Today in the global marketplace, we are witnessing an unprecedented paradigm shift. After OpenAI's ChatGPT caught the attention of consumers and investors, companies across industries are racing to integrate AI capabilities. Among the giants with a market capitalization of over $1 trillion in the U.S. market, Apple tops the list with a market capitalization of $3.08 trillion, followed by Microsoft ($2.51 trillion), Google's parent company, Alphabet ($1.67 trillion), Amazon ($1.35 trillion), and Nvidia ($1.15 trillion), with the exception of Apple's reliance on consumer devices such as iPhones, the other four tech giants are pushing full steam ahead to integrate with the AI space.

For example, Microsoft's recent announcement that it will begin charging monthly service fees for its enterprise AI software tools to business users is a signal that big companies are on the vanguard of successfully "cashing in" on the AI business opportunity with their clients. In addition, Alphabet is also integrating several products to introduce generative AI to help expand the potential market. With these tech giants investing large amounts of capital expenditures, AI is stirring the pot.

AI this east wind, also makes the chip supply chain in the enterprise benefit a lot, first of all, nvidia with GPU monopolize the entire generative AI chip market, SK Hynix and samsung and so on because of HBM and benefit, is responsible for the packaging and foundry TSMC is also in short supply, production capacity straight line emergency, sun moonlight / SPIL and other sealing and testing factory from the hands of TSMC to share the packaging outsourcing orders. There are also many AI chip players eyeing, even IBM is pushing its AIU chip that has been studied intensively for 5 years. Generative artificial intelligence "gold rush" is taking the lead in making a part of the "shovel seller" rich.

Do you think the dividends of AI have been eaten by them? In fact, in addition to these intuitive dividend payers, many device manufacturers and EDA/IP vendors have accidentally become indirect beneficiaries. If AI chip makers are selling shovels, then they can be called "shovel makers", they are also enjoying the opportunities brought by this change.

Equipment Vendors Unexpectedly Benefit

Most of the chips required in areas such as generative AI use advanced processes, and ASML, as the sole provider of EUV lithography, the equipment used to produce advanced process wafers, is surely one of the winners.2023 In the second quarter of the year, ASML achieved net sales of €6.9 billion, with a gross margin of 51.3% and a net profit of €1.9 billion. New orders in the second quarter of the year amounted to €4.5 billion, of which €1.6 billion were EUV lithography orders. What is more surprising is that ASML still has 38 billion euros of undelivered orders.ASML expects sales in 2023 will grow by 30%.

"Advanced AI servers have significantly higher leading-edge logic, memory and storage requirements compared to traditional servers, and every 1% increase in AI server and data center penetration is expected to drive $1 billion to $1.5 billion in additional (chip device) investment. With AI currently in its infancy, more investment in factory and corporate equipment will be critical in the coming years." Lam Research CEO Tim Archer said.

For the quarter ended June 25, 2023, Lam Research reported revenues of $3.21 billion and net income of $803 million, exceeding the range of earnings expectations. In terms of regional sales, China remains Lam Research's largest revenue generating region with 26%, South Korea with 24%, Taiwan with 20%, Japan with 10%, the U.S. and Europe both with 8%, and Southeast Asia with 4%.

For the remainder of 2023, Archer said he expects the market for chip manufacturing equipment to total about $70 billion. China's domestic demand for equipment purchases and high-speed storage tools is likely to drive the industry. Chinese companies have shifted purchases to equipment used for older logic and memory chips since October 2022, when the U.S. imposed export control restrictions.

Test equipment vendors are also on the beneficiary side, as many AI chips are required to utilize 2.5D stacking, 3D stacking, and Chiplet technology, which structurally increases the demand for chip test equipment in order to help manufacturers ensure performance and quality, and these chip tests require increasingly sophisticated test equipment to identify where manufacturing errors are occurring. Hideki Yasuda, an analyst at Toyo Securities, said, "Server chips are going to get bigger and more complex, requiring more time for testing. There is no magic way to shorten test time. The only solution for chipmakers is to buy more tools to test more chips at the same time. Global revenues from high-performance GPU chip test equipment could surpass smartphone chip test equipment in a few years."

Thanks to the growing demand for semiconductors in artificial intelligence technology, Q2 2023. U.S.-based chip test equipment giant Teradyne reported revenues of $684 million, including $475 million from its semiconductor test business, $94 million from its systems test business, $44 million from its wireless test business, and $72 million from its robotics business.

Greg Smith, Teradyne's chief executive officer, said, "Our revenues came in at the high end of our expected range, with increased semiconductor test shipments more than offsetting weak robotics demand during the quarter, and profits exceeded plan, primarily due to higher gross margins. Demand for DRR5 and HBM memory device testing for data center applications remained strong as we entered the third quarter, and SOC test demand for automotive applications is gradually strengthening. In robotics, we expect order rates to decline as customers respond to slowing global industrial activity and macroeconomic pressures."

Mihashi, co-chief strategy officer of Advantest, another Japanese test equipment major, recently said in an interview, "We are a dominant player in the industry, so we will benefit when ChatGPT and others expand the use of high-performance computing." They also believe that AI demand is helping the recovery of chip test equipment.

Semiconductor back-end equipment vendors are also enjoying big dividends, with demand for chips behind generative AI outstripping supply, forcing TSMC to repeatedly ramp up its CoWos capacity and even spend $90 billion on a new CoWos advanced packaging and testing plant in Taiwan. As a result, equipment vendors are being pulled up. To meet the growing demand for CoWoS packaging, TSMC is working with a number of global suppliers, including Rudolph Technologies in the US, Disco in Japan and SUSS MicroTec in Germany, as well as Taiwanese specialists Grand Process Technology (GPTC) and Scientech. According to DigiTimes, these suppliers have been asked to provide nearly 30 kits by mid-2024.

EDA/IP vendors enjoy "double benefits"

In the past, when the industry has been in a down-cycle phase, the EDA and IP markets have typically fallen before the overall market slowdown, but recovered faster than the market. This did not happen in this down cycle, and EDA vendors showed strength throughout the epidemic and after the recovery.

Looking specifically at the EDA vendors' earnings, Synopsys reported revenue of $1.395 billion for the second fiscal quarter of fiscal year 2023 ended April 30, 2023, compared to $1.279 billion for the same period a year earlier, an increase of 9.07%, and net income of $273 million. For the third fiscal quarter, Synopsys forecasts revenue to be between $1.465 billion and $1.495 billion, roughly better than market expectations. Meanwhile, Synopsys raised its full-year FY2023 earnings guidance, projecting revenue of $5.790-5.830 billion.

Cadence delivered excellent results in the second quarter of 2023, realizing revenue of $977 million in its second-quarter earnings report for the period ended June 30, compared to revenue of $858 million and net income of $221 million for the same period in 2022. said Anirudh Devgan, president and chief executive officer of Cadence: "With its unrivaled promise, generative AI is beginning to make a significant impact globally. Our focus on AI over the past few years, combined with our computational software expertise and the valuable data at the heart of AI, puts us in a unique position to realize the enormous potential of this transformative technology." Cadence has also raised its full-year revenue forecast to slightly above Wall Street expectations, with Cadence expecting full-year revenue to be in the range of $4.05 billion to $4.09 billion, an increase of 14 percent year-over-year from last year.

And when it comes to the impact of generative AI development on EDA vendors, unlike vendors that only sell equipment and chips, EDA vendors have at least two ways to benefit from generative AI: on the one hand, they can provide EDA tools for AI chip design; on the other hand, they can also take advantage of generative AI and add it to their own software to further help complete chip design.

As more and more system vendors such as Google, Meta, and Alibaba develop their own AI chips, they are part of the army of EDA buyers. walden C. Rhines, executive sponsor of SEMI's Electronic Design Market Data Report, said, "The electronic design automation (EDA) industry continues to see double-digit growth in Q1 2023 The electronic design automation (EDA) industry continued to see double-digit growth in the first quarter of 2023, with growth across all product categories and geographies. These product categories include computer-aided engineering, IC physical design and verification, printed circuit boards and multi-chip modules, and services all showed double-digit growth."

The use of AI in EDA software is not new, with the EDA triumvirate of Synopsys, Cadence, and Siemens all now launching their own AI tools. Existing AI tools are already providing chipmakers with significant improvements in productivity and speed in the present day, gradually showing advantages. Therefore, the development of generative AI will be even more of an icing on the cake for EDA vendors in the long run.

In April 2023, Siemens and Microsoft officially announced that the two are working together to use generative AI in the full lifecycle of industrial product design, engineering, manufacturing, and operations to improve innovation and efficiency. The two companies are integrating Siemens' product lifecycle management software, Teamcenter, with Microsoft's collaboration platform, Teams, language models in Azure OpenAI services, and other Azure AI capabilities.

Synopsys engineers are exploring how cutting-edge large-scale language models (LLMs) like those used by ChatGPT can help streamline internal processes and enhance existing solutions.

Generative AI can help build learning datasets, said KT Moore, vice president of corporate marketing at Cadence, during a workshop. In turn, these datasets can be used to create other future designs.

However, while generative AI does have excellent results in language and images, its development is still in its infancy and there are still flaws in using it exclusively for chip design. Actual chip designs need to be accurate enough (9 out of 9) that even the tiniest mistake could have huge consequences in terms of efficiency, yield, time to market, etc.

Today in the global marketplace, we are witnessing an unprecedented paradigm shift. After OpenAI's ChatGPT caught the attention of consumers and investors, companies across industries are racing to integrate AI capabilities. Among the giants with a market capitalization of over $1 trillion in the U.S. market, Apple tops the list with a market capitalization of $3.08 trillion, followed by Microsoft ($2.51 trillion), Google's parent company, Alphabet ($1.67 trillion), Amazon ($1.35 trillion), and Nvidia ($1.15 trillion), with the exception of Apple's reliance on consumer devices such as iPhones, the other four tech giants are pushing full steam ahead to integrate with the AI space.

For example, Microsoft's recent announcement that it will begin charging monthly service fees for its enterprise AI software tools to business users is a signal that big companies are on the vanguard of successfully "cashing in" on the AI business opportunity with their clients. In addition, Alphabet is also integrating several products to introduce generative AI to help expand the potential market. With these tech giants investing large amounts of capital expenditures, AI is stirring the pot.

AI this east wind, also makes the chip supply chain in the enterprise benefit a lot, first of all, nvidia with GPU monopolize the entire generative AI chip market, SK Hynix and samsung and so on because of HBM and benefit, is responsible for the packaging and foundry TSMC is also in short supply, production capacity straight line emergency, sun moonlight / SPIL and other sealing and testing factory from the hands of TSMC to share the packaging outsourcing orders. There are also many AI chip players eyeing, even IBM is pushing its AIU chip that has been studied intensively for 5 years. Generative artificial intelligence "gold rush" is taking the lead in making a part of the "shovel seller" rich.

Do you think the dividends of AI have been eaten by them? In fact, in addition to these intuitive dividend payers, many device manufacturers and EDA/IP vendors have accidentally become indirect beneficiaries. If AI chip makers are selling shovels, then they can be called "shovel makers", they are also enjoying the opportunities brought by this change.

Equipment Vendors Unexpectedly Benefit

Most of the chips required in areas such as generative AI use advanced processes, and ASML, as the sole provider of EUV lithography, the equipment used to produce advanced process wafers, is surely one of the winners.2023 In the second quarter of the year, ASML achieved net sales of €6.9 billion, with a gross margin of 51.3% and a net profit of €1.9 billion. New orders in the second quarter of the year amounted to €4.5 billion, of which €1.6 billion were EUV lithography orders. What is more surprising is that ASML still has 38 billion euros of undelivered orders.ASML expects sales in 2023 will grow by 30%.

"Advanced AI servers have significantly higher leading-edge logic, memory and storage requirements compared to traditional servers, and every 1% increase in AI server and data center penetration is expected to drive $1 billion to $1.5 billion in additional (chip device) investment. With AI currently in its infancy, more investment in factory and corporate equipment will be critical in the coming years." Lam Research CEO Tim Archer said.

For the quarter ended June 25, 2023, Lam Research reported revenues of $3.21 billion and net income of $803 million, exceeding the range of earnings expectations. In terms of regional sales, China remains Lam Research's largest revenue generating region with 26%, South Korea with 24%, Taiwan with 20%, Japan with 10%, the U.S. and Europe both with 8%, and Southeast Asia with 4%.

For the remainder of 2023, Archer said he expects the market for chip manufacturing equipment to total about $70 billion. China's domestic demand for equipment purchases and high-speed storage tools is likely to drive the industry. Chinese companies have shifted purchases to equipment used for older logic and memory chips since October 2022, when the U.S. imposed export control restrictions.

Test equipment vendors are also on the beneficiary side, as many AI chips are required to utilize 2.5D stacking, 3D stacking, and Chiplet technology, which structurally increases the demand for chip test equipment in order to help manufacturers ensure performance and quality, and these chip tests require increasingly sophisticated test equipment to identify where manufacturing errors are occurring. Hideki Yasuda, an analyst at Toyo Securities, said, "Server chips are going to get bigger and more complex, requiring more time for testing. There is no magic way to shorten test time. The only solution for chipmakers is to buy more tools to test more chips at the same time. Global revenues from high-performance GPU chip test equipment could surpass smartphone chip test equipment in a few years."

Thanks to the growing demand for semiconductors in artificial intelligence technology, Q2 2023. U.S.-based chip test equipment giant Teradyne reported revenues of $684 million, including $475 million from its semiconductor test business, $94 million from its systems test business, $44 million from its wireless test business, and $72 million from its robotics business.

Greg Smith, Teradyne's chief executive officer, said, "Our revenues came in at the high end of our expected range, with increased semiconductor test shipments more than offsetting weak robotics demand during the quarter, and profits exceeded plan, primarily due to higher gross margins. Demand for DRR5 and HBM memory device testing for data center applications remained strong as we entered the third quarter, and SOC test demand for automotive applications is gradually strengthening. In robotics, we expect order rates to decline as customers respond to slowing global industrial activity and macroeconomic pressures."

Mihashi, co-chief strategy officer of Advantest, another Japanese test equipment major, recently said in an interview, "We are a dominant player in the industry, so we will benefit when ChatGPT and others expand the use of high-performance computing." They also believe that AI demand is helping the recovery of chip test equipment.

Semiconductor back-end equipment vendors are also enjoying big dividends, with demand for chips behind generative AI outstripping supply, forcing TSMC to repeatedly ramp up its CoWos capacity and even spend $90 billion on a new CoWos advanced packaging and testing plant in Taiwan. As a result, equipment vendors are being pulled up. To meet the growing demand for CoWoS packaging, TSMC is working with a number of global suppliers, including Rudolph Technologies in the US, Disco in Japan and SUSS MicroTec in Germany, as well as Taiwanese specialists Grand Process Technology (GPTC) and Scientech. According to DigiTimes, these suppliers have been asked to provide nearly 30 kits by mid-2024.

EDA/IP vendors enjoy "double benefits"

In the past, when the industry has been in a down-cycle phase, the EDA and IP markets have typically fallen before the overall market slowdown, but recovered faster than the market. This did not happen in this down cycle, and EDA vendors showed strength throughout the epidemic and after the recovery.

Looking specifically at the EDA vendors' earnings, Synopsys reported revenue of $1.395 billion for the second fiscal quarter of fiscal year 2023 ended April 30, 2023, compared to $1.279 billion for the same period a year earlier, an increase of 9.07%, and net income of $273 million. For the third fiscal quarter, Synopsys forecasts revenue to be between $1.465 billion and $1.495 billion, roughly better than market expectations. Meanwhile, Synopsys raised its full-year FY2023 earnings guidance, projecting revenue of $5.790-5.830 billion.

Cadence delivered excellent results in the second quarter of 2023, realizing revenue of $977 million in its second-quarter earnings report for the period ended June 30, compared to revenue of $858 million and net income of $221 million for the same period in 2022. said Anirudh Devgan, president and chief executive officer of Cadence: "With its unrivaled promise, generative AI is beginning to make a significant impact globally. Our focus on AI over the past few years, combined with our computational software expertise and the valuable data at the heart of AI, puts us in a unique position to realize the enormous potential of this transformative technology." Cadence has also raised its full-year revenue forecast to slightly above Wall Street expectations, with Cadence expecting full-year revenue to be in the range of $4.05 billion to $4.09 billion, an increase of 14 percent year-over-year from last year.

And when it comes to the impact of generative AI development on EDA vendors, unlike vendors that only sell equipment and chips, EDA vendors have at least two ways to benefit from generative AI: on the one hand, they can provide EDA tools for AI chip design; on the other hand, they can also take advantage of generative AI and add it to their own software to further help complete chip design.

As more and more system vendors such as Google, Meta, and Alibaba develop their own AI chips, they are part of the army of EDA buyers. walden C. Rhines, executive sponsor of SEMI's Electronic Design Market Data Report, said, "The electronic design automation (EDA) industry continues to see double-digit growth in Q1 2023 The electronic design automation (EDA) industry continued to see double-digit growth in the first quarter of 2023, with growth across all product categories and geographies. These product categories include computer-aided engineering, IC physical design and verification, printed circuit boards and multi-chip modules, and services all showed double-digit growth."

The use of AI in EDA software is not new, with the EDA triumvirate of Synopsys, Cadence, and Siemens all now launching their own AI tools. Existing AI tools are already providing chipmakers with significant improvements in productivity and speed in the present day, gradually showing advantages. Therefore, the development of generative AI will be even more of an icing on the cake for EDA vendors in the long run.

In April 2023, Siemens and Microsoft officially announced that the two are working together to use generative AI in the full lifecycle of industrial product design, engineering, manufacturing, and operations to improve innovation and efficiency. The two companies are integrating Siemens' product lifecycle management software, Teamcenter, with Microsoft's collaboration platform, Teams, language models in Azure OpenAI services, and other Azure AI capabilities.

Synopsys engineers are exploring how cutting-edge large-scale language models (LLMs) like those used by ChatGPT can help streamline internal processes and enhance existing solutions.

Generative AI can help build learning datasets, said KT Moore, vice president of corporate marketing at Cadence, during a workshop. In turn, these datasets can be used to create other future designs.

However, while generative AI does have excellent results in language and images, its development is still in its infancy and there are still flaws in using it exclusively for chip design. Actual chip designs need to be accurate enough (9 out of 9) that even the tiniest mistake could have huge consequences in terms of efficiency, yield, time to market, etc.