Semiconductors are the bedrock of the digital economy, core enablers of the data revolution and at the forefront of technological innovation. The semiconductor industry could be set for exponential growth driven by the mega forces of AI (artificial intelligence) and digitalisation, as well as those of geopolitics and economic competition.

Investors looking to gain semiconductor exposure have the option of investing in multiple chips stocks directly, such as the sector’s superstar Nvidia (NVDA:NASDAQ), Broadcom (AVGO:NASDAQ), or UK chip architecture champion ARM (ARM:NASDAQ), all listed in the US but easily bought on a decent investment platform. Alternatively, there are several ETFs that target the entire semiconductor value chain, including companies focused on the design, fabrication, and assembly of semiconductors.

There are also a small number of UK companies with exposure to elements of the semiconductor industry which can be found in the table opposite.

CHIP SECTOR BACKCLOTH

After experiencing an oversupply in 2023, as the world’s economy reopened after Covid, the semiconductor sector is primed for a recovery this year as demand for AI chips continues to accelerate.

It is on exciting growth potential that many chip stocks have rallied hard. Taiwan Semiconductor Manufacturing Company (TSM:NYSE), usually simply called TSMC, is the world’s largest foundry business and its share price hit a recent record high of $149.20 for its New York-listed stock.

Advanced Micro Devices’ (AMD:NASDAQ) next-generation Ryzen 8000G chips power over 90% of AI PCs, according to company chief executive officer (CEO) Lisa Su. ARM has raised its full-year revenue guidance due its exposure to AI. The CEO of Dutch chip equipment supplier ASML (ASML:AMS) Peter Wennink believes the AI boom will fade without the Dutch company’s involvement.

At Cisco Live in Amsterdam earlier this year, the enterprise networking technology giant Cisco Systems (CSCO:NASDAQ) announced a series of hardware and software products in partnership with Nvidia. The focus of the collaboration will be on making it easier to deploy and manage AI systems using standard ethernet connections.

The majority of AI deployments use a high-performance computing alternative – Infiniband – which Nvidia acquired as part of its $6.9 billion buyout of Mellanox in 2020. However, ethernet is starting to make inroads into the AI connectivity market.

‘Companies everywhere are racing to transform their businesses with generative AI,’ says Nvidia’s boss Jensen Huang. ‘We’re making it easier than ever for enterprises to obtain the infrastructure they need to benefit from AI, the most powerful technology force of our lifetime.’

Thanks to soaring interest in ChatGPT and other large language models (LLMs) being developed by the likes of Alphabet (GOOG:NASDAQ), Microsoft (MSFT:NASDAQ), and many others, generative AI is helping to breathe new life into a stagnating semiconductor sector, which had swung from a supply shortage during the pandemic to an oversupply in some sections of the market last year.

Generative AI and LLMs require massive amounts of computing power and that is underpinned by microchips. According to the World Semiconductor Trade Statistics’ market forecast last November, a ‘robust recovery’ in demand for semiconductors is on the cards in 2024, with an expected annual growth rate of 13.1%, compared to a 9.4% decline in 2023. IDC is even more optimistic, with its projection of 20% growth this year. Recent releases from the industry have followed a pattern of companies focused on chips with more traditional applications falling behind those with an emphasis on AI.

DIFFERENT TYPES OF CHIP FIRM

Some businesses in the sector like Taiwan Semiconductor Manufacturing Company and Samsung (005930:KRX) own and operate semiconductor fabrication plants or foundries which make the physical chips themselves. Samsung and Intel also design microchips making them integrated device manufacturers. Many companies, Nvidia included, outsource manufacturing and simply design the microchips. This means they avoid the high capital costs involved in building a foundry. Beyond chip design and manufacturing, the likes of ASML and US-based Lam Research (LRCX:NASDAQ) provide crucial equipment to the industry.

ASTONISHING NVIDIA RETURNS

While Nvidia grabs most of the AI headlines, its share price is up 80% already this year after 2023’s 240% gains, other semiconductor stocks are generating plenty of AI excitement.

For example, AMD is targeting AI PCs, rolling out the MI300X chip in December 2023, its rival to Nvidia’s H100 graphics processing unit (GPU). The chipmaker is also betting on AI PCs, with president Victor Peng telling CNBC recently that the market is expanding and that the company is expecting the adoption of AI PCs to pick up in the second half of the year.

In January, AMD announced its next-generation Ryzen 8000G series desktop processors, which promise to deliver ‘immense power and dominant performance for intensive workloads including gaming and content creation’.

Speaking on the fourth quarter 2023 earnings call (30 January) AMD CEO Lisa Su said that ‘Ryzen CPUs (central processing units) power more than 90% of AI-enabled PCs currently in the market’. AMD’s stock has also enjoyed a strong run that stretches back to early last year and is up around 50% in 2024 to date. 

Elsewhere, shares in ARM soared close on 50% over a few days after a February announcement that the British firm anticipates capturing significant growth from red-hot demand for all things AI. ARM-based processor chips are used to train LLMs, including Nvidia’s GH200 Grace Hopper Superchip.

The firm’s performance and power-efficient CPU platform is used by more and more software developers, making it easier for OEMs (original equipment manufacturers) to adopt ARM technology, which generates further demand for ARM-based chips, the company explained.

The expectation is that growing demand for energy-efficient AI applications will drive ARM’s long-term growth prospects. In the short term, the company has raised its full-year revenue guidance from a range of $2.96 billion to $3.08 billion to between $3.16 billion to $3.21 billion.

Marvell Technology (MRVL:NASDAQ) is another aptly named chipmaker that has been elevated by AI. In Q3 2024, its data centre segment, which includes its custom AI chip business, beat revenue guidance, despite data centre storage demand remaining ‘depressed’ and industry hopes for a recovery being pushed out.

Data centre revenue grew 20% sequentially in the three months to the end of September, and the company is expecting it ‘to grow in the mid-30% range’ sequentially in Q4 2024, said president and CEO Matt Murphy in November 2023. The company’s Q4 2023 earnings report revealed largely flat growth year-on-year of $1.43 billion, yet the shares remain on the front foot this year, up 30%, as investors bet on a growth return in 2024.

ALTERNATIVE WAYS TO INVEST IN CHIPS

There are relatively inexpensive semiconductor themed ETFs for investors to consider. The £701 million iShares MSCI Global Semiconductor (SEMI) ETF has Broadcom, ASML and Nvidia as its largest three holdings, making up roughly 25% of the portfolio combined. Intel (INTC:NASDAQ) and Texas Instruments (TXN:NASDAQ) are perhaps less sexy but still important chip stocks in the portfolio.

Similarly, the VanEck Semiconductor (SMGB) has more than 11% of its £1.4 billion assets invested in Nvidia, it’s largest single stake, while also owning significant holdings of chip kit suppliers Lam Research and Applied Materials (AMAT:NASDAQ).

The smaller £292 million Amundi MSCI Semiconductor ESG Screened (SEMG) ETF has 75 sector holdings, but nearly 30% of assets in Nvidia, with another 9.7% in TSMC and 10.2% in Broadcom, making it a sector option more skewed to the success of a small number of companies.  

Finally, the HSBC Nasdaq Global Semiconductor (HNSS) ETF was only launched in January 2022, which helps explain its relatively small scale, with just £27 million of assets. That’s likely to change over time if investors select its’ 80-stock portfolio for chip exposure with a more evenly dispersed portfolio.

All of these ETF options are available on a 0.35% ongoing charge.

‹ Previous2024-03-14Next ›