Vice-President and Head of Research at the IMANI Centre for Policy and Education, Bright Simons, has raised concerns about Ghana’s national artificial intelligence (AI) strategy, questioning the accuracy of its computing capacity targets, the feasibility of its infrastructure plans and the affordability of its projected investments.
In a post on X, Mr Simons revisited concerns he raised during Joy News’ Digital Economy Forum on AI and the Ghana Economy on Wednesday, arguing that the strategy contains what appears to be a significant mathematical error in its computing capacity targets.
According to him, the document sets Ghana’s computing capacity target at “1021 FLOPS” for 2027 and “1025 FLOPS” for 2035.
Mr Simons argued that, if interpreted as ordinary numbers rather than powers of ten, the figures would fall far short of the computing capacity required to achieve the government’s ambitions.
“If this was indeed correct, it would mean that the Ministry would like the entire computing capacity of the country in 2027 to be equivalent to a 1970s desktop computer. Or 2% of the capacity of the 1980 Intel math chip,” he wrote.
He suggested that the intended targets were more likely (10^{21}) and (10^{25}) floating-point operations per second (FLOPS), respectively, rather than the figures as presented.
The apparent discrepancy, he said, raises questions about the technical expertise and scrutiny involved in preparing the strategy.
“Somebody who didn’t fully understand how FLOPS work must have been involved in the drafting,” he stated.
Questions over the cost of computing infrastructure
Mr Simons argued that interpreting the 2027 target as (10^{21}) FLOPS would imply a substantial investment in computing hardware, electricity and advanced semiconductor chips.
He estimated that achieving this level of computing capacity could require approximately 200,000 of Nvidia’s newest chips and 500 megawatts of electricity, at an estimated cost of US$24 billion.
He said the projected expenditure would represent a substantial share of the government’s programmed spending for 2026.
“The US$250 million AI data centre now being procured supplies about 1% of this 2027 target,” he noted.
The estimates form part of his broader criticism of the strategy’s proposed scale and affordability. However, the assumptions underpinning the computing requirements, chip numbers, power demand and cost estimates would require independent technical and financial assessment.
Mr Simons also questioned the process through which the strategy was reviewed and approved, arguing that the apparent error suggested inadequate scrutiny by technical experts and policymakers.
“Clearly, no serious critique of the plan would have okayed such a vision. Throughout the Ministry’s own review, cabinet approval, and a high-profile media launch, it is absolutely clear that not one critical reviewer was engaged,” he wrote.
He linked his concerns to what he described as “classic KATANOMICS”, referring to his criticism of a sugar factory project that he believes proceeded without adequate scrutiny.
“This is exactly how a sugar factory that was ‘designed from the start to fail’ was allowed to be built. Many programs and initiatives in Ghana fail for precisely this absence of critical review,” he added.
Energy targets also under scrutiny
Beyond the computing capacity figures, Mr Simons questioned the strategy’s stated ambition to increase energy capacity by “0.5 to 1 OOM per year”.
He interpreted the wording as referring to an increase by orders of magnitude and argued that, if taken literally, the projection would imply an extraordinary expansion in Ghana’s electricity output over the strategy’s implementation period.
“Did anyone in Cabinet not protest that this projection in 8 years would mean expanding Ghana’s electricity output by 10,000x?” he asked.
He also expressed doubt that the Energy Ministry would have committed to such an expansion, saying he expected the minister responsible to deny making such a commitment.
The precise meaning of the energy target remains an important point requiring clarification from the government, particularly regarding the intended unit, measurement and projected increase.
GhanaChat and the national AI ambition
Ghana’s national AI strategy includes plans to develop GhanaChat, a platform intended to compete with established AI systems such as ChatGPT, Gemini, Kimi, Qwen and Claude.
Under the proposal, GhanaChat would be trained and hosted in a new AI data centre, with 48 public-sector registries expected to contribute data to improve its ability to respond to Ghana-specific needs.
The proposed AI data centre and an accompanying AI Fund are estimated to cost approximately US$270 million, with additional expenditure anticipated for registries and applications.
The strategy is being pursued by the Ministry of Communication, Digital Technology and Innovations under Minister Sam George as part of the government’s ambition to position Ghana as a leading artificial intelligence hub in Africa.
Simons proposes a smaller starting point
Rather than immediately pursuing a large-scale national AI model, Mr Simons has proposed beginning with a sovereign AI assistant designed specifically for the civil service.
He estimated that such a system could operate with approximately 144 graphics processing units (GPUs), delivering around (7 \times 10^{17}) operations per second at eight-bit precision.
He put the initial cost at US$13 million, with annual operating expenses estimated at US$6.7 million.
“That is under a thousandth of the 2027 target, and a far more realistic initial-term strategy,” he wrote.
Mr Simons also questioned the need for Ghana to train its own foundational AI model from scratch. He suggested that adapting existing open-weight models could provide a more practical route, allowing the government to retain control over its data while benefiting from established AI capabilities.
His proposal points to an alternative approach centred on targeted public-sector applications, potentially reducing the initial financial and infrastructure requirements while allowing the government to build expertise and assess results before expanding its ambitions.
The debate raises broader questions about the technical assumptions, financing requirements, energy needs and implementation priorities underpinning Ghana’s AI strategy. Clarification of the disputed targets, alongside independent assessments of the proposed infrastructure and costs, would help establish whether the country’s ambitions are technically and financially achievable.
