be able to increase production of leading-edge chips and what little it does have will be hobbled by poor yields and insufficient capacity.

Yet, both of the aforementioned companies continue to bring new products on the market when they should not be able to based on these industry projections. Both are also rapidly expanding production capacity by investing tens of billions, which goes against the notion of them not sticking around as assumed by various industry reports. SMIC, for instance, was one of the few chipmakers to increase their capex budget to well over 100% of sales, which is much higher than other foundries like TSMC and UMC.

Granted, it's technically possible Chinese chipmakers will become handicapped by the lack of access to necessary tools and technology needed to produce semiconductor chips, depending on what China can achieve on its own. Still, many industry projections are predicated on the assumption export controls will keep China in check and that China will not be able to come up with alternatives to whatever has been denied to it. What some of these reports refer to as "established fact" is actually just an assumption on their part. And if the premise is flawed to start with, then the conclusion should be doubted.

It's very risky to have so much riding on this one assumption regarding China because if China were to succeed, a whole bunch of industry names could have a major challenge to confront a few years from now. SMIC, for instance, has three new gigafabs, defined as having a capacity of 100,000 12-inch equivalent wpm or more, that have recently started production or about to in 2023/2024. It will take time to reach full capacity, but this capacity increase is more than what all but a handful of chipmakers have in total worldwide.

With all the above in mind, the case for semis as an investment is a tricky proposition to make. There are forces pushing semis upwards, chiefly in the form of the Fed or its current policy to be more exact. On the other hand, there are also forces out there that could propel semis downwards, including several that have the power to affect inflation and monetary policy by extension. The former have been able to have the upper hand on the latter, but it required a promise on the part of the Fed that it may not be able to honor.

These conditions favor volatility with big stock moves in both directions, depending on which side has the upper hand at the moment. The use of long/short pairs can be useful during these times since they can be profitable, whether stocks go up or down. Suitable pairs include AMD/NVDA, AMD/INTC, SWKS/QRVO, AMAT/LRCX, ADI/TXN and UMC/TSM.

If someone still insists on placing directional bets, then HPC names are the best bet. Semis most likely to underperform are probably those tied to the automotive sector, especially those with heavy exposure to China. China is also the most likely source of curveballs for the industry, especially if China succeeds in progressing faster than the industry expects it to.

Nothing is set in stone, but it is unlikely semis will able to achieve a repeat of 2023 in 2024. Most semis are sitting on huge gains for which an encore is not likely. However, while temporary drawdowns are possible, especially after the recent run-up heading into 2024, semis are unlikely to stay down for too long, not with the Fed holding out the prospect of easing in front of the market.

Trillions of liquidity have been injected in the last 15 years and stocks will be hard to keep down as long as this liquidity needs some place to stay. On the other hand, it also means most people will see the cost of living go up. This should remain the case as long as this liquidity is not withdrawn, whether by the Fed itself, a major financial crisis or anything that has the ability to drain or even outright destroy large amounts of liquidity and in the process of doing so deflates asset prices, including stocks.

The path of least resistance is for semis to stay flat to slightly down by the end of 2024. The Fed is likely to come up short in terms of easing and the semiconductor market is likely to disappoint in terms of growth, but the prospect of continued Fed easing should prevent too much damage. Big stock moves in either direction, up and down, are to be expected in 2024. Moreover, if there is one thing to look for in 2024, it's most likely volatility.

For further details see: