$WLD According to the latest report from Reuters, OpenAI is accelerating the development of its own AI chips, aiming to reduce dependence on the market-leading Nvidia chips and thereby enhance its bargaining power in the supply chain. The company plans to complete the design of its first self-developed chip within this year and submit the design to TSMC for manufacturing, a process commonly referred to as 'tape out'.
Chip design and production progress
Design finalization and tape-out process:
Reports indicate that OpenAI will finalize the design of the first generation of chips in the coming months and then send the design documents to TSMC for production. Generally, the tape-out process can cost tens of millions of dollars and take about six months to transition from design to finished product production; if issues arise during the first tape-out, the process must be re-evaluated and repeated.
Mass production target:
OpenAI's long-term goal is to achieve mass production at TSMC by 2026, indicating significant progress in the company's in-house chip design and suggesting more self-controlled hardware support for large-scale AI model operations in the future.
Technical and strategic considerations
Reducing dependence on Nvidia:
Due to Nvidia's extremely high market share of GPUs and the potential supply bottlenecks and high costs associated with related chips, OpenAI hopes to mitigate supply risks by independently developing chips and further reduce reliance on a single supplier.
Team and partners:
This plan by OpenAI is led by former Google engineers, and the team has expanded to about 40 people in recent months, collaborating with Broadcom on chip design. This cross-company collaboration model not only facilitates technological exchange but also ensures advanced process support from TSMC during the production phase.
Application scenarios and initial deployment:
Although this chip is specifically designed for AI training and will initially be deployed on a limited scale to run AI models, if the trial production goes smoothly, OpenAI may test the feasibility of using it as an alternative to Nvidia chips later this year, laying the groundwork for large-scale production.
Future outlook and industry impact
This strategic initiative by OpenAI reflects the industry-wide trend towards self-manufactured chips. Like major players such as Google, Microsoft, and Amazon, designing hardware in-house not only allows optimization for specific AI workloads, enhancing computing speed and energy efficiency but also provides greater control and bargaining power amid supply chain tensions and rising costs.
If OpenAI can successfully complete this process and achieve mass production by 2026, it is expected to further reduce its operational costs and possibly drive more innovation in chip design and production within the entire AI industry, thereby having a profound impact on the overall ecosystem.
Conclusion
In summary, Reuters' report reveals that OpenAI is actively promoting the development of its own AI chip and plans to submit its design to TSMC for production within this year. This strategy not only helps reduce dependence on Nvidia but also enhances OpenAI's autonomy and competitiveness in AI infrastructure. With ongoing technological development and industry collaboration, we may see more major companies moving towards self-manufactured chips in the coming years, thereby advancing the entire AI industry.
