From silence to sound: MEMS brings a breath of fresh air – An interview with xMEMS

In the audio sector, one of the latest devices to turn to MEMS technology is the speaker. After decades of development, the first MEMS microspeakers are integrated in consumer TWS earbuds and are expected to penetrate even more end-systems in other markets. This was made possible by the development of piezoelectric materials and the improvement of PiezoMEMS technologies (Source: PiezoMEMS Technologies 2025 report, Yole Group), which are widely used for RF filters, inkjet printheads, and, more recently, for optical MEMS, fingerprint sensors, timing oscillators, and micro-cooling solutions.

With the ability to be integrated into TWS earbuds, smart glasses, AR headsets, and smartwatches, MEMS microspeakers offer several advantages over legacy technologies like electrodynamic speakers (ES) and balanced armature (BA) speakers. This includes lower power consumption, high repeatability and uniformity, greater assembly reliability, immunity to dust and water, and the potential for a lower average selling price (ASP) through economies of scale. According to the Status of the MEMS Industry 2025 report from Yole Group, the MEMS microspeaker market was worth $3.9 million in 2024, representing approximately 2 million units, and is expected to grow at a 79.8% CAGR_24-30.

Companies such as xMEMS, USound, and AAC Technologies are leading players in MEMS microspeakers, with xMEMS and USound together accounting for 93% of the market. While other companies, such as Bosch, SonicEdge, MyVox, Fortemedia, and Audiopixels, are developing their own solutions.

xMEMS has been one of the pioneering start-ups leading the development of MEMS microspeakers. Today, the company has already built a portfolio that includes four MEMS microspeakers, two loudspeakers, one MEMS acoustic vent, and one MEMS microcooling solution. The path forward is clear for xMEMS, now focusing on two hot topics: audio and electronic cooling.

Yole Group’s analysts, Clyde Midelet and Pierre-Marie Visse, are pleased to have the opportunity to speak with Joseph Jiang, Co-founder and CEO of xMEMS. Check out this interview to see how xMEMS is addressing the audio industry, positioning itself as a leading innovator in the future of MEMS technologies.

Clyde Midelet (CM): Please briefly introduce xMEMS and its core mission in the MEMS industry.

Joseph Jiang (JJ): Founded in 2018, xMEMS set out to bring groundbreaking MEMS innovations into everyday life. From the beginning, we set the bar high: to create small, low-power, full-range loudspeakers capable of replacing conventional smartphone speakers, and to develop MEMS speaker arrays that could transform home entertainment by eliminating the need for bulky woofers in the living room.

Today, we are halfway toward those goals — and new, urgent needs are expanding our mission. Demand is growing for advanced MEMS speakers in wearables and for enabling natural AI-driven voice interfaces, further reinforcing our commitment to redefining how people experience sound.

CM: Should we expect similar growth for MEMS microspeakers as witnessed with MEMS microphones?

JJ: For the sake of clarity, it is important to point out that we are using the term “MEMS speakers” instead of “MEMS microspeakers”. Many people associate microspeakers with earbuds and tiny wearable devices. Since day one, our technology has been aiming at loudspeakers, which many thought wasn’t possible with MEMS.

Back to your question… Yes, indeed. From the very beginning of xMEMS, we have believed that MEMS speakers will follow a growth trajectory similar to, and potentially greater than, MEMS microphones. The benefits are simply too compelling. Unlike traditional dynamic speakers, MEMS speakers are manufactured on a foundry-based semiconductor process rather than a labor-intensive factory line. This means scaling production does not require massive manpower, enabling faster ramp-up and more predictable supply. In addition to delivering superior audio and more consistent chip-to-chip performance, MEMS speakers offer a compact size and SMT compatibility — both of which are critical for customers.

Price-wise, we expect that MEMS speakers will eventually be comparable to conventional coil speakers. With the rapid adoption of xMEMS µCooling technology and MEMS speakers, the wafer price will come down in the next year or two, which will lead to a higher adoption rate, similar to MEMS microphones.

Just as MEMS microphones reshaped the audio capture market, we believe MEMS speakers are poised to redefine audio playback across consumer electronics.

Pierre-Marie Visse (PMV): What are the key opportunities and challenges, and how does xMEMS plan to address market needs? What is its roadmap?

JJ: The key challenge for MEMS speakers has always been physics. To produce audible sound, speakers must move a large volume of air, and this requirement has traditionally favored larger diaphragms and greater displacement — neither of which are natural advantages for MEMS devices.

Recognizing this from day one, xMEMS pioneered an ultrasonic piezoMEMS platform designed to overcome this limitation. Today, we can demonstrate performance that is up to 40x louder at low frequencies than conventional “air-pushing” approaches of other MEMS speakers, proving the viability of solid-state microspeakers in applications once thought unattainable. After seven years of development, we are proud to announce the success of this breakthrough platform.

The opportunities ahead are enormous. Virtually every voice interface and AI-enabled device requires microspeakers — from AI glasses to earbuds and wearables, to smartphones and even next-generation loudspeakers. These markets are rapidly converging on the need for smaller, lighter, more uniform, and more reliable speakers that also deliver superior sound quality.

Our roadmap is built on scaling our Sound-from-Ultrasound™ platform across multiple product families — such as Sycamore for glasses and wearables, Cypress for ANC earbuds, and future MEMS speaker arrays for home and entertainment systems.

In short, the challenges of physics drove our innovation, and today they are becoming our competitive advantage. The result is a clear roadmap: leverage the unique strengths of MEMS to transform audio and thermal solutions across the full spectrum of consumer and enterprise electronics.

PMV: What are the main technical challenges in scaling MEMS-based speakers for mass production?

JJ: Scaling MEMS speakers for mass production brings both technical and market challenges. In November 2023, we achieved a major milestone with our first MEMS microspeakers shipping in consumer-grade earbuds, the Creative Aurvana Ace and Ace 2. By 2024, we had shipped over one million units into earbud applications as high-frequency tweeters. While the Aurvana Ace 2 won multiple global awards and was praised for its “amazing sound,” acclaim alone does not drive mass adoption.

The core challenge is that for MEMS speakers to scale, they must move from being a “nice-to-have” to being a “must-have” component. The first key step is to deliver speakers capable of full-bandwidth sound for the target application. With our Sound-from-Ultrasound™ platform, we now offer a new class of full-range speakers that are up to 10x smaller and 10x lighter than traditional drivers, while enabling superior audio performance in form factors where conventional speakers simply cannot fit. This shifts MEMS speakers from niche innovation to essential enabler across wearables, glasses, and next-generation earbuds.

As adoption grows, volumes naturally scale wafer consumption, which in turn lowers wafer costs through manufacturing efficiencies. This virtuous cycle drives down price points and removes lingering doubts about MEMS speaker viability, paving the way for broader market penetration and increased market share. In short, the biggest technical challenge — proving MEMS speakers can scale — has become our biggest advantage as we align performance, form factor, and cost with the needs of high-volume consumer electronics.

CM: The microspeaker sector is highly competitive. So, what differentiates xMEMS’s technology and products from other manufacturers (both legacy and MEMS technology)?

JJ: The microspeaker market is indeed competitive, with conventional coil-based speakers still enjoying a cost advantage due to decades of scaling. However, pricing alone will not define the future. The real opportunity for MEMS is in defying traditional physics—delivering greater sound output from dramatically smaller and lighter form factors. These advantages are not incremental; they are transformative. MEMS speakers are quickly moving into emerging categories like AI glasses, where thin, light, and reliable solid-state solutions are critical to adoption.

CM: Why are almost all the players involved in MEMS microspeakers new or emerging companies?

JJ: It’s true that most MEMS microspeaker players are relatively new entrants, though there are a few established companies like AAC that have explored this space. Years ago, when I worked at InvenSense, we also evaluated MEMS speakers, which shows the idea has been circulating for some time. What is different now is that xMEMS has proven the technical and commercial viability of MEMS microspeakers at scale. This demonstration will inevitably attract more established players into the market. In other words, it’s not just about startups—success breeds competition, and as MEMS gains traction in audio, we expect increasing investment from across the industry.

CM: As the leader in the MEMS microspeaker segment, what other opportunities do you see for xMEMS to leverage its piezoelectric technology and expand into new markets over the coming years?

JJ: Our µCooling™ air-pump-on-a-chip is a prime example of how we can extend our piezoelectric MEMS platform beyond speakers. The concept dates back to 2017, when our work on MEMS loudspeaker architecture showed the potential to move meaningful volumes of air at the micro-scale. That insight led directly to µCooling, which is now positioned as a breakthrough solution for active thermal management in smartphones, XR glasses, SSDs, and data center applications.

Looking ahead, we see additional opportunities for piezoMEMS in areas where size, efficiency, and solid-state reliability are critical. There are a couple of new applications that we are working on, and we will share details with you once they are ready.

PMV: Is xMEMS working on alternatives, and why is it particularly challenging to replace PZT?

JJ: We have explored alternative materials to replace PZT in our piezoMEMS devices, and we have seen some promising results in collaboration with our foundry partners and select customers. However, despite years of research across the industry, PZT remains unmatched in its combination of electromechanical efficiency, maturity of processing, and reliability. For now, it continues to provide the best performance for our applications, though we remain committed to evaluating new materials as they mature.

CM: Looking at the long term, how do you envision xMEMS’s business evolving? What are your next milestones in both the short and long term?

JJ: We anticipate the speaker business to grow exponentially in the next few years, which will drive the wafer price down and enable more customers and applications to adopt MEMS microspeakers and loudspeakers. Our near-term milestones are to secure a majority share in newly emerging AI interface applications such as glasses, watches, and other wearables, where MEMS microspeakers provide clear competitive advantages over legacy coil technology.

Longer term, our roadmap includes next-generation loudspeakers for smartphones, automotive, and home entertainment, extending our leadership into markets that demand both scale and uncompromising performance.

Our µCooling platform will also be a significant pillar of growth and diversification for the company, giving xMEMS a strong foundation in both audio and thermal management. With the majority of electronic devices being thermal-limited, not compute-limited, we expect µCooling, the world’s smallest active thermal management solution, to become a standard across multiple electronic categories, where conventional solutions simply cannot fit. µCooling will enter production in 2026.

Source: www.xmems.com