Shipping millions of RF and satellite chips worldwide, a firm is blending scale with a startup mindset. Discussing India’s path to building its own semiconductor products, Dr Naveen Yanduru from Axiro tells EFY’s Nidhi Agarwal about innovations and global operations.
Q. What does your company do, and how did its global presence evolve?
A: We are a semiconductor design company working across wireless infrastructure, satellite communications, industrial systems, and certain medical and defence applications. We originated as part of Integrated Device Technology (IDT), which was later acquired by Renesas Electronics. About a year ago, this business unit was divested, and the Murugappa Group formed Axiro Semiconductor as a new Indian company. The business itself is not new. It was built over about 14 years, starting with design centres in Boston and San Diego, and later expanding to Bengaluru, Shanghai, and Istanbul. Today, we continue to operate across these locations.
Q. Do you consider yourself a startup?
A. Not really. We are not a startup, but we still move like one. We are closer to a mature semiconductor business, but we retain a startup mindset for speed and flexibility. We already have scale, over 100 parts in production, more than one million units shipped every month, and over 120 unique products. We serve 80-plus customers, including Ericsson, Nokia, Hughes, and L3Harris. We also have over 145 patents, a team of around 120 people, and four global sites. Many of our customers came from our earlier association with Renesas, and now engage with us directly.
Q. What types of chips do you design, and where are they used?
A. We serve four main markets. First is wireless infrastructure. Our chips go into radio units on towers and building tops, enabling communication between mobile phones and the core network. We design radio-frequency (RF) front-end chips, including power amplifiers, low-noise amplifiers, RF switches, digital step attenuators, and variable-gain amplifiers.
Second is satellite communication. Our chips are used in ground terminals connecting to low Earth orbit (LEO), medium Earth orbit (MEO), and geostationary Earth orbit (GEO) constellations, enabling broadband connectivity. Companies like Hughes use our beamformer and up-down converter chips in systems like OneWeb. We also support in-flight connectivity using phased array antennas.
Third is RapidIO-based systems, where we provide switches and bridges for high-speed data transfer and cache coherency in applications such as medical and compute infrastructure. While PCIe is replacing RapidIO in some areas, these solutions remain relevant.
Fourth is Wi-Fi. We design front-end modules for enterprise networking equipment and provide modem-plus-front-end solutions for 60 GHz WiGig used in fixed wireless access.
Q. Which markets are you prioritising right now and why?
A. Right now, satellite communication is our top priority. With the growing number of satellite systems, including LEO constellations and in-flight connectivity applications, this market is very active. Our products in this space, beamformers, up/down converters, and RX modules, hold a leading position, and our customers value them highly. We are leveraging this advantage to focus on and further expand in satellite communications. At the same time, we continue monitoring other markets, such as computing, for future opportunities, which we may discuss later.
Q. What gaps in satellite infrastructure do you think are still not fully addressed?
A. There are three main challenges: signal-to-noise performance (G/T), power consumption, and cost. Low noise is essential due to the long distance between satellites and Earth. Power efficiency is critical because large antenna arrays can overheat, especially in remote deployments. Cost is equally important because phased-array systems require many chips, making affordability a key factor in commercial viability.
Q. How do you handle Doppler shift and fast-moving satellite links at the silicon level?
A. Doppler shift is a critical challenge, especially as video data increases. In traditional systems, analogue beamformers with phase shifters using two-time delays were used, which can address some Doppler issues. Implementing two-time-delay-based phase shifting at the silicon level helps, but it increases terminal costs and size because these components are bulkier and harder to design. At the system level, our customers also apply techniques to reduce Doppler effects when communicating with fast-moving satellites.
Q. What are the key technical challenges in building up-down converter chips for satellite links?
A. The main challenges are managing spurious signals and out-of-band noise, maintaining low phase noise, and ensuring strong image rejection. Keeping the intermediate frequency low helps downstream processing but makes filtering more difficult due to the proximity of image frequencies.
Q. You mentioned working on in-flight connectivity. How do you see its future from a semiconductor perspective?
A. In-flight connectivity will move entirely to electronically steered arrays, or phased array systems. Mechanical or legacy systems will be replaced. The focus will be on making these phased-array modules more efficient and cost-effective, but all future in-flight connectivity will use electronically steered antenna systems (ESAS) technology.
Q. What differentiates your technology and RF approach?
A. Our strength lies in deep expertise across RF technologies, complementary metal-oxide semiconductor (CMOS), gallium arsenide (GaAs), gallium nitride (GaN), bipolar-CMOS-DMOS (BCD), and radio frequency silicon on insulator (RFSOI), combined with strong circuit-level innovation. We focus on system-level improvements. For example, our RF switches use constant-impedance technology to maintain a stable VSWR and reduce reflections. A key differentiator is our use of silicon germanium (SiGe), which gives a strong balance of power efficiency, noise performance, and cost. We continuously evaluate alternatives like RFSOI and GaSb, but SiGe currently offers the best trade-off. Alongside performance, we place a strong emphasis on reliability through rigorous testing and qualification processes.
Q. What do you think are the biggest technical bottlenecks in the RF industry?
A. RF spans multiple markets, but the main challenge today is in cellular. While 5G is still being deployed, 6G lacks a strong business case, slowing the transition. However, opportunities remain across satellite, defence, aerospace, and wireless infrastructure, allowing diversification despite slowdowns in cellular.
Q. What are the biggest technical challenges in designing chips for 5G, 6G, and beyond?
A. The power amplifier is the most critical component. It must handle high peak-to-average ratios while maintaining linearity. It also needs to work with digital pre-distortion to meet efficiency, gain, error vector magnitude (EVM), and adjacent channel leakage ratio (ACLR) requirements. While other components matter, the power amplifier has the greatest impact on RF front-end performance.
Q. What changes do you expect in RF front-end architecture as networks move towards 6G?
A. The key trend is higher integration. As frequencies increase, systems require more elements, moving from 32×32 and 64×64 to 128×128 and beyond in massive multiple-input multiple-output (MIMO). Efficiently managing this scale requires tightly integrated RF front-end architectures.
Q. In Wi-Fi front-end modules, how do you manage coexistence and interference, especially in dense enterprise deployments?
A. We address interference through highly linear components with strong compression characteristics. Technologies like Zero Distortion minimise signal leakage, while Kalen technology maintains linearity across power levels, and Flat Noise ensures consistent receiver performance. Together, these improve coexistence in dense spectrum environments.
Q. What role does integration play in reducing costs for access point manufacturers?
A. When we talk about wireless infrastructure, I am usually referring to base stations and radio units. But for access points like Wi-Fi units, integration is equally important. Today, front-end components such as the power amplifier, filter, low-noise amplifier (LNA), and gain blocks are still discrete. As cost pressures increase alongside performance requirements, integrating filters with the power amplifier and LNA becomes essential. We are already seeing companies like Broadcom move in this direction. It not only reduces cost but can also improve overall performance.
Q. What design tools and flows are critical to your productivity and innovation?
A. We rely heavily on Cadence tools for integrated circuit (IC) development. For electromagnetic simulation and finite element analysis, we use tools like Ansys HFSS. Cadence Analogue Artist supports analogue design, while Mentor tools are used for digital back-end work, alongside Synopsys, where required.
Q. Are you using AI or ML in your design testing or performance optimisation?
A. Not in a significant way yet. Our work in RF, millimetre-wave, and analogue design remains difficult for artificial intelligence (AI) to handle because it relies heavily on practical engineering experience. We do use AI and machine learning (ML) selectively, mainly in digital back-end flows, but RF and analogue design remain largely unexplored for now.
Q. How do you protect your semiconductor IP and manage supply chain risk?
A. Most of our intellectual property (IP) resides in engineering expertise, which makes it inherently difficult to replicate. Even if someone reverse engineers a part, by the time they understand it, we have already moved ahead. Supply chain is a bigger concern. We depend on foundries and outsourced semiconductor assembly and test (OSAT) partners, many of which are in Taiwan, which introduces geopolitical risk. We are addressing this through second sourcing, but it remains an ongoing challenge.
Q. How do you differentiate yourself from competitors?
A. Our differentiation comes from a combination of engineering depth, innovation, and reliability. We compete with players like Qorvo, Analogue Devices, and Broadcom, where performance is critical. We focus on delivering measurable performance advantages through innovation, backed by rigorous testing and qualification. That combination of performance and reliability allows us to compete effectively with much larger companies.
Q. What is your go-to-market strategy with global customers, and how do partnerships fit into it?
A. Our approach is selective rather than aggressive. We do not chase every opportunity; customers come to us. Given limited capacity, we focus on choosing the right products, customers, and markets, and building long-term strategic relationships. Partnerships are central to this. Over 40 per cent of our business comes through distributors like EVV, RFMW, DTD, and Riosan. We actively support them through training, demand creation, and performance tracking to ensure strong market reach.
Q. How is your broader ecosystem, including foundries and tools, shaping up?
A. Our ecosystem is well established and running smoothly. On the foundry side, we work with TSMC, GlobalFoundries, Tower Semiconductor, Dongbu HiTek, Win Semiconductor, and UMC. For OSAT, we partner with ASC, MCOR, UTL, and Carsem, and are exploring opportunities with CG Semi within the Murugappa Group. For electronic design automation (EDA), Cadence and Synopsys are our primary partners.
Q. Are you looking for new distribution or channel partners?
A. Not actively. We already have a strong global network of distributors and independent design houses (IDH) partners. We may selectively onboard new partners in specific regions, but overall, the current network meets our needs.
Q. Which emerging technologies do you see as the most promising opportunities?
A. Silicon photonics stands out, particularly for optical communications. It is still in an early stage but offers significant potential. However, it is a system-level play that requires strong ecosystem partnerships, especially in packaging.
Q. What do you see as the next big opportunity in wireless or satellite silicon?
A. There are multiple opportunities, but the key is aligning technology roadmaps with market evolution. We are focusing on radio frequency transceivers as systems scale in frequency and element count, across both satellite and wireless infrastructure, including hybrid and fully digital beamforming. Power amplifiers are another critical area, especially around efficient load modulation and linearisation. Companies that solve these well will lead the next generation.
Q. What role are startups playing in India’s semiconductor transition, and what challenges do they face?
A. Startups are essential. We want to see more of them reach the stage of shipping real products, what I call becoming ‘ex-zeros’. The biggest challenge is investment. Semiconductor businesses require significant upfront capital and long timelines before revenue. While India has strong talent, much of it is still in multinational companies. As more product companies succeed, both investment and ecosystem support should improve.
Q. Is government support sufficient, and what more is needed?
A. Government support is improving. Initiatives like ISM have strengthened manufacturing, particularly in OSAT and foundries. Support for fabless companies is evolving, but there is room to expand.
Going forward, incentives linked to actual product shipments, not just early-stage funding, could make a big difference. Encouraging companies like Ericsson, Tejas, or Jio to adopt domestic chips would help drive real productisation and attract investment.
Q. Where does India stand today in semiconductor product IP, and what needs to change?
A. India is not yet a true semiconductor product nation. Axiro is among the few companies shipping its own ICs at scale. Most devices still use chips from companies like Qualcomm, Broadcom, Renesas, Sony, and MediaTek. That said, things are starting to change; our chips are already used in systems from Ericsson, Nokia, and even ASML. India has the talent and global exposure; many leading semiconductor firms are run by Indian engineers. The missing piece is investment to build and scale more Indian product companies.
Q. What should India do to strengthen its position globally?
A. The fundamentals are already in place: talent, connectivity, and government intent. The key challenge is unlocking investment and scaling companies. In terms of standards, these are largely driven by industry players like Qualcomm, MediaTek, HiSilicon, and Samsung. To have influence in bodies like 3GPP, India needs more domestic companies actively participating.
Q. What does a globally competitive Indian semiconductor ecosystem look like?
A. It means having many companies like Axiro. If you build enough, one or two can reach the scale of global leaders like Qualcomm or Broadcom. It is ultimately a numbers game; more ‘shots on goal’ increase the probability of global success.
Q. What are the current challenges your company faces in growing fast?
A. Talent acquisition is the main constraint. Growth depends not just on design but on operations, testing, qualification, and yield engineering, which form the bulk of the work in a product company.
Q. What are your plans for the future?
A. It would include many companies like Axiro, with at least one reaching the scale of global leaders. Achieving this requires a broad base, increasing the probability of global success.
Q. Do you have any advice or message for investors, engineers, customers, or emerging entrepreneurs in this technology?
A. Entrepreneurs should stay resilient despite the odds. Investors should recognise semiconductors as a high-potential sector. Experienced professionals should explore alternative paths to build Indian product companies without starting entirely from scratch.
