Outlook Supplement - Flipbook - Page 15
Forecast
photonic-MEMS manufacturing
ready-to-go, and strong market
demand from data centers are in
alignment to produce an exciting
new wave of products: integrated
MEMS-photonic devices.
Precision motion plus light equals
new solutions for data center networking, LiDAR systems, photonic
computing and many other advanced
applications.
Optical MEMS Boosts the
Microphone in 2026
KIERAN HARNEY, C E O ,
sensiBel
Across mobile devices, conferencing
solutions, spatial recordings, and
AI-driven interfaces, the demand for
higher-performance audio capture
is clear: Systems need microphones
with dramatically lower noise, higher
overload capability, and tighter partto-part matching to support multi-microphone beamforming and advanced
signal processing. Yet the traditional
capacitive and piezoelectric MEMS
microphones that have dominated
the industry for the past twenty years
have made incremental gains at best.
Saddled by fundamental architectural
constraints — micron-scale diaphragm
gaps, squeeze-film damping, backplate
noise, and mechanical clipping —
these older microphones limit both
signal-to-noise ratio
(SNR) and dynamic
range.
In 2026, this will
change as optical
MEMS microphones
deliver the first high-fiKIERAN HARNEY
delity audio capture for
consumer, prosumer,
and studio applications.
Optical MEMS devices are free
from the structural constraints
of capacitive and piezo MEMS
technology. Having replaced the
capacitive readout with an interferometric optical detection system in its
www.semiconductordigest.com
SBM100 series microphones, sensiBel
will begin volume production of a new
category of MEMS microphone that
breaks through longstanding performance ceilings.
With 80 dB SNR, a 14 dB noise
floor, and a 146 dB SPL acoustic
overload point (AOP), sensiBel’s
optical MEMS mic achieves a 132
dB dynamic range — a 20+ dB leap
past today’s best capacitive solutions.
The technology’s mechanical architecture, with diaphragm separation
in the tens of microns, eliminates
squeeze-film noise and prevents
stiction, enabling both exceptional
low-noise sensitivity and virtually
unlimited diaphragm movement.
These performance gains do more
than improve audio quality; they
simplify semiconductor system design.
A highly linear 24-bit digital output
(PDM, I²S or TDM) reduces signal-chain complexity and minimizes
calibration overhead in multi-microphone arrays. Applications once
hindered by microphone noise — such
as 3D spatial audio, AR/VR capture,
advanced ANC, AI voice interfaces,
and edge-based sensing — now support
redesigns with greater accuracy,
robustness, and user immersion.
Looking to 2026, optical MEMS
will expand beyond premium
adoption. Ongoing development
promises smaller packages and lower
power and will support scalable highvolume manufacturing for consumer
markets. As audio and acoustic
sensing become increasingly critical
to next-generation devices, optical
MEMS is poised to become a foundational semiconductor technology
— akin to the transition from electret
condenser mics (ECMs) to capacitive
MEMS more than two decades ago.
In an industry defined by breakthroughs, optical MEMS represents not
just an incremental improvement, but
the opening of an entirely new performance package.
Why RF’s Future in 2026
Lies in the Third Dimension
PAOLO FIORAVANTI, C E O ,
Circuits Integrated Hellas
As we head into 2026,
the radio-frequency
(RF) segment of the
semiconductor industry
is poised for robust
growth. Demand
is being driven by
advanced 5G deployments, expanding
PAOLO
non-terrestrial networks
FIORAVANTI
(NTNs), satellite
communications, and
defense modernization programs.
According to Fortune Business Insights,
the global RF semiconductor market is
expected to nearly double from about
$25.6 billion in 2025 to $50.3 billion by
2032, underscoring sustained demand
for RF solutions across applications.
Yet beneath these promising numbers
lies a persistent technical challenge.
Much of the RF industry today still relies
on planar, 2D integration of discrete
components — each optimized through
incremental material and process
improvements. While these efforts,
such as scaling gallium nitride (GaN)
and refining gallium arsenide (GaAs)
performance, have extended capabilities,
they are bumping up against the physical
limitations of 2D architectures.
In typical RF systems, separate
dies — for power amplification, signal
processing, and control — sit side-by-side
on a PCB and connect via wire bonds or
long traces. At high frequencies, these
planar interconnects introduce parasitic
effects that degrade signal integrity and
increase power loss. As RF designs push
into the Ka-band, and millimeter-wave
beyond, these inefficiencies become more
pronounced. Simply put, we are building
more capable transistors but connecting
them with longer, lossy “wires.”
The next leap in RF performance
— especially for systems where size,
weight, and power (SWaP) matter
Semiconductor Digest Supplement to January 2026
| 13
