SoftBank, Sam Altman and Nvidia are among those now seeking to enter new parts of the semiconductor fabrication market. Both SoftBank and OpenAI are seeking to create new artificial intelligence production facilities, while Nvidia, facing new competition from its own customers, wants to branch out into custom chip production and even “GPU as a service” roles within the computing ecosystem.
Perhaps the size of the future market indicates why much of that activity is happening.
Research Firm | Current Market Value (2023) | CAGR (2023-2032) | Estimated Market Value (2032) |
Allied Market Research | $15 billion | 38.4% | $384 billion |
Precedence Research** | $454.12 billion | 19% | $2,575.16 billion |
Grand View Research | N/A | 38.1% | $1.81 trillion |
MarketsandMarkets | N/A | 42.5% | $829.4 billion |
Mordor Intelligence | N/A | 36.4% (to 2027) | $128 billion |
** Precedence Research estimates for AI chips in 2023 is a bit less than $22 billion. The figure shown in the chart above obviously includes semiconductors of all types. The AI chip growth rate is estimated by Precedence as just a bit under 30 percent annually.
Profit margins can vary, however.
Company | Revenue (USD Billion) | Net Income (USD Billion) | Profit Margin (%) |
TSMC | 56.8 | 23.3 | 41.0% |
Samsung Electronics | 208.9 | 39.4 | 18.9% |
Intel | 70.8 | 8.0 | 11.3% |
Micron Technology | 29.1 | 8.2 | 28.2% |
NVIDIA | 29.0 | 12.7 | 43.8% |
AMD | 16.4 | 3.2 | 19.5% |
Qualcomm | 37.5 | 7.1 | 18.9% |
Still, there might be considerations beyond revenue and profit potential.
The increasing adoption of artificial intelligence is driving the demand for specialized AI chips designed for efficient processing of complex algorithms. And customers already have faced shortages of key products such as graphics processing units.
Existing chip manufacturers might not be able to fully meet this demand due to their broader product portfolios and priorities.
Owning and operating chip fabrication facilities could give companies like OpenAI greater control over the hardware foundation of its AI development, allowing it to optimize chip design for its specific AI models and algorithms or reduce reliance on external chip suppliers and their pricing models.
Strategic value might also be seen. For companies like SoftBank, investing in AI chip fabrication could be part of a broader strategy to build a comprehensive AI ecosystem.
For an AI chip leader such as Nvidia, moving into adjacent areas, such as custom chip fabrication or “GPU as a service” represents a diversification move as others start to build their own internal GPU and other capabilities.
