news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

he graduated from tsinghua university at the age of 22, and founded deephi technology at the age of 23, devoting himself to ai chip entrepreneurship. at the age of 25, the company was acquired for us$300 million. at the age of 29, he started a new business, crossed over to commercial aerospace, and co-founded the private rocket company oriental space. from ai chips to the stars and the sea, yao song, co-founder and co-ceo of oriental space, shared wonderfully at the cadencelive china 2024 china user conference, showing us that in today's rapid development of global technology, artificial intelligence (ai) is becoming the core force driving innovation in various industries. especially in the field of electronic design automation (eda), the integration of ai is constantly expanding the boundaries of design. from tiny chips to the vast universe, the innovation of eda is leading us to a brand new future.

yao song gave a wonderful speech titled "from eda to the universe: the simulation industry with an infinite future", describing his experience of moving from the artificial intelligence chip industry to the commercial aerospace industry, and especially introducing his thoughts on the future development of the technology industry when choosing to cross industries. commercial rockets are complex systems, in which simulation and intelligence play an important role, which is also one of the reasons why spacex has a huge gap with other commercial aerospace companies in the world.

yao song said that in the field of direct competition, all chinese commercial aerospace companies are more than 20 years behind spacex, founded by musk, and the speed of technological development is still lower than that of spacex. in terms of r&d efficiency, the spacex starship prototype can conduct an experiment about once a month and a half, which reflects that spacex's infrastructure is completely different from that in china, and its digital and information-based r&d system is far ahead. this is due to spacex's extensive r&d in the field of simulation algorithms and software, some of which have even subverted the original r&d process. yao song particularly pointed out the characteristics of commercial rockets as complex systems. he mentioned that simulation technology plays a vital role in the design and testing of these complex systems, not only helping engineers understand the behavior of the system, but also predicting and solving potential problems before the actual launch.

"with the technological evolution in integrated circuit technology, mobile internet and other fields, several core technology paradigms that humans have relied on for survival have gradually become invalid in the past few decades." yao song pointed out, "the first technology paradigm is moore's law in the field of integrated circuits. the chip manufacturing process has become slower and more expensive. the second technology paradigm is wireless communication (mobile internet). its cornerstone is shannon's formula. now it is difficult to significantly improve communication efficiency again. 6g is better than 5g not only in communication speed. the third technology paradigm is the entire internet. from ipv4 to ipv6, it has not been possible, and the internet business model has not made a bigger breakthrough."

based on his previous entrepreneurial experience in integrated circuit simulation software, hardware security, compilers, artificial intelligence-related architectures and algorithms, and his current experience in the aerospace field, yao song believes that with the advancement of technology, simulation will play a key role in more fields, especially in the commercial aerospace industry, where simulation will become an important tool to promote innovation and reduce costs. in particular, with the evolution of ai technology, there is great potential to improve the accuracy and efficiency of simulation. he believes that deep learning and large models may bring about a new generation of paradigms for future simulation technology, ushering in a new era characterized by intelligence and automation.

in fact, simulation is a key link in the eda industry, and the entire industry is at the forefront of technological innovation. trends such as ai-based ultra-large-scale computing, 5g, and autonomous driving are driving design needs in many vertical fields. this not only brings new market opportunities to the eda industry, but also promotes the in-depth application of ai technology in eda tools. from chip design to system integration, ai's intelligent algorithms and data analysis capabilities are helping the eda industry achieve more efficient design processes, more accurate simulation predictions, and more optimized solutions.

dr. jinqing teng, senior vice president and general manager of the digital and signoff division at cadence, mentioned that the integration of the semiconductor industry and the system industry has become an irreversible trend. this integration is not only reflected in the technical level, but also has a far-reaching impact on business models and market strategies. the emergence of ai, especially generative ai, has also brought huge growth space to the semiconductor industry. on the one hand, thanks to the continuous evolution of software and algorithms and the significant improvement in the performance of hardware such as gpus, ai large models continue to evolve, and in the future it is expected to imitate the magnificent scale of the human brain and reach the level of 100 trillion synaptic connections, which will greatly promote the demand for computing, data and energy infrastructure. on the other hand, the trend of model miniaturization has also quietly emerged, which aims to adapt to edge devices and promote vertical system integration and optimization. in the next 5 to 10 years, the core design of ai software and hardware will undergo amazing changes, and the ai ​​chip market will also usher in explosive growth.

in the ai-driven era, cadence has three directions of efforts: using ai chip infrastructure, using ai to update existing solutions, and exploring more new markets with partners. when using ai to help customers innovate, cadence's solutions have built a three-layer architecture. the bottom layer relies on jedai big data and ai platform to accelerate ai deployment; the middle layer provides diversified optimization ai solutions for digital, analog and simulation fields; the top layer is cadence copilot, which uses advanced technologies such as large language models (llm) to improve llm-based ai solutions.

teng jinqing emphasized that over the past three years, cadence has made significant progress in the field of optimization ai. for example, cadence cerebrus has been widely used in the field of digital design, greatly improving customer productivity. looking to the future, cadence will be committed to transforming the traditional single eda tool and operating environment into a multi-operation, multi-tool collaborative design environment, and combining ai to learn from design data, thereby reducing manual decision-making and promoting the automation of chip design.

the speeches of the two industry leaders jointly painted a picture of eda innovation enabled by ai. from yao song's in-depth analysis of commercial rocket simulation to teng jinqing's detailed introduction to the cadence ai platform, we can see that the innovation boundaries of eda are constantly expanding. ai is not only improving design efficiency and reducing costs, but also promoting eda technology to a higher level of intelligence.

with the continuous advancement of ai technology, the innovation boundary of eda seems to be becoming increasingly blurred. in many fields such as automobiles, data centers, and life sciences, the combination of ai and eda is opening up unprecedented design possibilities. cadence's co-pilot tools and digital twin technology, as well as oriental space's simulation applications in the commercial aerospace field, are vivid manifestations of this trend.

from chip design to space exploration, the combination of ai and eda seems to have no limit. as yao song said, the application of simulation technology in commercial rockets has allowed us to see the infinite possibilities from the earth to the stars. teng jinqing also mentioned that cadence's ai platform is shaping tomorrow's data center, heralding a more intelligent and automated future.

in this ai-enabled era, the innovative boundaries of eda may only exist in our imagination. as technology continues to develop, we have reason to believe that eda will continue to break through traditional boundaries and lead us to a broader universe.