Off-the-Shelf Chiplets Open New Market Opportunities
By Ravi Thummarukudy, Mobiveil (September 20, 2023)
With the reintroduction of semiconductor integration into system development, standards-based chiplets promise more efficient, cost-effective designs.
Twenty years ago, I watched a fledgling cottage industry spring up in which enterprising engineers built blocks of intellectual property to sell to design groups that were developing semiconductor chips. Many of those small enterprises are now big and formidable public companies. They serve a $4.58 billion market, according to Fortune Business Insights, that touches almost all of the semiconductor industry.
With chip design now more expensive than ever, sharing IP and information across companies has become a well-accepted practice. Only the largest, best-funded system companies, such as large mobile phone companies or hyperscalers, are able to build full-custom chip solutions. Differentiation moves these system companies to build their own chips, which become integral elements of a company’s success, instead of off-the-shelf, third-party chips that any company can purchase.
To read the full article, click here
Related Semiconductor IP
- AXI Interconnect
- 50G PON LDPC Encoder/Decoder
- Universal Multi-port Memory Controller for RLDRAM2/3, DDR4/3, DDR4 3DS and LPDDR3/2 and LPDDR3/2
- Universal NVM Express Controller (UNEX)
- ONFI/Toggle IP
Related Articles
- System-on-chip market to hit 1.3 billion units in 2004, says new report
- From IP re-use to open Innovation - A new trend in the industry
- 3-D IC design: New possibilities for the wireless market
- What's new in the Semiconductor IP market? The case of civilian Aerospace industry
Latest Articles
- LLM Inference on IMC-NoC Architecture with Balanced Dataflow and Fine-Grained Parallelism
- AI-Assisted Design of a Post-Quantum Cryptographic Accelerator: A Deployed-Silicon Case Study
- Hardware Acceleration of Block-Diffusion LLM for Edge Devices
- Scalable AXI4 Transaction Monitoring for Mixed-Criticality SoCs: From Phase-Level Precision to ID-Level Efficiency
- Hardware-managed heterogeneous high-bandwidth memory and flash in LLM inference systems