NVIDIA earlier announced a partnership with electronic design automation (EDA) provider Synopsys. To expand strategic cooperationFurthermore, NVIDIA directly invested $20 billion in the company. This deal, combined with NVIDIA's previous strategic layout with Intel in the fields of advanced packaging and foundry, highlights that NVIDIA CEO Jensen Huang's ambitions have long surpassed simply "selling GPU chips."
By binding together capital and technology, NVIDIA is attempting to leverage the CUDA ecosystem to bring the entire global semiconductor industry, from upstream design to downstream manufacturing, into its sphere of influence of "accelerated computing." This will be an unprecedented squeeze on the ecosystem for competitors like AMD.
Strategy 1: Investing in Synopsys to make chip manufacturing inseparable from the CUDA software ecosystem.
First, the core significance of NVIDIA's investment in Synopsys lies in establishing CUDA as the "standard language" for industrial design.
As chip manufacturing processes enter the ampere generation, traditional CPU computing is struggling to handle the massive simulation demands. NVIDIA, by pushing Synopsys' entire EDA tool suite to support NVIDIA CUDA-X and AI physics acceleration, is creating a development landscape that will stifle its competitors.
• Raising the design threshold:In the future, if chip design companies (including AMD, Intel, and even Google) want to use the most efficient Synopsys EDA tools (such as the AI-driven AgentEngineer), their underlying computing power must rely more heavily on NVIDIA GPUs.
• CUDA's moat extends:CUDA has almost "dominated" AI training and inference in the past, and now it has extended even further to "chip design itself". This means that even if competitors want to design a chip that "can beat NVIDIA", they will probably still have to purchase and expand NVIDIA's GPUs to run simulation tests during the design process.
Strategy Two: Strategically align with Intel to diversify risks and utilize competitor's production capacity.
In addition to partnering with Synopsys on the upstream design side, NVIDIA previously relied on Intel...Investment LayoutThis is also worth noting. Although the two companies are still competitors in the AI chip market, NVIDIA has shown a high degree of "flexibility" in manufacturing.
NVIDIA has previously confirmed its collaboration with Intel in the field of advanced packaging, and has even not ruled out using Intel Foundry for chip manufacturing in the future.
This move has two strategic implications:
• Supply Chain Resilience:With TSMC's CoWoS capacity tight, NVIDIA has secured a second supply chain for itself by "investing" in Intel's manufacturing ecosystem (whether through actual orders or technical cooperation).
• Suppressing AMD:Intel urgently needs foundry orders to maintain cash flow. If NVIDIA becomes a major customer of Intel Foundry, it will, to some extent, tie Intel's interests to NVIDIA's in the wafer manufacturing sector. This will make AMD, which also relies on advanced processes, face a more complex competitive and cooperative relationship in the competition for production capacity.
Conclusion: From "selling shovels" to "defining shovels," AMD faces the crisis of marginalization.
In summary, NVIDIA's strategy has shifted from simply selling AI hardware to defining the industry standards for the AI era.
• Software side:By acquiring a stake in Synopsys, the company ensures that its EDA tools are optimized primarily for NVIDIA GPUs, thereby building a technological barrier.
• Hardware side:By partnering with Intel for manufacturing, we ensure additional production capacity.
This strategy will undoubtedly put greater strategic pressure on AMD. Although AMD is currently pushing the ROCm open platform, in the highly specialized and closed software field of EDA, if existing partners such as Synopsys are gradually being penetrated by the CUDA ecosystem at the underlying level, it will become more difficult for ROCm to gain a foothold.
Jensen Huang's plan is very clear: in the AI gold rush, NVIDIA not only wants to be the one "selling shovels", but also wants to redefine "how to make shovels" through the combination of CUDA and Synopsys, so that all those who want to make gold, and even competitors who want to make shovels, will have to pay "tolls" to NVIDIA.
Whether this investment will attract the attention of regulatory agencies in the United States and the European Union, and lead to investigations into whether the investment cooperation involves market monopoly, remains to be seen.
Digression:When NVIDIA previously invested $1000 billion in OpenAI, OpenAI subsequently announced a multi-billion dollar GPU purchase agreement with AMD. Jensen Huang later stated in an interview that "this was a very smart move." Could this investment and cooperation agreement between NVIDIA and Synopsys be seen as NVIDIA trying to regain some ground with AMD?
