Why UAE President Gifted Cerebras AI Superchip To PM Narendra Modi

Right after hosting a high-profile conclave of the foreign ministers of BRICS nations in New Delhi on Thursday (May 14), Prime Minister Narendra Modi landed in the UAE the next day for the first-leg of his five-nation diplomatic tour. 

While the announcement of the West Asian nation committing $5 Bn investment in the country grabbed headlines, much more happened behind the scenes. During a tête-à-tête, UAE President Sheikh Mohamed bin Zayed Al Nahyan presented a Cerebras chip to PM Modi. 

The exchange was not just symbolic. It marked the formal execution of Condor Galaxy India, an ambitious 8-exaflop AI supercomputing partnership between India and Abu Dhabi-based tech major G42. The physical chip handed to PM Modi is the foundational building block of this infrastructure. 

Under this agreement, 64 Cerebras CS-3 systems will be deployed to build one of the largest AI compute clusters in India. 

For context, supercomputers provide the computational power necessary to train complex AI models, while AI, in turn, optimises the performance of these computing systems. Even the current breed of LLMs rely on supercomputers, specifically massive clusters of GPUs. 

The partnership signalled the growing importance of AI infrastructure in India-UAE cooperation and the increasingly strategic role of compute in the next phase of cross-border cooperation.

For India, the partnership aligns closely with its domestic AI ambitions. It also helps bring sovereign AI compute onshore, reduce dependence on foreign cloud infrastructure, and provide startups and researchers access to frontier-scale computing power.

But what exactly are the Cerebras chips at the centre of the growing India-UAE tech bonhomie?

The AI Chips Powering India-UAE Tech Ties  

At the heart of this geopolitical tech partnership is a radical departure from conventional chip design. Traditional GPUs, which power most modern AI systems, are manufactured by printing hundreds of small chips onto a large circular silicon wafer. The wafer is then cut into individual chips for commercial use.

However, Cerebras Systems builds a single massive processor using an entire silicon wafer, an architecture it calls the Wafer-Scale Engine (WSE). This chip houses over 4 Tn transistors and close to 1 Mn AI-optimised cores on a single piece of silicon.  

Featuring a 44 GB on-chip SRAM that moves data at 21 petabytes per second and a proprietary on-wafer fabric, Cerebras’ WSE-3 claims to deliver an aggregate bandwidth of 214 petabits per second. It also claims to be able to train AI models with up to 24 Tn parameters without complex code splitting.  

By keeping the entire processing network on one continuous wafer, Cerebras reduces data communication latency, delivering up to 20X faster AI training and inference capabilities. 

Founded in 2015, Cerebras Systems is the brainchild of Andrew Feldman, Gary Lauterbach, Michael James, Sean Lie and Jean-Philippe Fricker. At the outset, the startup set out to solve a difficult challenge in semiconductor history – wafer-scale integration.

For decades, using an entire silicon wafer as a single processor was unviable. A single speck of microscopic dust during manufacturing could create a defect. Cerebras Systems solved this by building thousands of spare cores directly onto the silicon wafer, allowing the chip to automatically reroute data around any manufacturing flaws.

Since rolling out its first chip in 2019, the US-based startup has gone from selling individual hardware boxes to building giant cloud supercomputers.

Backed by the AI momentum, Cerebras Systems made one of the biggest tech debuts on the Nasdaq since 2019, raising $5.55 Bn just a couple of days ago. Its market capitalisation surged to as high as $95 Bn in the debut trading session. 

After President Al Nahyan gifted the wafer to PM Modi, Cerebras Systems cofounder and CEO Feldman took to social media to welcome the development. 

“We founded @cerebras with a vision to forever change AI compute… Today our AI super computer solution was the centerpiece of a collaboration between the Governments of the UAE and India. We are honored and humbled,” said Feldman in a post on X.

With chips in its kitty, work is all set to begin on one of the largest AI compute clusters in India.

Boost For India’s AI Ambitions 

The supercomputing initiative has been in the works for some time now. The project was announced by PM Modi and Al Nahyan during the latter’s visit to India in January this year. 

Subsequently, a term sheet was signed between G42 and India’s Centre for Development of Advanced Computing (C-DAC) to deploy an 8-exaflop supercomputer cluster in India during the India AI Impact Summit 2026.

Under the framework, G42, in partnership with C-DAC, will be responsible for the installation, deployment, operations, and maintenance of the system.

Once operationalised, the project will mark a big step forward for the Indian AI ecosystem. The proposed eight exaflop (or 8,000 petaflop) capacity is almost 19X compared to the combined peak 410 petaflop compute capacity of the country’s two flagship AI supercomputers AIRAWAT and PARAM Siddhi-AI, both hosted at C-DAC Pune. 

For starters, the supercomputer facility will help train and run large AI models much faster than conventional infrastructure, and offer greater control over data and latency. It will also offer affordable access to startups as well as other public and private sector entities for research, application development, and commercial use.

“The AI compute cluster will underpin a new era of joint R&D across sectors such as health and genomics, energy, and geospatial analytics, enabling researchers, institutions from both countries, and India’s emerging innovators to advance frontier science and address some of the most consequential challenges of our time,” G42 said in a statement. 

The project will likely be utilised to accelerate drug discovery, processing satellite data for disaster management, simulating efficient smart energy grids and fostering research.

A big part of the appeal is also data sovereignty. By hosting the infrastructure and data in India, the project helps address questions of data residency, security and national control. 

“India is one of the world’s great innovation economies. Deploying an instance of G42’s Intelligence Grid at this scale in such an important geography is what AI-native transformation looks like in practice. We are delivering infrastructure that converts energy and compute into sovereign governed nation-scale intelligence,” said G42 CEO Mansoor Al Mansoori during the latest meeting between the heads of the two states. 

The post Why UAE President Gifted Cerebras AI Superchip To PM Narendra Modi appeared first on Inc42 Media.