September 2025 Insights

AI Capex Spend

Artificial Intelligence is driving one of the largest capex cycles in history, as the world’s biggest tech companies are investing hundreds of billions into new data centres, chips, and power infrastructure. McKinsey estimates that by 2030, nearly $7 trillion will need to be spent worldwide to meet AI’s insatiable demand for compute power. For now, this spending is not just building servers; it is powering the economy. Barron’s calculates that AI-related capex contributed more than a full percentage point to US GDP growth in the second quarter alone, while Deutsche Bank noted that AI machines “appear to be saving the US economy right now.”

However, critics warn that today’s growth comes more from building the infrastructure of AI than from the products themselves. Bain & Company calculates that AI needs to drive $2 trillion in annual revenue by 2030 to profitably fund the investment and estimates that the current gap is around ~$800 billion. That shortfall, plus grid constraints and supply chain bottlenecks, raises the risk of a classic boom-and-bust capex cycle.

Still, the opportunity is undeniable. Goldman Sachs forecasts that AI could lift US productivity growth by 1.5% cumulatively in the years ahead. Meta CEO Mark Zuckerberg believes that the real risk is actually underspending, saying that they cannot afford to be behind in what he thinks will enable the most “innovation and value creation [in] history.” Alphabet CEO Sundar Pichai went further, predicting that AI “will have a more profound impact on humanity than fire, electricity, and the internet.”

At High Street, we believe AI adoption will not be a smooth or linear process, but that the structural demand for compute is undeniable. The clearest winners remain the enablers of this buildout, companies supplying the critical infrastructure rather than unproven end-use applications. Nvidia, whose Data Centre business brought in $115bn in revenue in the past 12 months, sits at the heart of nearly every AI deployment. Alongside it, cloud providers like Microsoft are seeing unprecedented demand, as companies like Barclays and UBS have moved from pilot projects to company-wide deployments of Microsoft’s AI tool 365 Copilot. By focusing on such enablers, we aim to capture the durable growth of this transformational trend, while balancing the risks of near-term overinvestment.

Unless otherwise stated, all performance and statistical figures provided in this article have been pulled from Bloomberg by the High Street Asset Management Research Team on 29 September 2025 and all the images provided in this article have been sourced from FreePik and have been used in line with their Acceptable Use Policy. The contents of our newsletters are frequently sourced from or verified through our various product providers and other third parties. Although every effort is made to ensure the accuracy of the information contained in the newsletter, it should not be construed as financial advice as defined in the Financial Advisory and Intermediary Services Act. Links are provided to third-party websites for convenience only. High Street Asset Management (Pty) Ltd cannot accept responsibility and does not endorse any information contained on a third-party site. For our full disclaimer, please see: https://hsam.co.za/legal/.