AI future

Last November, President Ali took a bold step to reorient Guyana’s developmental strategy into the brave new world of artificial intelligence (AI). He signed an MOU with Silicon Valley-based Cerebras to build a 100 MW data centre using the company’s revolutionary new wafer. The project would be the first under the company’s “Cerebras for Nations” programme, intended to help governments build, accelerate, and scale their sovereign AI initiatives.
It aims to co-design and build AI supercomputers powered by Cerebras hardware, develop AI models, and train local talent pools in AI and infrastructure skills. The ten-year-old company recently announced plans to launch six new data centres focusing on AI inference – the ongoing compute processing required for models to answer prompts – across North America and Europe to join inference clusters in California and Texas.
President Ali outlined his vision for the project – the first in the region: “This partnership is more than an AI data centre; it’s a declaration of Guyana’s ambition. Guyana is building a future where Guyanese talent powers global innovation, where its infrastructure supports frontier technologies, and where the nation leads the region in digital transformation.” The company confirmed it would deploy its WSE-3 AI clusters and infrastructure to serve international demand, as well as position Guyana as a “global destination for startups, researchers, and enterprises seeking high-performance compute capabilities in a favourable country environment.”
The MOU with Cerebras could have been seen as a gamble since the company was a goliath challenging the giant chipmaker Nvidia, which has dominated the AI industry with its postage stamp-sized GPU chip. Its core thesis is that GPU clusters are the wrong system architecture for frontier AI workloads, having to externalise communication across thousands of discrete chips connected by high-speed interconnects.
This turns AI into a networking problem: performance becomes constrained not by compute but by the cost of moving data between devices. Cerebras collapses that distributed system into a single, plate-sized piece of silicon – the WSE-3 (Wafer-Scale Engine 3) – with 4 trillion transistors, 19 times more than the Nvidia B200 and up to 15 times faster inference speeds. The WSE-3 is like a supercomputer on its own that integrates compute, memory, and interconnect onto one wafer-scale processor. It eliminates the need for most off-chip communication, trading network complexity for on-chip locality.
Cerebras was a startup that had been backed by several UAE companies, including G42 and MBZUAI, which will use WSE-3 to build the largest supercomputer in India. Most significantly, AI industry leader OpenAI (ChatGPT), which backstopped Cerebras with a US$20 billion contract needing 750 MW through 2028, with options for approximately 3 GW more by 2030 for inference services, had advanced Cerebras a $1 billion loan and received warrants for 33 million shares at near-zero cost. That gives OpenAI a direct economic stake in Cerebras’ IPO outcome. The strategic signal is still powerful: if the world’s leading AI lab is committing gigawatts to a non-Nvidia chip, the performance case is validated. And the need for a billion-dollar data centre in Berbice powered by gas from a second ExxonMobil pipeline becomes urgent.
Last Thursday, Cerebras finally went public on the stock market for the test on how it was evaluated by some of the smartest investors on the planet. Even as it increased the size of the offering to 30 million shares, it raised US$5.5 billion in its Initial Public Offering (IPO) of shares on NASDAQ, with the shares issued at $185. This was way higher than the initial range of $115 to $125, later raised to $150 to $160, recommended by its underwriters. The two founders were made instant billionaires. This US$5.5 billion is now available as cash to Cerebras that can fund the Berbice Data Center.
The stock then opened on Friday to public trading at $385, more than double (up 108 per cent), as retail investors bid up the price to grab shares, and ended the day at $311 and a $66 billion valuation for the company.
President Ali made a good bet.