Verbs first. In this announcement only one describes something done: the systems are live.
The rest are plans, expects, remains on track. That does not make them lies. It makes them promises, to be measured later.
Fact
The facts: on 31 August 2026 AMD, Cisco and HUMAIN, a PIF company, announced that AI infrastructure with AMD Instinct MI355X GPUs, AMD EPYC processors and Cisco Silicon One-based networking is live in Saudi Arabia and serving HUMAIN customers in the Kingdom and beyond. The Cisco N9000 platform with 800G optics interconnects the accelerators, and HUMAIN sells compute as a service, from model training to inference. The next phase, through the companies' previously announced joint venture, foresees up to 250 megawatts (250 MW) of infrastructure with AMD Instinct MI400 accelerators, EPYC processors and ROCm software; deployment is planned to begin in 2027, with first capacity expected in the second half of the year. According to the companies, the joint venture remains on track for up to 1 gigawatt by 2030.
One working megawatt says more than a promised gigawatt.
Why the word open keeps repeating
In the announcement open sounds like a chorus: open models, open software, open network. As we read it, behind the repetition is a commercial calculation. Sovereign AI is sold on the promise of greater control over where the customer's data lives and how its models are customized, and open software is how AMD buys into that promise.
The test comes in the second half of 2027. That is when we will see whether the 250 megawatts are plugged in or moved to the next announcement.
Official primary sources
AMD - AMD, Cisco and HUMAIN Expand Saudi Arabia's AI Infrastructure as AMD Instinct Systems Go Live, 31.08.2026The visual is generated code-art, no third-party images.