Why colocation should be top-of-mind for African trading firms in the AI Era
Trading floors across the globe once buzzed with an almost palpable energy – a chaotic ballet of traders shouting orders, phones ringing off the hook, and papers rustling in frantic haste. From New York to Johannesburg's trading pits, the scenes of frenzied activity painted the very essence of market economics in action.
Every trade necessitated meticulous verification, and access to market data was significantly slower, with traders scribbling down details on slips of paper and brokers manually reconciling these transactions at the end of each day – a labour-intensive process that was not only prone to errors but also limited by the capabilities of those involved. The inherent delays, sometimes lasting from minutes to hours, underscored a system ripe for evolution.
The landscape of trading has since undergone a seismic shift, ushered in by the digital era of high-frequency trading (HFT). Today, trading firms harness sophisticated algorithms capable of executing orders in mere fractions of a second, capitalising on minuscule price discrepancies that may appear across different markets and exploiting these fleeting opportunities faster than any human could.
In response, stock exchanges have significantly upgraded their technological infrastructures, offering new services that support the ultra-fast trading activities which now dominate the markets. This includes enhanced data feeds, improved latency in transactions and more robust security measures to safeguard against the complexities introduced by algorithmic trading, to name a few.
Yet, the financial landscape stands on the cusp of another monumental advancement: the integration of third-generation Artificial Intelligence within the capital market environment. AI is being used to accelerate trade executions and refine strategies by analysing extensive datasets, from breaking news to in-depth financial reports, enabling trading firms to respond swiftly and accurately to shifting market dynamics and enhancing both the speed and sophistication of their operations.
What this means for Africa’s trading firms
To fully leverage AI’s capabilities, the integrity of underlying data is crucial – and requires accurate, real-time and low-latency information. This is especially critical for African trading firms who want to compete effectively on the global stage.
For them, colocation emerges as a vital enabler.
Colocation refers to a practice where trading firms rent a space in a data centre to house their computer servers in the same physical location as an exchange’s own systems. This proximity drastically reduces the time it takes for data to travel between servers, eliminating geographical disadvantages and equalising the playing field between clients regardless of their physical location, while minimising latency and enabling transactions to occur almost instantaneously.
To illustrate just how swiftly colocation facilitates transactions, consider this: the blink of an eye takes about 300 milliseconds, roughly one-third of a second. In contrast, the Johannesburg Stock Exchange’s (JSE’s) colocation data centre, for example, has an average round-trip latency of under 20 microseconds. This makes the transfer of data 15,000 times faster than a human blink.
At such velocities, the difference between profit and loss – or competitive edge and obsolescence – is measured in microseconds and in some instances, nanoseconds.
Colocation in the AI era
As AI becomes more embedded in capital markets, colocation will offer trading firms unparalleled speed and efficiency. By housing servers directly within exchange facilities, these firms benefit from reduced data transmission times, allowing algorithms to operate at peak performance and capitalise on real-time market data. Moreover, colocation facilities come equipped with robust infrastructure ensuring reliability and enhanced security. High uptime rates are standard, minimising potential disruptions and maintaining a steady trading presence.
Security protocols are also rigorously enforced to shield sensitive data from cyber threats, while scalable architectures enable firms to increase their computational capacity seamlessly as their trading volumes grow.
Traditionally, colocation was a costly endeavour, often accessible only to the top-tier trading firms. However, recognising the competitive disparity this can create, exchanges around the world are now innovating ways to democratise the service.
For instance, at the close of 2023, the JSE unveiled Colo 2.0, a reimagined iteration of its colocation data centre. This initiative marks a significant shift as it introduces an Infrastructure as a Service (IaaS) model, offering pre-assembled infrastructure tailored to varying client needs. From basic entry-level virtual machines to advanced high-spec servers, Colo 2.0 is designed to streamline the technical and financial barriers typically associated with setting up and maintaining high-frequency trading operations, making colocation more accessible to a wider range of clients.
As African trading firms look to carve out a more pronounced role on the global stage, the strategic adoption of technologies like colocation becomes imperative. This will not only enhance their current operations but also prepare them for future advancements as AI continues to evolve, ensuring they remain competitive in an increasingly digital trading arena.