Who holds the power in the client–foundry relationship?

Semiconductor tooling is extremely expensive, leaving fabless companies with little choice but to work with a well-resourced foundry. These foundries typically have full sales pipelines and can therefore be selective about whom they allow into their facilities.

You need representation that understands this dynamic from both perspectives, and has done foundry negotiations before, ensuring your proposed statement of work (SOW) demonstrates thoroughness and a nuanced understanding of the foundry’s needs—thereby avoiding red flags that could lead to project rejection.

You need someone who understands the difference between a reasonable request versus an unreasonable demand and avoids the latter in all foundry communications.

It can be difficult to recognize project dysfunction and suggest course correction without a diversity of perspective that comes from experience with many unrelated technologies. High specialization, high pedigree subject matter experts within your company may lack this alternate perspective. Our technical consultants have the pedigree to understand your device and the experience to ALSO help get the most effective foundry interaction. If your foundry relationship is contentious today, that is symptom of a problem that a consultant can help diffuse at costs substantially below the cost of ongoing dysfunction.

Foundries often staff with less experienced employees that might need to be guided by the client. So you don’t always get that kind of experience you will need from your foundry. Consultants can help bridge that technical communication divide.

High difficulty projects typically increase demand for experience to provide clarity on course corrections and solutions needed for a seeming technical impasses. This generalist perspective is rare to find in the permanent employee population exposed to fewer projects during their careers.

It is incorrect to think that consultants are an expensive option. Managing foundry relations is typically part-time with two weekly meetings: 1) to review data internally, and 2) to guide the foundry toward meaningful future technical actions. Consultants can offer a variable burn rate, increasing support at times of urgency for progress, and decreasing during contract negotiations or funding cycles, thus avoiding furloughs of permanent employees. And consultant presentations can often improve foundry understanding of technical and business goals in foundry native language, often resulting in faster alignment and lower cost of execution.

Please request a private consultation under NDA to gain a diverse perspective on your specific project.

A prototype is typically not a product. It is more accurate to think of a prototype as a test vehicle enabling the collection of important data to quantify risks, validate a working model of the device, and start testing risk mitigation strategies.

The physics of failure represents an infinity of problems to go solve — your device’s risk profile. A subset of financially imperative problems must be solved before production. A prototype is your telescope into the expanse of work to address risks.

This risk assessment phase is very early in the maturity of the project — usually concurrent with creating prototype samples. It is generally characterized as the testing of hypothetical failure modes one can usually surmise in advance. But, there is often a few problems that were not anticipated that change the working knowledge of the device. There can be a reality check at this milestone if budgets were too optimistic to fund the next phase.

The risk mitigation phase is the tough engineering work to resolve those financially imperative problems. It can take multiple tries (cycles of learning), before an acceptable path forward is identified, resulting in unexpected timeline slip, and cost overruns. A consultant can help try to minimize the number of pivots by paralelizing the solution discovery process.

You are still not done, because next is a risk production phase to test out the process control plan in support of production launch, and to uncover the low probability failure modes that were hidden during the risk assessment phase. Hidden issues may have been obscured by a much bigger problem that dominated the data early in the project, or maybe early runs were lucky.

In some cases the identified risk(s) may require a product redesign to design for manufacturability or improve reliability.

Therefore, it is more likely that 90% of the work has yet to be completed when a prototype exists, because it will highlight everything that is not right in your plans.

Please request a consultation under NDA to discuss your specific project risks and a structured approach to discovery and resolution. Also learn how to communicate this uncertainty to investors effectively.

No doubt you are brilliant and capable. No doubt you are the subject matter expert for the device you invented.

Are you also experienced at taking a prototype to production while managing the risk profile of your device? If not, do you really want to take unnecessary risks with the future of your employees and investor capital during your first attempt?

It is one thing to learn about FMEAs, Ford 8D, maturity level specifications, and statistical design of experiments applied hypothetically to textbook problems. It is another thing entirely to have navigated difficult projects, pivoted in response to new data, and apply these tools efficiently under real-world pressures.

Most risk mitigation work is the responsibility of the client who understands their own device, and of lesser concern to the foundry that is focused on reproducibility. It is recommended that you own this role, since it is your device, and your financial future.

At Persimmon Ridge Ventures, we have guided multiple disruptive technologies through difficult phases from prototype to production with financial success at front of mind in the process.

It’s the difference between having good camping gear, and having been back country backpacking multiple times in adverse conditions and survived. If this is your first backpacking trip, maybe an experienced guide would be a good idea when you survival is in question? We are your semiconductor industry sherpa.

The biggest mistake is assuming that anyone has any certainty of the scope, cost, and timeline of a custom process integration project to make a novel device during the quoting phase. I see CTO’s grasping for certainty, when everyone knows it is going to be time and materials, with a best guess as to approximate scope.

That includes the uncertainty with production wafer pricing, which depends on the difficulty and cost of processing and control measures put in place. That will not be known until the risk mitigation work is completed near the end of the project. Your financial margins at the end of the project are an educated guess. Don’t bet your business model on that initial wafer price, because it is an assumption.

Everyone loves a good physical model for the operation of a device, a process, or a project timeline, or a business case. However, incorrect assumptions are one of the largest project risks that explodes like a grenade later on, because the team often does not challenge assumptions. Everyone needs assumptions to model the working world. It is also true that all assumptions need to be identified as a risk as early as possible in the project, and validated as factual with good supporting data, so that it can continue to be used safely going forward for technical and business judgement calls.

A good bottoms up project plan that anticipates, prioritizes, and validates as many risks and assumptions as possible also helps to avoids unnecessary drama. The closer to production, the more resistant everyone becomes to unintended consequences from necessary changes (e.g. software update, system redesign), and the more visible supply disruption from yield issues, and reliability customer returns.

You need the risk mitigation plan before approaching a foundry to obtain an accurate quote, and an accurate quote before seeking funding from investors to scale. It is common to see clients attempting the reverse order.

Please request a consultation to develop your statement of work and secure multiple foundry quotes before engaging with investors about series A funding.

Many people do not realize the significant difference between fully custom deep-tech nano-fabrication and mass-produced, highly standardized semiconductor fabrication.

No computer chip manufactured today uses a custom process flow. Foundries compete to create process platforms, known as “a node,” to provide design companies with standardized process hopefully offering a competitive advantage for customer loyalty.

Custom process flows present a risk to standard flows. A facility contamination and resulting yield excursion is one such example. Therefore, your device WILL NOT share a clean room with Apple, Nvidia, Google or Amazon computer chips.

You may be able to get a spot in one of their older cleanrooms, but they may also want to be convinced that your company is a unicorn that will scale to high volume production. Custom process integrations and novel device fabrication require much more engineering support than those foundries may be willing to hire for. So your business model may be vetted as much or more than your device technical fabrication needs.

There is an alternative foundry ecosystem specializing in custom process integrations, novel materials, and engineering services to manufacture your device—though their names are not widely known.

Request a consultation to find out more about the matchmaking process to identify the right foundry partner for the capabilities you will need.