Maldistribution of AI multiplies the global tech divide

[[{“value”:”

In the world of cloud computing, a silent revolution is underway, but not the kind that propels innovation. Instead, it’s a stranglehold tightening its grip on organizations and the global technology industry as a whole, creating counter-innovation, and crippling technological progress for some, while advancing it for a handful. The cloud hyperscale oligopoly casts a shadow over the entire industry, impeding global technology innovation, without much awareness of the direct victims, let alone the indirect victims – think of civilians beyond the borders of the US tech hubs. As AI is about to propel worldwide innovation, the stakes are up.

The allure of seamless and excellent services promised and delivered by the leading cloud service providers (CSPs) masks a silent issue. The convenience, expertise, integrated services, and familiarity provided by a single cloud provider has created a huge dependency issue. This trend has become industry best practice since the mid-2010s, with the growing success of Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP). Cloud services operate smoothly within one provider, but there are no universal standards forced on the three US-based Cloud Hyperscalers, hindering mobility and integration between clouds. Businesses therefore have no choice than to surrender their digital futures to a single cloud provider, and somehow governments and investors have accepted that.

The ongoing antitrust trial – U.S. et al. v. Google – is one example where the US government challenges the monopoly of one provider, but this is an isolated case. The fact is that 80% of businesses are locked in to services from one of the top three cloud providers. The underlying dependency, cloaked by comfort and convenience, concentrates power and control, making it challenging for companies to envision a digital future beyond these cloud confines.

Overwhelming trust in the Cloud partners, the lack of resistance while the dependency is so immense, is potentially more alarming than the lock-in itself and could be described as mass Stockholm Syndrome playing out. While that is the case, resolve is not coming from the industry itself.

AI can turn the wildfire into an inferno

The advancement and potential of AI intensifies the problem by magnitudes. At the heart of this issue lies the artificial inflation of GPU prices, dictated by the dominance of the hyper-scale cloud service providers. This pricing power that sits with the ones in charge isn’t merely an inconvenience – it’s stifling innovation and stifling it most outside of the US tech capitals.

As the majority of GPU chips, crucial for advanced AI applications, find their way into the hands of the hyper-scalers, the consequences ripple through the business landscape. Startups, particularly those venturing into AI, face an uphill battle in presenting viable business cases due to exorbitant GPU-related costs. Even as businesses relentlessly cut costs elsewhere, the price of cloud computing continues to soar. It’s not merely an expense; it’s a lifeline for businesses, making it impossible to simply switch off and explore alternative options.

The centralized control over critical resources hampers global innovation, creating a scenario where tech hubs in regions beyond the West Coast and East Coast of the US are left behind, let alone the rest of the world, not to even think about the development regions in the world.

The disparity is glaring – while cities like Amsterdam and London might feel the pinch, accessing GPUs becomes an insurmountable challenge in Eastern Europe, Italy, Spain. This inequality in resource access inadvertently shifts the focus of global innovation towards issues that matter most in regions that benefit the hyper-scalers, leaving the rest of the world to grapple with its unique challenges, hugely growing the gap between the digital and economical ‘haves’ and ‘have nots’.

The impact of this GPU scarcity is not confined to the business world. In highly regulated industries like healthcare, where the need for AI-driven innovations is paramount, the limitations of GPU access present a tangible challenge. The difficulty in complying with regulatory requirements forces organizations to resort to physical GPUs deployed in local data centers, reintroducing operational risks that cloud computing sought to eliminate.

The stark reality is that this GPU scarcity is a crisis with far-reaching consequences. As businesses struggle to secure the resources needed for AI innovation, the risk of a global business and humanitarian disaster looms large.

Averting a looming business and humanitarian crisis

It’s high time industry players, governments and investors understand their responsibility and act quickly to avoid the global crisis caused by the global cloud lock-in leaking into the AI forcefield.

Governments should take proactive measures to ensure equal distribution of GPU, foster competition within the cloud computing and AI sectors, and force all Cloud Providers to support Cloud-interoperability standards. By promoting regulatory frameworks that encourage open market practices, they can create an environment where there is a good amount of availability and choice for buyers and where suppliers compete without creating lock-in, breaking up the monopolistic control by a few Cloud offerings and worse; AI offerings.

The importance of clear guidelines on data ownership – sovereignty -, transfer, and protection has become clear to governments and companies since the start of the global cloud transition. Yet even data portability is barely solved for. With the risk of AI becoming ‘an added cloud service’ it is even more crucial that using a particular cloud for applications still allows freedom in choosing where to store data, how to manage compute resources, and how to implement specific use cases, including those involving Large Language Models (LLMs). Collaborative efforts between industry stakeholders to develop open standards can reduce dependency on proprietary technologies, promoting interoperability and choice.

Digital sovereignty initiatives have gained in importance for ensuring that nations have control over their critical digital infrastructure and data. Governments can support the development of national or regional cloud initiatives, providing secure and competitive alternatives to reduce reliance on a small number of global providers.

And businesses themselves can proactively adopt practices that align with open cloud standards. Betting your company on an infrastructure you don’t own is extremely risky, which is why adding an abstraction layer with global and neutral standards between yourselves and the underlying infrastructure is imperative. Supporting the development of alternative cloud providers, participating in the establishment of open standards, and embracing digital sovereignty initiatives are ways businesses can contribute to a more equitable and competitive environment.

The AI revolution has begun and could contribute up to $15.7 trillion to the global economy by 2030, with huge consequences for society as a whole. Today marks the historical moment where the world either accepts and deepens existing segregation putting more power in the hands of specific leaders, or this is the historical moment in time when governments and the industry address this issue head-on, limiting the unlawful reign of silent powers.

We’ve featured the best cloud management software.

This article was produced as part of TechRadarPro’s Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

“}]]