At the MOSIS 2.0 Developer Workshop in Arlington, VA, attendees gathered to explore how MOSIS 2.0 is shaping the future of semiconductor prototyping. The event took place on September 12, 2025, and showcased the platform’s ability to provide access to multi-project wafers (MPWs), a network of research fabs, and expert support for design teams, making it an invaluable gateway for innovators in the field.

Rehan Kapadia, director of MOSIS 2.0, opened his talk with a nod to the past. The original MOSIS, founded in 1981, was revolutionary for its time. By pioneering multi-project wafer (MPW) services, it gave designers access to silicon fabrication without requiring them to master the intricacies of foundry processes. This disaggregation of design and fabrication lowered barriers and democratized chip innovation. Designers could submit their work by email, which was remarkably advanced in a pre-internet era, and receive working circuits back in six to eight weeks. The service was affordable, automated, and often cost only a few hundreds of dollars. That model directly enabled some of the most significant architectural innovations of the 1980s and 1990s, from Berkeley’s first RISC processor to Stanford’s MIPS CPUs and the geometry engines that powered Silicon Graphics. MOSIS was reducing risk, cost, and turnaround time dramatically, expanding who can participate in innovation.

Accessible paths from concept to hardware

It is this ethos that MOSIS 2.0 seeks to revive. Kapadia emphasized that the goal is not to replicate the original service menu, but to reimagine how MOSIS 2.0 can reduce barriers for today’s challenges. The world of microelectronics is far more complex than it was in 1981, with multiple materials systems, heterogeneous integration, and system-level demands that extend well beyond silicon logic. For Kapadia, the mission is clear: MOSIS 2.0 must give designers faster, cheaper, and more accessible paths from concept to hardware, while expanding into new domains like photonics, advanced packaging, and heterogeneous prototyping.

MOSIS 2.0 operates within the Department of War’s Microelectronics Commons framework, specifically the California DREAMS Hub. This ecosystem combines a powerful mix of academic, government, and industry partners. On the academic side, seven universities—including USC, Caltech, and five UC campuses—contribute research facilities and nanofabs. Industry partners bring both fabrication and systems capabilities: Northrop Grumman, HRL, Teledyne bring their medium-volume fab capabilities to the hub, Raytheon, Boeing, Lockheed Martin, and Northrop Grumman represent to our system developers. PDF Solutions adds a layer of analytics expertise, while workforce development organizations support training and growth of talent pipelines.

PIES: Tailored Guidance Across the Fabrication Network

The complexity of this network requires coordination, and Kapadia positioned the PIES team (Prototype Integration and Engineering Team) as the linchpin. The PIES team acts as the interface between users and partners, ensuring that customers can navigate a diverse landscape of fabs, tools, and processes. Process engineers handle nanofabs and low-volume fabs, while circuit and layout aggregation engineers manage engagement with mid- and high-volume foundries. Engineers are also embedded directly within partner organizations, so MOSIS 2.0 can tap into distributed expertise without duplicating staff. The model is flexible: some customers arrive with fully finished designs and simply need access to a shuttle run. Others come with a concept and need advice on whether to pursue silicon, III-V, or another path altogether. In those cases, the PIES team plays a consultative role, evaluating options and guiding decisions.

Kapadia outlined the current MOSIS 2.0 service structure. Customers can access MPWs for both silicon and compound semiconductors, pursue custom prototypes through split-flows across nanofabs, and use a secure online portal for PDKs, GDS files, and verification checks. Packaging capabilities are being added so customers can receive packaged devices ready for testing and integration. Importantly, MOSIS 2.0 benefits from the USC Information Sciences Institute ecosystem. Beyond its own staff, it can draw on professional engineers and scientists across ISI for customer projects as needed with high granularity on a contract basis. This model provides flexibility and scale without the financial burden of a large permanent design team. Kapadia noted that even faculty and students may contribute on specific projects, allowing MOSIS 2.0 to mobilize specialized expertise.

The talk then turned to how MOSIS 2.0 differentiates itself in today’s global landscape. For Kapadia, simply aggregating access to fabrication services is not enough. What will set MOSIS 2.0 apart is its focus on system-level prototyping and innovation. He described efforts to develop semi-customizable interposers, rapid prototyping systems, and frameworks that combine ASICs with FPGAs to accelerate design iteration. By allowing designers to focus on novel IP while offloading routine functions to pre-validated modules or FPGA integration, MOSIS 2.0 reduces barriers not just at the chip level, but at the system level.

Reproducibility Should be the Norm

Automation is another priority. Kapadia highlighted collaborations with faculty at USC and UC Irvine on spec-to-GDS algorithms, which aim to reduce the time and expertise needed to move from high-level design specifications to manufacturable layouts. This work, still in early stages, represents the kind of innovation that MOSIS 2.0 hopes to capture and scale: connecting academic research with industry needs in practical, customer-ready form.

Research fab services represent a particularly challenging but critical area of growth. Kapadia, who also directs USC’s nanofab, spoke candidly about the realities of process development in research fabs, where reproducibility is not the norm and multiple attempts are often required before success. MOSIS 2.0 aims to “professionalize” this environment by building infrastructure that makes process flows portable and predictable across multiple facilities. He pointed to partnerships with PDF Solutions, which bring data analytics into nanofab operations. By analyzing past runs, applying predictive maintenance, and using machine learning to optimize process conditions, MOSIS 2.0 reduces the trial-and-error burden. Kapadia cited a recent example where an etch process developed at UCSB was transferred to USC’s nanofab with only a single trial and a 5% variance in performance between the fabs, a milestone that shows how data can make multiple independent facilities operate more like a single, cohesive network.

Focus on Reliability and Lowering Barriers

The roadmap for MOSIS 2.0 is ambitious. In the near term, the team is focused on reinforcing MPW services, expanding design and IP support, and completing packaging offerings. In the medium term, MOSIS 2.0 will work to standardize rapid prototyping systems and expand into photonics, building on early partnerships with Sandia National Labs and others. Over the long term, Kapadia envisions a networked, data-driven ecosystem where disparate fabs are connected not just by shared customers, but by interoperable processes, analytics, and design frameworks. This would allow innovators to move seamlessly from early ideas through system-level prototypes and, eventually, into scalable manufacturing.

Kapadia closed on a pragmatic note. The success of MOSIS 2.0 will depend less on vision statements and more on execution and on delivering consistent, professional experiences for both designers and foundry partners.

MOSIS 2.0, by contrast, is committed to building trust through phased, carefully managed rollouts of new capabilities. Every new service, he stressed, must provide real value to customers from day one, because in the end, reputation rests on the experience that users take away.

For Kapadia, this work is deeply personal. He has stepped away from his faculty role to lead MOSIS 2.0 because he believes there is an urgent opportunity to accelerate U.S. innovation in microelectronics. By lowering barriers, connecting emerging technologies with real customers, and building a network that spans academia, industry, and government, MOSIS 2.0 aims to do what MOSIS 1.0 once did: make it possible for more ideas to reach hardware, faster and at lower cost. In today’s environment of global competition and rapidly evolving technologies, that capability may prove just as transformative now as it was in 1981.

Distribution Statement A: Approved for public release. Distribution is unlimited.

Published on November 17th, 2025

Last updated on November 17th, 2025