Introduction

When most investors/potential investors think of AI hardware companies they often lump their interest into just a few major players such as Nvidia, Taiwan Semiconductor, or Advanced Micro Devices, for example, and perhaps a few others, but the AI hardware market is much more extensive than that. Rather, it's a complex ecosystem comprising 8 distinct sub-segments, each with its own unique dynamics and growth trajectories which MunAiMarkets tracks in 8 model portfolios as to their constituent stock performances, the reasons behind any changes, and analyses of their current situation and future prospects. 

What Is A Semiconductor?

According to an edited and abridged version from Wikipedia, a semiconductor (aka computer chip or wafer) is a material that has an electrical conductivity value falling between that of a conductor, such as copper, and an insulator, such as glass. In many cases their conducting properties are altered by introducing impurities (doping) into the crystal structure and when two differently doped regions exist in the same crystal, a semiconductor junction is created. Connecting the two materials causes creation of a depletion region near the boundary, as the free electrons fill the available holes, which, in turn, allows electric current to pass through the junction only in one direction. and is the basis of diodes, transistors, and most modern electronics. Silicon is a critical element for fabricating most  electronic circuits and is the most common semiconductor. After silicon, gallium arsenide is the second-most common semiconductor and is used in laser diodes, solar cells, microwave-frequency integrated circuits, and others.

The Global AI Hardware Market Size, Growth Projections, and Key Segments

The global Artificial Intelligence (AI) hardware market size was valued at approximately $25.1 billion in 2024, with projections indicating a surge to $76.7 billion by 2030 at a compound annual growth rate (CAGR) of 20.5%. Separately, the global AI computer market, estimated at $51.10 billion in 2024, is expected to grow at a CAGR of 34.4% from 2025 to 2030, reaching $281.67 billion by 2030. Within the broader AI accelerator market, a robust CAGR of 28.20% is projected, with the market expanding from $20.01 billion in 2024 to $240 billion by 2034.

What Is the Semiconductor Value Chain?

The semiconductor value chain refers to the 4 stages involved in the production and distribution of semiconductor devices such as microprocessors, memory chips, and sensors which are used in a wide range of electronic devices, including smartphones, computers, and automobiles. The semiconductor value chain is a complex and dynamic process that involves multiple companies, each with their own specialized roles. (Source)

The 4 Semiconductor Value Chain Model Portfolios

#1: The Pure-Play EDA Software Chip Design Model Portfolio: UP 0.6% in May; DOWN 4.4% YTD 

• Focus: its constituents are involved exclusively (i.e. pure-play) in the custom design of complex semiconductor chips using sophisticated electronic design automation (EDA) software.
• Constituents: its 2 constituents are as follows: Cadence Design Systems (CDNS); and Synopsys (SNPS).

#2: The Pure-Play Semiconductor Supplier Model Portfolio: UP 7.6% in May; UP 8.6% YTD  

• Focus: its constituents provide essential equipment for manufacturing chips such as lithography machines, etching, test and packaging equipment, automation and inspection systems and production materials such as chemicals, gases, wafers and packaging.
• Constituents: its 5 constituents are as follows: Lam Research (LRCX); ASML Holding (ASML);  KLAC Corp. (KLAC); Applied Materials (AMAT); and Entegris (ENTG).

#3: The Pure-Play Semiconductor Foundries Model Portfolio: UP 12.9% in May; DOWN 7.4% YTD 

• Focus: it consist of factories (called foundries in semiconductor jargon) that concentrate all their efforts in the manufacture of chips based on the designs provided by other semiconductor companies.
• Constituents: its 4 constituents are as follows: Taiwan Semiconductor (TSM); Tower Semiconductor (TSEM); United Microelectronics (UMC); and GlobalFoundries (GFS).

#4: The Pure-Play Outsourced Semiconductor Assembly and Test Model Portfolio: DOWN 7.7% in May; DOWN 23.4% YTD 

• Focus: it consists of companies that assemble chips into finished semiconductor components, tests for defects and do the very specialized packaging of the chips for shipping.
• Constituents: its 2 constituents are as follows: ASE Technology (ASX); and Amkor Technology (AMKR).

#5: The MunAiMarkets Pure-Play Fabless Semiconductor Companies Model Portfolio: UP 13.1% in May; UP 0.1% YTD

• Focus: it consists of manufacturers who design and sell hardware devices and semiconductor chips themselves while outsourcing their fabrication (i.e. are fabless) to a specialized manufacturer.
• Constituents: its 6 constituents are as follows: Nvidia (NVDA); Broadcom (AVGO); Advanced Micro Devices (AMD); Monolithic Power (MPWR); Marvell (MRVL); and Qualcomm (QCOM).

#6: The MunAiMarkets Integrated Device Manufacturer (IDM) Model Portfolio: UP 11.6% in May; DOWN 0.3% YTD

• Focus: it consists of companies that do everything themselves in house to control the entire production process.
• Constituents: its 8 are as follows: Microchip Technology (MCHP); Micron Technology (MU); Infineon Technologies (IFNNY); Texas Instruments (TXN); Analog Devices (ADI); STMicroelectronics (STM); NXP Semiconductors (NXPI); and Intel Corp. (INTC).

#7: The MunAiMarkets High Band Memory (HBM) Chips Model Portfolio: UP 28.1% in May; UP 12.2% YTD

• Focus: High Band Memory (HBM), in simple terms, is a certain number of conventional DRAM (direct random-access memory) chips stacked together to achieve greater performance and efficiency. 
• Constituents: Only one (1) HBM company trades on a North American stock exchange, namely Micron Technology (MU), and, as such, is the only constituent in the model portfolio.

#8: The MunAiMarkets Pure-Play Quantum Computing Model Portfolio: UP 60.3% in May; DOWN 2.3% YTD

• Focus: it consists of companies  involved exclusively in the research and development of quantum computers.
• Constituents: its 4 constituents consist of: D-Wave Computing (QBTS); Quantum Computing Inc. (QUBT); IonQ (IONQ); and Rigetti Computing (RGTI).

Summary

On average, the 31 constituents in the 8 Model AI Hardware Stock Portfolios were UP 9.6% in May; UP 1.9% YTD. The iShares Semiconductor ETF (SOXX), with 35 constituents, in comparison, was UP 15.7% in May and is DOWN 4.9% YTD.

Please note that hyperscalers such as Google's TPU, Amazon's Inferentia and Trainium, Meta's MTIA, and Microsoft's Maia accelerators, are developing and deploying their own proprietary AI chips. This vertical integration strategy aims to optimize costs, enhance performance, and reduce reliance on third-party suppliers, signaling a potential fragmentation of the market and a shift towards more integrated, ecosystem-specific solutions. These hyperscalers are not presented in a portfolio of their own as their operations are not exclusively AI hardware operations but their major contributions to the research, development and deployment of AI chips are hereby acknowledged.

More By This Author:

Is Canadian "Elbows UP!" Boycott Of U.S. Grocery And Alcohol Products Hurting Company Stock Prices?
American Cannabis MSOs Rode Wave Of Wishful Thinking In April: Up 20%, On Average
Ai Hardware And Generative Ai Model Stock Portfolios Were Up 4%, On Average, In April