Give us a brief history of Akoustis Technologies and what prompted you to start the company.
I have had a history going back to my time as a Howard Hughes Doctoral Fellow at UC Santa Barbara of working with novel, group III nitride materials.
My first foray as an entrepreneur began with my first company, RF Nitro, which was one of the first to develop a 50 V GaN RF transistor, roughly 20 years ago. RF Nitro was acquired within 18 months for 6x the invested capital by RF Micro Devices (RFMD), which is now Qorvo. I went on to run a division at RFMD for 13 years before starting Akoustis.
The idea behind Akoustis is to use high purity and single crystal piezoelectric materials to build a superior bulk acoustic wave (BAW) filter for ultra-high band (UHB) frequencies, as the two other BAW filter companies were focused on using polycrystal technologies.
I had worked with single crystal materials at UC Santa Barbara and believed that they would be able to produce BAW filters that could handle higher frequency, wide bandwidth and higher power and would, therefore, make superior products, as we transition to 5G and high frequency Wi-Fi 6 and Wi-Fi 6E.
The growth of wireless technology — principally cellular and Wi-Fi — has led to a proliferation of frequency bands from 400 MHz to now above 7 GHz, prompting an explosive demand for filters. While there is clearly a market opportunity, the market is already being served by formidable companies, most able to integrate the filter with the active components in the radio front-end. What is the proverbial mousetrap that will enable Akoustis to succeed?
We have advantages in each of the markets you mentioned:
In 5G massive MIMO (mMIMO) base stations, our advantage involves our high power handling capability, where our single crystal materials can handle twice the amount of power as polycrystal materials. This puts Akoustis in a class by itself for higher power 5G base station equipment. As a result, we expect to be a leader in BAW filters for the mMIMO base station market.
In 5G small cell network infrastructure, our advantage is our XBAW® performance and time to market. As a new market entrant, Akoustis must innovate and react quickly, relative to large competitors which compete using their scale.
In Wi-Fi, we were the first company to introduce the 5.2/5.6 GHz tandem coexistence BAW solution for high frequency Wi-Fi. This summer, we became the first company to introduce the 5.5/6.5 GHz solution for Wi-Fi 6E, the next generation of Wi-Fi that was ratified by the FCC in April of 2020. While we do expect competition in high frequency Wi-Fi, our XBAW solutions outperform our rivals, and we expect to gain significant share in both Wi-Fi 6 and Wi-Fi 6E, given our first-to-market positioning and superior performing filter solutions.
On the mobile front, we recently announced a second RF front-end manufacturing partner that we are developing a filter for, for use in mobile 5G handsets by the second half of 2021, earlier than we were projecting. We continue to see increased interest from other RF front-end makers for our solutions.
While we do have much larger competitors that may be able to provide high frequency BAW solutions, there are many others that do not have internal high frequency, high performance BAW. This includes some of the largest RF front-end and component manufacturers in the world. We believe we will be able to attract one or more of these manufacturers over time, as we show our ability to scale over the next 12–24 months.
Lastly, I would like to mention that while many focus on the continued integration of front-end modules, there is still a significant mobile market for discrete solutions, particularly among the Asian handset manufacturers. Where there is a focus on low cost handset options, manufacturers still develop the RF front-end on their own, rather than purchase full front-ends from the major manufacturers. This allows these handset manufacturers to develop phones that only use radios for specific markets.
Where a world phone may have 45 radios, and now more with the addition of 5G, a handset going into a specific province in China may only have 10 radios for users that do not typically travel outside of the city in which they live. This market is still 300–400 million handsets per year, which represents a significant opportunity for Akoustis, as the major RF front-end manufacturers tend to focus on the larger handset makers.
Tell us more about XBAW technology and what differentiates it from the acoustic filter technologies offered by your competitors.
XBAW is a proprietary and patented manufacturing process that we have developed over the past seven years, with the addition of the MEMS-based manufacturing facility we acquired in Canandaigua, New York, nearly four years ago.
There are significant differences in the way we make our XBAW filters relative to our competition, particularly with our high purity piezoelectric materials. We combine a MEMS process with high volume, silicon manufacturing scale to generate our high frequency resonator device, which uses an air cavity similar to other dual air cavity approaches and quite different from the other lossy, solid mounted resonator devices.
The XBAW process provides more flexibility than other dual air cavity approaches, enabling Akoustis to continue advancing the performance to achieve best-in-class results.
Where do you get your high purity and single crystal starting material and what other IP do you have?
Akoustis starts with bare silicon substrates and designs and manufactures its high purity piezoelectric and single crystal piezoelectric materials within our New York wafer fab operation. We take these custom, high performance piezo materials and utilize our in-house XBAW process to fabricate RF resonators and filters.
The only outsourced part of the supply chain to deliver completed XBAW products is our back-end packaging, although we work closely with these suppliers to develop optimal package solutions. As far as our outsourced partners, Akoustis uses some of the largest, most well-known packaging suppliers in the world to help deliver that final piece of the product.
Regarding IP, we have been awarded 34 patents and have another 75 patents pending. These patents start at the platform substrate level, move up through the piezoelectric materials, our manufacturing process and acoustic device level, the packages which we design, our complete filter product and the subsystem hardware solution. We also hold numerous trade secrets regarding our process that would make it extremely difficult for anyone to replicate our manufacturing flow.
How do the processing complexity and manufacturing cost of XBAW compare with traditional SAW and BAW?
SAW technologies have been around much longer than BAW technologies, and there is far greater competition in the SAW market. As such, the cost to produce SAW is lower than BAW, which is true for our BAW competitors that have been in the market for some time.
BAW filters are made with multiple photolithography layers, typically between 10–15 mask levels per filter, and this is consistent across polycrystal and single crystal BAW processes. Therefore, we believe our costs are in-line with our competitors.
What led to buying your own wafer fab?
When we started the company, we used an outsourced manufacturing partner to develop our filter technology. As we neared the production stage of our development, it became clear that we needed to develop our own internal capacity to improve cycle times and have greater control over our scale.
We were extremely fortunate that we were able to find and acquire the facility in upstate New York. The facility, a former Xerox fab, had over $88 million invested in the plant and equipment over the years before we acquired it.
Perhaps more importantly, and often overlooked, the acquisition enabled us to retain the roughly 30+ employees that were already working in the facility and knew how to operate the equipment. They quickly transferred and were able to improve our existing silicon process and worked closely with the Akoustis device team to develop our new XBAW process, which has set us up for significant commercial success.
Through the acquisition of the facility, we have been able to reduce our cycle times from three months to less than one month, and we are currently adding equipment and a second shift that will increase our manufacturing output by 5x over the next six months.
What markets are you targeting, and where are you achieving design wins?
We are targeting the 5G mobile, 5G infrastructure, Wi-Fi and defense markets.
In 5G mobile, we are engaged with a tier 1 RF front-end module customer and just announced a new RF front-end module customer that is expected to go into production in the second half of calendar 2021. Our focus is on the 5G UHB spectrum that will require coexistent filter solutions where our XBAW filter technology is differentiated from traditional technology.
In 5G small cell network infrastructure, we have received three design wins from a tier 1 customer that is currently ramping in the December 2020 quarter, and we expect an additional design win from this customer in the near term. We expect this customer to ramp with multiple filters as we go through calendar 2021. This has led to a second customer design win, which is expected to enter production in the first half of calendar 2021. Additionally, we have four other active engagements with 5G network infrastructure makers, two of which are tier 1.
In U.S. 5G infrastructure, with the recent auction of the Citizen’s Broadband Radio Service (CBRS) private access licenses (PAL) over the summer, the network operators that purchased the frequencies are looking to build out networks across the United States. We have received our first two design wins from a customer that we have worked with developing CBRS filters for over two years, given that the auctions were expected much earlier. This customer has ordered preproduction parts to develop both network infrastructure equipment and consumer premise equipment with our XBAW filters. Additionally, we are currently in active engagements with more than eight other network equipment providers and system on chip (SoC) vendors for the use of our CBRS XBAW filters in future products.
Finally, in the Wi-Fi market, we started to ramp production for Wi-Fi 6 with a tier 1 consumer-focused brand in the September quarter of 2020, which is shipping volume production today. We recently announced an additional Wi-Fi 6 win in the December 2020 quarter. In the next generation Wi-Fi 6E market, we have announced a first order with a tier 1 enterprise-class customer that is expected to go into production in the first half of calendar 2021. We have also signed a strategic purchase agreement with a second tier 1 enterprise-class customer that is expected to begin production in the second half of calendar 2021.
You mentioned working with a couple of mobile customers with opportunities that could ramp to production fairly quickly. You have said this would require a partner to initially support the required capacity. How would that work? Would you port your process to a foundry to scale volume and then retain them as a second source?
While we are scaling our production capacity by 5x by next summer, we would still not be able to produce enough filters to service a tier 1 handset manufacturer. Our expansion will bump our capacity into the 100s of millions of filter units, but the tier 1 handset market requires output in the billions.
Our current facility, however, can scale to 3–5 billion filter units per year once fully equipped with 8 in. wafer equipment. We will be well equipped to service the tier 2 handset market once the 5x capacity expansion is complete.
An RF module partner would be ideal as a bridge to enter the tier 1 5G mobile market, as it would help ease any reservations a tier 1 handset company might have regarding our ability to scale. Additionally, a tier 1 partner would help us acquire the equipment we would need to scale to billions of units per year.
Our current plan does not include outsourcing to a manufacturing partner, as it makes more sense to scale our current facility than take the time and bandwidth necessary to teach our process to a partner. A manufacturing partner would also add to the costs associated with our manufacturing process through margin stacking, which would cut into our profits over the intermediate to long term. Additionally, we do not want to take on the greater IP and trade secret discovery risk associated with outsourcing our process to a partner.
Are you using commercial CAD packages or your own proprietary software to design your filters? How accurate are these tools for filter design?
Yes, we are using commercial CAD packages to design our filters. As we continue to run our XBAW process across multiple filters at varying frequencies and bandwidth, we are growing our understanding of our own process using proprietary software and modeling refinement. This has led to significant improvements in our ability to model future filters, and we have experienced first-pass success on multiple filters over the last year.
What is the upper frequency capability of XBAW today? What determines that limit? Is that a focus of your R&D?
We believe the upper limit of our XBAW filters is somewhere between 10 and 15 GHz, limited by the thickness and uniformity of the piezoelectric materials, which are challenging with higher operating frequency. Such higher frequencies are not currently a focus of our R&D, as we are only at the very first stages of the 5G roll-out, Wi-Fi 6 expansion and Wi-Fi 6E roll-out.
To put that in perspective, revenues from the BAW incumbents have grown within this filter market to over $3 billion per year between 1.8 and 2.7 GHz. Looking ahead, 5G will add 3.2 to 5.0 GHz, nearly tripling the bandwidth for BAW over the next several years. Adding in 5.1 to 7.1 GHz Wi-Fi, the bandwidth for BAW is increasing nearly 5x with 5G and high frequency Wi-Fi. While it is likely that frequencies above 7 GHz will eventually come into play, we expect that 5G and Wi-Fi 6E will take several years to become mature, much like we have seen with previous generations of mobile technologies, including 3G and 4G. Right now, our R&D is focused squarely on the 3–7 GHz range.
The wafer fab you purchased in 2017 included STC-MEMS, a DoD accredited trusted foundry for MEMS processing, packaging and assembly. How does that business fit in your strategy?
With the ongoing pandemic, we have delayed a DoD trusted foundry classification, and we are focused on entering the high volume commercial markets. Additionally, we have not needed the designation to win our two current defense customers.
We may revisit this later if the defense market demands greater BAW filtration, but for now, we want to make sure we stay focused on the enormous commercial markets which we believe are ripe for a new entrant like Akoustis.
Tell us more about your background and the path that led you to acoustic filters.
My background begins with my education, which led to a Ph.D. degree in electrical engineering with emphasis on solid-state physics. In this discipline, I learned the power of leveraging advanced nano materials with superior physical properties to enable state-of-the-art electronic, optical and acoustical micro devices. In the case of acoustical resonator and filters, the selection of the piezoelectric material can directly impact the power handling and bandwidth of the filter.
In parallel to the technical path, I observed the migration of wireless standards from 3G to 4G and, now, 5G and how the RF front-ends in our consumer wireless devices are dominated by the selectivity of the acoustic RF filter. Combining the technical and marketing paths with a novel process flow and in-house manufacturing scale has enabled Akoustis to compete for sockets in some of the most advanced consumer wireless devices connecting our world.