Dr Andile Ngcaba, executive chairman and co-founder of Convergence Partners and Solcon Capital. (Photograph by Strike a Pose)
Artificial intelligence (AI) agents could outnumber humans operating on the internet within the next two to three years, as rapid innovation forces businesses and governments to rethink how they approach the emerging technology.
This is according to Dr Andile Ngcaba, executive chairman and co-founder of Convergence Partners and Solcon Capital, who recently spoke at Digital Council Africa’s Conext 2026 annual event.
Ngcaba, who is also chairman of Digital Council Africa delivered a presentation on developments in the AI industry, noting the world is entering unfamiliar territory as advances in computing power, AI models and autonomous agents accelerate.
“There is a view that I strongly believe in: agents will supersede humans on the internet before 2030,” noted Ngcaba.
“The number of living humans will be a fraction of the number of agents operating on the internet within the next two to three years. The scale and speed of these developments make it increasingly difficult to predict what the technology landscape will look like in the coming years.”
See also
The global AI landscape is evolving rapidly as companies race to develop more advanced models, deploy AI agents and expand the computing infrastructure needed to support them.
AI agents are increasingly advancing beyond simple chatbots, with the ability to use tools, execute multi-step tasks and interact with digital systems with varying levels of human supervision.
These developments have raised concerns about rogue or misbehaving agents that may act outside their intended parameters, exploit system vulnerabilities, or pursue unintended outcomes when given excessive access or poorly-defined instructions.
Ngcaba stated it is difficult to forecast what the AI landscape will look like in 2027, underlining the uncertainty facing governments, businesses and technology developers.
Ngcaba explained the intensifying AI competition is shortening innovation cycles, forcing organisations to adapt faster, while addressing the challenges of governance, security and responsible AI deployment.
“Businesses had previously planned around product launches separated by years or months, whereas changes in AI are now happening over days and weeks. This creates pressure on organisations to reconsider how they develop, deploy and govern technology.”
Ngcaba described a technology stack built around computing infrastructure, AI models and agents, with governance and science also playing important roles. He pointed to the development of AI factories equipped with graphics processing units (GPUs) to support the computing demands of these systems.
He also highlighted the growth of tools and protocols that allow AI systems to connect with services and other capabilities.
“There are three layers in this, what I call the velocity stack: one at the compute level, at models and agents, and also what I call governance and science sits above it clearly. People are building these AI factories, with GPU compute layers, with all of these. This is the world we are living in.
“Awareness of Model Context Protocol (MCP) was previously limited, but the ecosystem has since expanded, with marketplaces and large numbers of MCPs emerging alongside agents that operate through them.”
The infrastructure supporting AI is also changing, he continued. While traditional hyperscale cloud providers remain part of the market, Ngcaba pointed to specialised AI cloud companies offering access to high-end GPUs, GPU-as-a-service and inference-as-a-service.
“The emergence of these providers reflects a shift towards more specialised computing services as organisations sought the resources needed to develop and run AI applications.”
Ngcaba said the rapid development of AI raises questions about how Africa could position itself in a global technology race increasingly shaped by the US and China.
He warned against African governments simply adopting regulatory approaches developed elsewhere, arguing that the speed of change required a different response. He was particularly critical of Europe’s tendency to regulate technologies before fully developing them, suggesting that frameworks could be more adaptable than rules written into law.
African authorities should consider how to support people and businesses building AI technologies, rather than focus primarily on regulation. He acknowledged, however, that AI safety remains a critical concern.
The potential misuse of AI, including its weaponisation, is another risk, he noted. “AI is being used in ways that worry some of us, including its weaponisation. We do not know how these issues can be addressed.
”Some argue there should be treaties similar to those on nuclear weapons, negotiated through institutions such as the United Nations. I am not sure how that would work, but the reality is that we are seeing this happening.
“At the same time, there are possible benefits in fields such as life sciences, where AI-powered virtual cell models could contribute to the discovery of treatments and help shorten drug development and trial processes.”
For Africa, the challenge is to respond to both the risks and opportunities while the technology continues to evolve, he concluded.