WOLF - Semiconductors Winners And Losers At The Start Of Q4 2023

2023-10-05 03:27:06 ET

Summary

• Semis dodged a bullet in Q2 thanks to AI, but there was no such luck in Q3 with most semis ceding ground, particularly in September.
• Q3 was a quarter of change in various ways, but in general, the tailwinds that enabled the rally in the past year lost strength and the headwinds gained strength.
• China is in the spotlight heading into Q4, especially since it has the potential to bring about lasting changes to the semiconductor industry that could prove to be permanent.
• The potential is there for more surprises, like the recent one from Huawei, as semis head into Q4, but there are also clues as to what's likely to come next.

There were signs semiconductor stocks were on their last legs in Q2 2023, but the uptrend that started in Q4 2022 came to an end in Q3 2023. Semis as a whole lost ground for the first time in Q3 2023 after three consecutive quarters of strong gains. The iShares PHLX semiconductor ETF ( SOXX ), for instance, lost 6.6% in Q3. The tailwinds that enabled the rally off the October 2022 lows lost strength for various reasons, pulling down most semis with the exception of a few who seem to be in a better position heading into Q4. In contrast, the headwinds gained strength with China playing a leading role. In fact, China may have heralded changes that could prove to be lasting for the semiconductor industry. Why will be covered next.

The recent trend in semiconductors is no more

A previous article at the start of Q3 2023 ended by stating that "nothing goes up forever" after concluding that while semis remained in an uptrend, semis were running on fumes with the rally driven by a decreasing number of stocks and an increasing number of stocks sitting out the rally. This turned out to be pretty much spot on because the uptrend in semis that started almost one year ago came to an end in Q3. The chart below shows how SOXX started its rally in October 2022, only to have the uptrend broken in Q3, September to be exact.

As elaborated on in the last article, semis as a whole actually struggled for much of Q2 2023, but artificial intelligence or AI came to the rescue by triggering a resurgence in investment interest in semis, and a few selected names in particular. This allowed semiconductor stocks as a group to finish Q2 with a double-digit gain after struggling for much of Q2.

SOXX, for instance, managed to erase its losses from earlier in the quarter to wind up with a gain of 14.8% in Q2. SOXX continued to build on the gains from Q2 with a good start in Q3, but it was unable to keep it going, resulting in a loss of 6.6% in Q3. Nevertheless, SOXX is still up 36.1% YTD heading into Q4 after posting gains of 9.2%, 27.7% and 14.8% in Q4 2022, Q1 2023 and Q2 2023, respectively.

However, while most semis lost ground, some did not. An ETF like SOXX with its 30 stocks shows this as much. The 30 companies in SOXX are Advanced Micro Devices ( AMD ), Broadcom ( AVGO ), Nvidia ( NVDA ), Intel ( INTC ), Texas Instruments ( TXN ), NXP Semiconductors ( NXPI ), Analog Devices ( ADI ), Qualcomm ( QCOM ), Microchip ( MCHP ), ON Semiconductor ( ON ), Micron ( MU ), Marvell ( MRVL ), KLA Corp ( KLAC ), Applied Materials ( AMAT ), Lam Research ( LRCX ), TSMC ( TSM ), ASML ( ASML ), Monolithic Power Systems ( MPWR ), Skyworks ( SWKS ), Teradyne ( TER ), Entegris ( ENTG ), Lattice Semiconductor ( LSCC ), STMicroelectronics ( STM ), Qorvo ( QRVO ), Rambus ( RMBS ), MKS Instruments ( MKSI ), United Microelectronics Corporation ( UMC ), Axcelis ( ACLS ), ASE Technology ( ASX ) and Wolfspeed ( WOLF ). Note that RMBS and ACLS were added to SOXX, having replaced Silicon Laboratories ( SLAB ) and Synaptics ( SYNA ). The table below shows their gains or losses.

Source: iShares

Who was hot and who was not in the world of semiconductors

From the list one can observe a change in trend. Semis decelerated after rallying earlier. Most semis retain their big gains from the rally over the past four quarters, but the gains are skewed towards the first two rather than the last two. Most semis have lost steam and off their highs. Semis as a whole are heading into Q4 without the momentum from earlier in the year.

One notable exception that did not lose any gains because it didn't have any to lose is WOLF. WOLF never really participated in the semiconductor rally of the past year like the others did for at least a while. So not surprisingly, WOLF is ranked dead last, both YTD and TTM. Others did rally at some point in the last 12 months and were up at one point, only to lose all their gains. This includes stocks like TXN and QCOM. These three names are the only ones to be down compared to a year ago after the rally during the preceding months.

In general, the month of September was a particularly bad month for semis with pretty much everyone suffering setbacks. A notable exception to this was INTC, a name which is often seen as the main beneficiary of U.S. government efforts to increase local chip manufacturing through initiatives like the U.S. Chips and Science Act.

Not all semis are included of course, but in the list above just three out of thirty stocks ended Q3 with gains, namely MU, INTC and NVDA. INTC, in particular, has been able to trend higher at a time when most semis have faltered despite disappointing results from INTC. MU, the most recent one to report earnings, has also disappointed, especially in terms of guidance. However, rising memory prices, DRAM in particular, have raised optimism the worst may be behind for MU.

All others finished Q3 with losses, led by WOLF, MKSI and ASML. On a YTD basis, NVDA remains on top, a position it has held for all of 2023. Investors have flocked to NVDA due to AI, especially as the leading supplier of server GPUs. Outstanding earnings growth from NVDA have reinforced NVDA's position as the top AI play.

On the other hand, it's worth mentioning that most other AI plays have not fared as well as NVDA. MRVL, for instance, has yet to show much of an AI boost in its quarterly results despite being touted as an AI play, which could explain why the stock has retreated after an initial AI-induced boost. In fact, while AI remains a major factor, it lost some of its potency. This, in combination with other tailwinds that started to lose strength earlier, paved the way for semis to have their first down quarter in a year.

What conspired to bring down semis

The last 12 months or so have been very good for semis, September notwithstanding. What makes the rally in semis even more remarkable is that it happened even though the semiconductor industry went into a downturn during this time. This has led to, for instance, major declines in earnings for most semis.

Three factors helped makes this possible. A less restrictive Fed policy, the expectation of a relatively quick end to the industry downturn, followed by a strong rebound and AI. It's no coincidence, for instance, semis started to rally in mid-October 2022 because that's when the Fed started to signal its intention to start pivoting towards less monetary tightening with inflation starting to recede.

This provided relief for a sector that was badly hit by interest rate hikes in 2022. At one point, expectations were for the Fed to begin cutting rates as soon as late summer/early Fall 2023. This did not happen, which tempered the rally, but both Fed Funds futures and the Fed's own dot plots still call for interest rates to be lower than where they are now in the next few years. Futures suggest the first rate cut is most likely to come in mid-2024.

Similarly, the industry downturn has been worse than anticipated. Most forecasts, for instance, expected worldwide semiconductor sales to contract in the low single digits in 2023 after a record year in 2022, but more recent forecasts have revised that to the low teens. What has not changed, though, is the expectation of a major expansion in 2024. Forecasts have lowered their expectations, but the expectation is for growth in the low teens in 2024, which would essentially negate the dip in 2023. In other words, a V-shaped recovery is expected.

Still, while both factors remain and continue to serve as tailwinds for semis, they have lost some of their strength. The Fed, for instance, has eased up on tightening, certainly compared to 2022 when 75bps hikes were the norm, but not as much as expected due to inflation being more stubborn than expected, which affects how much room the Fed has to move.

Can AI keep boosting semiconductor stocks?

However, up until fairly recently, AI was able to make up for the other two. AI requires semiconductor chips and increased use of AI should raise demand for semiconductors in general. AI thus provided a lift to semis and actually kept the sector afloat when the other aforementioned tailwinds started to recede. In fact, if not for AI, and the outsized gains for a small group of stocks, NVDA especially, semis as a group would very likely have ended Q2 with losses as noted in the previous article.

However, there are signs people are starting to temper their enthusiasm about AI. This paved the way for semis to decline as AI was essentially the last pillar supporting the rally. As mentioned earlier, with the exception of NVDA, most supposedly AI plays have seen little to no real gain in quarterly results from AI. Keep in mind there have been many who jumped on the AI bandwagon by touting their AI credentials, even if their connection to AI was borderline at best.

Furthermore, a number of prominent companies and top executives have come out to explicitly temper expectations of what AI can bring to the table. There are a number of examples, but probably none so more than TSM, which is noteworthy since TSM is often seen as the bellwether for the semiconductor industry. TSM was actually fairly optimistic heading into 2023, but TSM was forced to revise its stance by progressively lowering its outlook as the year went by.

TSM originally called for FY2023 revenue to grow in the single digits, but the most recent outlook sees revenue shrinking by about 10% YoY after stating semiconductor demand was weaker than anticipated. TSM also stated in the most recent earnings call that "while we have recently observed an increase in AI-related demand, it is not enough to offset the overall cyclicality of our business."

AI accounts for about 6% of total revenue according to TSM. AI is much less important as a driver of semiconductor demand than say the mobile industry. TSM believes AI may grow at a 50% CAGR in the next 5 years to reach the low teens, but TSM tempered this by stating the future cannot be accurately predicted and AI demand could actually fall off from recent levels.

So while AI is definitely a plus for the industry, especially for a few selected names like NVDA, AI cannot hold up the industry on its own as AI lacks the size to do so. It may have been a coincidence, but it's interesting to note how the peak in semis this year roughly coincided with the aforementioned statements from TSM concerning AI in terms of what it can and, more importantly, what it cannot do.

Why China could shape the path forward for semis in the short term and possibly the long term

Fed policy, an industry downturn and AI all had a role to play. However, an argument can be made that in Q3, it was China that was the most consequential for the semiconductor industry, especially in terms of long-term ramifications for the industry. Unlike say Fed policy or industry downturns that should prove fleeting over time, China has the potential to bring about changes to the semiconductor industry that could be more lasting and more permanent in nature.

The semiconductor industry is now down, but that does not change the fact that the 2017-2022 period was a period of great expansion, arguably the greatest ever. There are many factors cited as leading to the rise in semiconductor demand, but an article from early 2023 identified two factors that overshadowed all others.

One was unprecedented fiscal and monetary stimulus to combat COVID-19. For instance, the Fed balance sheet went from around $4 trillion in early 2020 and before the COVID-19 outbreak to almost $9 trillion by mid-2022. The U.S. national debt was around $22 trillion in late 2019 and before COVID-19 struck. The debt is now estimated to be around $33 trillion. All this liquidity served to boost demand for all things, semiconductors included.

The second factor pointed out in the article was the building up of semiconductor inventories in China due to the trade/tech war with the U.S. For instance, at one point, China's imports of semiconductors reached an amount equal to about 80% of worldwide semiconductor sales. This for a country that itself is one of the biggest producers of semiconductors, although it needs pointing out that a number of the chipmakers in China are foreign companies with local fabs like Samsung, SK Hynix, TSMC, UMC and so on.

It's also worth mentioning that while both drivers have receded, they have not gone away completely. For instance, while official numbers have yet to be released, the U.S. government is estimated to have spent around $2 trillion on deficit spending in the recently concluded fiscal. This amount exceeds every other year except the COVID-19 years, including the financial crisis of 2008/2009. This is equal to four times the $500B in worldwide semiconductor sales projected for 2023.

There's reason to believe that while China has reduced inventory building, China continues to stock up on specific items like chips for HPC, AI in particular, and certain semiconductor manufacturing equipment. The former is often done indirectly through third parties to hide its eventual destination in China. For example, NVDA chips, which are ostensibly sold to a non-China buyer, are rerouted through an elaborate network to get past existing export controls. It's possible the recent explosion in demand for NVDA's chips may in part be due to this, which could explain why others are not seeing the same AI benefit that NVDA is seeing.

Why China may have closed the door on a V-shaped recovery for the semiconductor industry

As mentioned earlier, one of the reasons why semis have done so well in 2023, even though most reported big declines in earnings, is because of expectations of a V-shaped recovery. The general consensus at the start of 2023 was that while the semiconductor industry was in the midst of a downturn, the bottom would come in H1 2023, followed by a strong rebound in H2 2023.

This would narrow the YTD losses to the low single digits for 2023 and pave the way for a record-breaking year in 2024. Semis were priced with this in mind, which helps explain why the market bought up semis despite poor earnings. The downturn has been more severe than expected and demand has been weaker than anticipated, which has resulted in forecasts being lowered accordingly, but a V-shaped recovery has not been written off.

However, China may have significantly lowered the possibility of a quick recovery for the industry, if not entirely, especially in light of recent events. As mentioned in the article at the start of Q3, China had begun making moves in Q2, marking perhaps the beginning of a new chapter in the ongoing tech war between the U.S. and China, which started in May 2019 with Huawei being put on the U.S. Entity List.

China, for instance, imposed sanctions on MU and it announced its own export controls, but it was Huawei that stole the show in September with its first 5G smartphone in years, which contributed to the decline in semis in September. Keep in mind the U.S. government imposed various export controls designed to limit the ability of Huawei, and China in general, to produce semiconductors, especially at the leading edge.

The industry was thus taken by surprise by Huawei's ability to release products using mostly domestic technology that China was not thought to possess. For instance, Huawei worked with SMIC to produce a SoC with a build-in 5G modem. This impacted a number of semis seen to be affected by this revelation. TSM, for instance, lost 8.3% in September with SMIC offering an alternative to TSM. While SMIC is currently not in the same league as TSM, it has the potential to become its foremost competitor over time.

A lot of attention was focused on the SoC from Huawei, but the recently released smartphone showed a number of other advancements that are very relevant to the semiconductor industry, even if they did not get the attention they deserve. For instance, not much attention was paid to 5G RFFE modules, which China must have produced to enable 5G frequency support, a major milestone since almost everything these days needs an Internet connection.

Huawei introduced a new wireless technology called NearLink, which is touted as an alternative to Bluetooth and Wi-Fi to a lesser extent with superior specs, at least on paper. All this is very relevant to stocks like QRVO, SWKS and AVGO. China must have developed EDA tools that support the design rules used by the new process nodes to design chips like the recent 5G SoC, which opens the door for other chip designs. Al this means extra competition at a time when many industry players are dealing with insufficient demand.

In short, the return of Huawei shows much progress has been made in China, suggesting existing players in the semiconductor industry need to get ready for more competition. If demand is an issue, as it is for many at the moment, then the last thing a company needs is for new competition to emerge, which will only serve to make market conditions more difficult.

How China could affect the semiconductor market in the short run and potentially in the long run

The semiconductor industry is highly cyclical with periods of booms followed by busts and vice versa. To help mitigate the fallout companies can resort to various measures to ease the transition. One of the more common strategies used to deal with, for instance, oversupply is to agree to production cuts to bring supply in line with demand.

This is often used in say the DRAM industry. However, this only works if all players are on board. If only some cut back and others raise production, the former risk losing market share. This is one area where China could shake up the industry. China is interested in increasing its production output of all sorts of semiconductors and agreeing to production cuts to help balance the semiconductor market is not likely to be something China will be keen on.

The emergence of China as a semiconductor producer makes it less of a consumer and more of a competitor to existing players with all sorts of consequences for companies. In addition, assuming China can greatly increase its fabrication ability, China will make it harder to end industry downturns. Current players will likely not be able to employ the same tactics used to counter downturns.

The silver lining here is that increased competition from China should spur progress, especially in those segments currently controlled by a few or even a single supplier, which ultimately benefits the average consumer who will be able to get more for less.

How much of a challenge is Huawei and chip development within China?

However, while China's progress is a headwind, it's worth mentioning that it is difficult to quantify by how much exactly as much remains unclear. For instance, while SMIC has been able to produce a chip that is roughly equal to those produced using 7/6nm process nodes in terms of PPA, it remains unclear what the yields are like, which could determine how competitive the new chips are against those from existing players.

It's also unclear as to how much fab capacity China has available for leading-edge nodes and if that capacity can be expanded with what China has access to. Nevertheless, the same industry sources that tipped off Huawei's return to the 5G market are saying to expect additional handsets from Huawei, including mid-range ones. This suggests yield and capacity are not a problem since dedicating limited fab capacity to producing lower margin chips instead of higher margin ones does not make sense if capacity and/or yield is insufficient.

Still, many assume U.S. export controls will restrict China's ability for further advances. For instance, China can no longer order high-end DUV/EUV lithography machines from ASML. This is why a recent forecast from IDC predicted that China's share of chips produced using leading-edge nodes will remain in the 1-2% range in 2023-2027.

But this is based on the assumption that China will not be able to develop replacements for key technology, tools, equipment and so on needed to produce semiconductors. This assumption could be misplaced since China is making rapid progress in coming up with replacements based on all the technical papers coming out of China. This includes systems similar to existing ones, but also novel ideas like SSMB-EUV and 3D-DRAM. On paper, these have major advantages compared to existing solutions, which could become a gamechanger if China succeeds that is.

What makes it difficult to assess the ultimate impact on semis is the lack of reliable information as to where China really is. There is a real need for more information on China and there are many who are looking for it, but finding it is not so straightforward. There are lots of sources, but whatever information is out there is more often than not unreliable. For example, one should really question how a supposed expert on China matters can have access to confidential information about specific Chinese companies without having any link whatsoever to that company and when the company in question is a strict follower of the idiom of "loose lips sink ships".

It doesn't help that an entire cottage industry has sprung up catering to the need for more information about China. In return for top dollar, the "inside scoop" can be provided, but the information is usually highly dubious to say the least. Still, there are people willing to pay, which only serves to further muddy the waters.

Investor takeaways

An earlier article after the first quarter noted how even though semis had done very well with a defined uptrend, the situation was likely to change towards the latter part of 2023. As it turned out, things have changed. The uptrend that started in October 2022 has come to an end. The tailwinds that contributed to the rally have faded in strength for various reasons.

Still, the tailwinds remain, which makes a sustained drop in semis difficult. Keep in mind most of the liquidity injected in recent years is out there and as long as all that liquidity is in need of a place to stay, like stocks for instance, semis will find it hard to stay down for long, even if declining earnings due to an industry downturn say they should go down.

The industry is in a slump overall, but some segments are doing better than perhaps they should. For instance, semiconductor manufacturing equipment suppliers have seen their earnings shrink, but not as much as they probably should have, considering the state of the industry. This is due to most chipmakers sticking with planned investments in new fabs, which has been made possible by all the government subsidies being doled out to build new production capacity around the world.

This has kept equipment demand afloat, but it is worth pointing out that all this added capacity is very likely to come back to haunt the industry several years down the road. Still, some might want to consider those "national champions" that have been singled out for government assistance in order to carry out specific objectives set out by the politicians.

It's also worth mentioning that 2024 is an election year in the U.S. The Fed is likely to stick with its current policy of holding out the prospect of rate cuts in the future, but not make any significant changes. This is to ensure the Fed is not seen as helping one side with rate cuts. On the other hand, the Fed will also not want to raise rates as they could be seen as helping one side at the expense of the other. The Fed will thus likely stay on hold until at least until late next year to avoid being seen as favoring one side or the other.

However, this does not mean semis cannot retreat more than they have in the last few months. If, say China or Huawei is able to pull another rabbit out of the hat like they did recently, that may trigger another drawdown, especially in individual stocks. For example, if Huawei is able to quickly recapture the mobile market share it lost in recent years from Apple ( AAPL ), starting in Q4, then a host of semis stand to lose, especially those that depend on AAPL for most of their sales like SWKS.

China is likely to be the source of additional volatility for semis in upcoming quarters, whether due to the U.S. government imposing more export controls or China progressing faster than expected in creating homegrown alternatives. Sticky inflation is the biggest wildcard out there due to its ability to directly influence monetary/fiscal policy globally, but China is a more distant second.

In conclusion, semis are being pulled in opposite directions. On the one hand, there are tailwinds like the prospect of easing from the Fed, the expectation of a rebound in semiconductor demand and AI, all of which continue to make semis a favored destination. On the other hand, there are headwinds like weak earnings due to a continued industry downturn, weak macroeconomic conditions that is negatively affecting end-user demand, stubborn inflation and all the uncertainty associated with China.

This makes it hard to keep rallying, but a sustained selloff is also unlikely. The most likely outcome is for semis to see enhanced volatility in Q4 2023 with rallies followed by selloffs and vice versa, but no major gains or losses. The current stock environment is not great for semis, which argues against taking lots of risks, but if one insists on placing bets, then bet on increased volatility for semis.

For further details see: