- Follow me on Twitter @microcapreturns
As the next generation fabs start to ramp up in 2022-2025 the royalty potential with the adoption of Atomera’s MST technology must be fully understood to appreciate the transformational effect on the company if only one new node adopts. Recent published data estimated that 5nm fabs will have wafer costs of $17,000/wafer. My estimate is the 3/2nm fabs will likely have wafer costs closer to $25,000/wafer. Numbers like these should make it very clear the 370 older fabs will be used for years to come and are desperate for ways to improve capacity which is also a huge benefit of adopting Atomera’s technology.
- TSMC’s Wafer Prices Revealed: 300mm Wafer at 5nm Is Nearly $17,000 | Tom’s Hardware
As seen in the above article wafer costs will be quite high for next generation fabs. Using the slides from Atomera’s website the estimate for a 28nm fab($3,000/wafer), a 2% royalty rate and 80k wafers/month led to a $58m/year royalty rate. See below.
Using a slide from Atomera’s slide deck I have added both the 5nm and 3nm wafer estimates to the 2019 data. As can be seen using a range of 20k-80k wafers/month and 2% royalties the numbers explode. For a 5nm fab royalties could be $81m – $326m/year for one new node and between $125m – $500m/year for a 3nm node. While these numbers are huge for a company like Atomera, given its 23m shares outstanding and zero debt, the royalties are noise to a large fab selling wafers at $17,000 per wafer. Which means there is no reason not to adopt the technology if it provides advantages that can not be achieved by any other means.
There are three key parts to any technology adoption.
- Does the technology do what it needs to do i.e. improve performance, reduce cost etc..without impacting reliability
- Is it cost effective i.e. cost benefit analysis
- Can it be implemented into a high volume production environment
The first item is by far the easiest to answer. They have half the worlds largest semiconductor companies working with them. If the technology didn’t work they would have dropped out years ago. The first thing any large company does when testing a new technology is to try and prove it does not work. They don’t want to waste precious development dollars on something that has no benefit. The fact none of these 19 companies have dropped out tell us they see the benefits and this represents hundreds of engineers across the space. Not just a few people but teams of the top engineers in semiconductor development, testing and manufacturing.
The next question is really the business case. The greatest technology in the world will never be adopted if there is no business case. With the royalty model, 1-3%, adopting MST really is no cost to fabs. The numbers easily show that the fab can make a lot more money, increase capacity or reduce water, power or emissions by adopting MST. Its a win/win for both companies.
Given the first two items indicate a high likelihood of adoption the real question is can it be implemented into today’s high volume semiconductor environment? Besides the obvious realization that no company would work and spend money on a new technology if they didn’t strongly think it could be implemented into production there are a few other indicators from the first JDA that convinced me it is very likely. The first is what has been mentioned about the JDA customer. The central engineering group working on approving MST for usage by the entire company(3-6 nodes to start) are not your typical engineers. These are the top gray beards who have been with the company for decades. They have the most experience taking new technology into production. They are the best of the best and their reputations are on the line. This group would never have signed off on the tech transfer if there was any doubt about viability. But it takes time. Your building, testing, analyzing hundreds of thousands of parts. Looking at lot to lot variability, doing life tests and many other items to prove it out. Keep in mind these parts are hundreds of millions if not billions of dollars in future revenue for them and they can’t find out years down the road there are issues. Another key piece that the company has talked about with the JDA customer is the central engineering group becomes the advocates within the company for adoption. This is very important because manufacturing groups work with these engineers and if they support the technology it is very likely to be adopted once given the approval.
The real question investors keep asking themselves is why is it taking so long for deals to be signed and adoption. For the deals to be signed one item often overlooked is the size of the numbers being discussed. They aren’t signing a deal for a few million dollars/year. These are hundreds of millions of dollars per year that go for decades. When the potential is to become the industry standard both sides want to be protected. The key is this opportunity is not going away. MST is applicable to new nodes(see Mears blog below) and legacy nodes so the company should not rush into deals unless they get what they want. Sure they could probably lower the royalty rate or give up future benefits but that would leave a lot of money on the table.
The other issues with timing could be as simple as two of their licensees being delayed with covid and a fab fire. STMicro has delayed their new fab about a year and the other licensee, AKM, had a fab fire. The good news with STMicro is they will start moving equipment into their new fab this quarter. The other reason there may also be delays in adopting is the current state of the chip shortage. To start the Phase 4/5 developments you need to be in manufacturing which means that whatever is being currently built is not being built. Given all these fabs are running at full capacity this has to be worked in. On one of the recent calls it was mentioned that multiple customers have completed all their Phase 3 testing and had all the data they needed to move forward. I strongly suspect they are waiting for manufacturing to open up scheduling space to start phase 4/5. They also may be waiting on new equipment, which also has seen long delivery cycles, that will allow easier adoption of new technology as capacity catches up to demand.
As the industry grows at almost a 10% CAGR going forward and are already at capacity fabs will find ways to grow capacity. While the newer fabs coming online in the next few years will be part of that solution the existing fabs will need to come up with ways to increase capacity and the adoption of Atomera’s MST is by far the most cost effective way to do it.
- MST® Dopant Blocking for Advanced 3D Applications
- Whitepaper by Robert Mears, released 6/29/21
- The paper puts the new MST data in context for advanced 3D architectures for the 2/3nm node, such as nanosheets or gate-all-around (GAA) devices.
Any due diligence from this site is for entertainment only and not a solicitation to buy or sell Atomera stock. Any estimates are just examples of what is possible and should not be considered financial advise. I have not been compensated in any way and will never be compensated for my reports.