Where does India's Vikram 32-bit microprocessor stand against global peers?

While India hails its first made-in-India space chip, global peers work on building 64-bit processors and even chips with built-in AI support

India has built its first space-grade computer chip for rockets, called Vikram 3201. It is the country’s first indigenous 32-bit microprocessor designed to survive the extreme conditions of space launches.

 

The chip was designed by the Indian Space Research Organisation's (Isro) Vikram Sarabhai Space Centre and made at the government’s Semi-Conductor Laboratory (SCL) in Chandigarh. It has already been tested on PSLV-C60’s POEM-4 platform and was formally handed over for production use in March 2025 before being showcased at Semicon India 2025 on Tuesday.

 

All about Vikram 32-bit microprocessor

Unlike consumer chips, space processors prioritise radiation tolerance, deterministic behaviour and certifiability over raw speed. Vikram 3201, built on a 180 nm CMOS process, falls in line with international practice where slightly older nodes are preferred for their radiation resilience. It supports floating-point arithmetic and is backed by an Ada programming toolchain, with a C compiler under development.

   

The chip is intended mainly for launch-vehicle navigation, guidance and control, but its heritage could extend to satellites and planetary missions.

 

Before Vikram 3201, Isro’s had developed Vikram 1601, a 16-bit part, which was in use since 2009.

 

How does it compare with other chips in other countries?

For decades, the US and Europe have relied on similar 32-bit chips for space missions:

 

RAD750 (US): Used in Nasa’s Mars rovers and deep-space missions.

LEON (Europe): Powers many European spacecraft.

 

These are older by consumer standards, but still widely trusted because they don’t fail under radiation or extreme temperatures. Vikram-32 is now in the same league.

 

Does chip size matter?

The size or process node (like 12nm, 7nm, 5nm) of the chip is also important as it can provide a sense of how small the transistors and other features on the chip can be made. A smaller chip size means the chip can be faster, use less power, and fit into smaller devices. Overall, this can also impact weight of the system which is crucial in aircrafts.

 

Vikram 3201 is built on a 180 nm CMOS process. In comparison, China's Loongson CPUs launched last year is fabricated on a 12nm process. US' HSPC size is not known, however, its predecessor RAD750 processor was based on IBM’s PowerPC 750 design, was originally fabricated on 250 nm and later improved to 150 nm.

 

For context, Intel's Panther Lake (Core Ultra 300) series is fabricated on the advanced 18A process node -- which is approximately 1.8 nm. Their 16-bit x86 processors also fall within the 1.8–3 nm range. However, Intel chips are primarily used for ground operations in space missions, not in the aircraft themselves.

 

Space chip race: Moving to 64-bit processors and AI

While India has caught up to legacy international practice, the US and Europe have already begun moving to faster, 64-bit processors with multiple cores, some based on the new RISC-V architecture, and even chips with built-in AI support. These are aimed at data-intensive payloads such as Earth observation satellites, astronomy instruments, and autonomous spacecraft.

 

United States: Moving from reliability to speed

Leading in space computing, Nasa’s High Performance Spaceflight Computer (HPSC) is being developed as a fault-tolerant, multicore RISC-V system-on-chip designed to deliver up to 100 times the performance of current flight processors. In 2022, Nasa’s JPL awarded the contract to Microchip and SiFive. First modules (PIC64-HPSC family) began emerging for testing earlier this year.

 

The new chips are expected to offer fault tolerance, AI-capable vector processing, Ethernet networking, and improved power efficiency. This should enable spacecraft to make smarter, faster decisions autónomously and be used in future missions (robotic and human) capable of edge computing through the 2040s.

 

Europe: Moving to 64-bit RISC-V

The NOEL-V, a 64-bit RISC-V successor to the long-serving LEON line, is gaining adoption as a modern and flexible architecture for space missions. The physical size of the chip has not been disclosed.

 

China: Indigenous CPUs and AI in orbit

Loongson CPUs were delivered to China's space station, Tiangong, via the Tianzhou-8 resupply mission in November 2024. Onboard, Loongson modules now support tasks like remote sensing, radiation-source tracking, and in-orbit chip testing—critical steps for validating indigenous hardware under space conditions.

 

In June 2025, China unveiled Loongson's 3C6000 CPU, which comes with 64 cores, 128 threads, 32MB of cache,and  four 72-bit memory channels, according to media reports.

 

China has also begun deploying the first steps of a planned AI supercomputer constellation in orbit, enabling real-time, onboard data processing -- reducing dependence on Earth-bound processing centres.

 

Russia: Slow modernisation

Russia's GLONASS-M navigation satellites continue using the K1839 processor, a design tracing back to the 1980s VAX-era computers. These chips remain proven and reliable, though outdated.

 

Russia developed the KOMDIV series, which are MIPS-based in both 32-bit and 64-bit in the mid-2000s and are been slowly upgrading them.

 

Japan

JAXA and Mitsubishi Heavy Industries began a joint effort to develop new space-grade processors domestically in 2020.

 

The bottom line

India has closed the gap to proven international standards with Vikram 3201, but the global race is now shifting toward high-performance, 64-bit and AI-enabled processors.