For retailers and digital platforms, these are some of the most commercially significant days of the year. What bargain-hunting consumers don’t see is the behind-the-scenes AI and infrastructure that’s responsible for keeping these experiences running smoothly, or not. 

“At Africa Data Centres, we regularly see how moments of extreme demand test the limits of the digital infrastructure supporting modern retail platforms and AI-driven services,” Adil El Youssefi, CEO at Africa Data Centres, said.

Every step of the digital shopping experience is carefully curated, with search engines pushing relevant products, recommendation systems personalising offers in real time, pricing algorithms adjusting dynamically, fraud detection models screening transactions instantly, and AI-powered customer service tools handling thousands of queries simultaneously. 

However, the quest to improve customer experience and drive revenue introduces a new infrastructure challenge. When millions of users arrive at once, AI workloads spike, placing immense pressure on the infrastructure supporting them. 

Without the right foundation, even the most sophisticated digital platforms can struggle to cope. The same goes for the infrastructure that houses them.

The infrastructure stress of peak demand

Peak demand creates a very different operational environment compared with normal daily workloads. 

During major retail events, e-commerce platforms frequently report traffic spikes of three to five times normal levels. Data from Capitec, for instance, shows that South Africa’s Black Friday weekend alone generated 55 million transactions valued at R25.3 billion. And that’s just one institution.

Each additional customer interaction triggers several AI processes in the background, which, when multiplied across millions of interactions, generate significant compute demand. 

Research from the International Energy Agency suggests that AI-driven data centre workloads are already contributing to rapid increases in power consumption globally, with demand expected to grow sharply as AI adoption accelerates.

For infrastructure operators, this translates into higher power needs, greater heat output, and dramatically increased data movement across networks during peak periods. If these are lacking, performance begins to degrade: latency increases, applications slow down, or services simply become unavailable. 

For businesses, that can translate directly into millions in lost revenue, while slow or unreliable platforms during major promotions can quickly erode trust and drive shoppers elsewhere.

Designing for performance under load

Maintaining consistent AI performance during peak demand, without sacrificing reliability, needs infrastructure designed specifically for these conditions. 

From ADC’s experience supporting high-performance digital environments, four elements consistently prove critical:

1. High-density compute capacity; the powerful GPU clusters capable of processing large volumes of requests simultaneously without bottlenecks. 
2. Power systems engineered for stability under sudden increases in demand, incorporating redundant power distribution and surge-handling capabilities.
3. Advanced cooling systems to deal with the significant rise in heat generation as servers operate at high volumes, maintain stable operating temperatures, and prevent thermal throttling, where hardware automatically reduces performance to avoid overheating.
4. High-bandwidth, low-latency data pipelines capable of moving vast amounts of information quickly and reliably to ensure that recommendation engines, search platforms, and pricing models respond instantly, even when demand surges.

Peak performance is an infrastructure challenge

With digital commerce expanding, high-demand events will only grow more intense, transforming peak performance from a software or cloud-scaling challenge into an infrastructure challenge.

Businesses are quickly recognising that the reliability of their AI infrastructure during peak demand times depends on the physical infrastructure supporting them. When infrastructure is designed to handle these demands, AI systems can deliver the consistent performance customers expect when it matters most.

In the digital economy, this is often the difference between revenue growth and missed opportunity, which is why infrastructure readiness should form part of AI strategy discussions at board level.

The post When AI Infrastructure Is Tested by Peak Demand appeared first on Tech | Business | Economy.

The fresh capital lifts the firm’s total funding to $16 million and enhances its real-time discovery technology deeper into global retail, marketplaces, and travel platforms.

The round was led by MMC Ventures with backing from Redalpine, Daphni and a group of angels. The company’s system already handles billions of live interactions each month and delivers tens of millions of predictions for partners across several industries.

The startup was created in 2024 by Dr Kevin Kahn and Dr Matteo Ruffini, both former Amazon AI leaders, alongside entrepreneur Johan Boissard. They argue that the growth of generative AI has overshadowed the need to understand how people behave, browse, and change their minds in real time.

Most recommendation engines still work off batch-trained models and recycled data. They guess based on what users liked yesterday. Albatross takes a different route. 

Its transformer-based architecture reads live activity, learns from it instantly, and adjusts results within milliseconds, without any manual retraining. According to the company, no previous system has been able to adapt at this speed.

Speaking on the company’s mission, Dr Kahn, co-founder and CEO, said: “Our mission is to close the discovery gap. Everyone has felt the frustration of seeing the same generic recommendations over and over. Our system perceives and adapts instantly, so every search and feed reflects the user’s intent at that very moment. 

“From our conversations with a broad range of CEOs, we know they all want to adopt AI, but most efforts fail because they treat it as an add-on. The real opportunity is to rethink how experiences are shaped, to make every interaction intelligent, adaptive, and alive. That’s what Albatross does: it learns and reasons in real time, understanding intent the moment it happens.”

The company’s flagship tools, the Real-Time Discovery Feed and the Multimodal Search engine, are designed to enhance what users see as their behaviour shifts. The system can connect online and in-store journeys, react to image inputs, and run at near-zero latency.

Early trials have delivered strong results, with some partners recording triple-digit growth in engagement and product discovery. Deployment takes less than seven weeks, and the platform now powers what the team presented at RecSys 2025 as the industry’s most advanced approach to cold-start discovery.

For investors, the timing shows a change in how personalisation technology is evolving. Mina Samaan, General Partner at MMC Ventures, said: “Personalisation is entering a new era. Albatross moves beyond static algorithms to build systems that understand context as it happens. Their real-time architecture represents a step change for how businesses engage customers online.”

Redalpine, which led the startup’s first round, also increased its commitment. Dr Marc Moesser, Investor at Redalpine, said: “We backed Albatross from the very beginning because of the team’s exceptional depth in AI and personalisation. In just a year, they’ve gone beyond what even the largest platforms achieve with real-time infrastructure that adapts instantly to user behaviour.”

Albatross seeks to build systems that react as quickly as the people using them, and remake the online discovery experience in the process.

The post Albatross: Former Amazon AI Execs Raise $12.5m to Launch Real-Time User Intent Platform appeared first on Tech | Business | Economy.

The round was led by Mercy Corps Ventures and Chui Ventures, with participation from investors such as Digital Currency Group, SOSV, Resilience17, growX ventures, Orbit Startups, Reflect Ventures, and angel investors/advisors from global tech giants like Google, Salesforce, and Zendesk.

Founded in 2022 by Kenfield Griffith (CEO) and Louis Majanja, tappi addresses the challenges faced by SMEs, which contribute significantly to Africa’s GDP and employment. The platform focuses on digitizing these businesses by streamlining the creation and management of online profiles, providing a hassle-free way for SMEs to showcase products, interact with customers, and facilitate payments.

The funds secured in this round will be used for tappi’s expansion, concentrating on talent acquisition, brand building, and solidifying its presence in current markets. SMEs often struggle with the digital transition, and tappi aims to bridge this gap by offering an intuitive platform that allows businesses to establish an online presence within minutes.

One of tappi’s standout features is its ability to generate SEO-optimized websites within two minutes, significantly enhancing the visibility of small businesses. The platform has successfully indexed thousands of business pages on Google, amplifying the online presence of these enterprises.

To tackle payment hurdles faced by SMEs, tappi has partnered with mobile network operators, currently working with MTN, to facilitate ad purchases using airtime. This addresses a common challenge where businesses encounter obstacles in paying for ads with traditional credit cards.

Kenfield Griffith emphasized tappi’s commitment to helping businesses achieve visibility, stating, “Our goal is to help businesses achieve visibility.” He added that tappi’s AI feature aids businesses in crafting effective ad copies, an important aspect often overlooked by resource-constrained SMEs.

While not yet profitable, tappi is on a trajectory toward profitability, buoyed by a growing subscriber base with plans ranging from $2 to $100. Griffith highlighted the startup’s substantial growth in business ads and data bundle subscriptions, with a notable surge in subscribers within the $2–$19 range.

Expressing gratitude for the support from investors, Griffith noted tappi’s mission to tap into the untapped potential within Africa’s informal SME markets. The startup aims to empower businesses in overlooked service industries such as food services, fashion, agriculture, health, and beauty, providing them with a trusted online identity to connect with customers.

The post Digital Commerce Platform, tappi, Secures $1.5 Million in Pre-Seed Funding appeared first on Tech | Business | Economy.