How we are thinking about AI

AI is reshaping entire industries, changing how companies are spending money and challenging traditional approaches to valuing businesses.

The remarkable rise of AI‑related companies over the past few years has made one thing clear: for fund managers, it is crucial to understand where this technology is heading and how its impacts – positive or negative – will ripple across the economy and society.

With increasing resources directed toward AI and a constantly shifting narrative around its applications and impact, below we explore the key areas where change is emerging, the signals we look for when making investment decisions, and the ways governments are beginning to address the implications of this new technological frontier.

In many ways, the resilience of the US labour market in the years following the pandemic underpinned economic growth, boosted consumer spending and played a significant role in the rise in equity prices.

Looking ahead, however, the labour market is poised to undergo some structural change as advances in AI and automation reshape job functions – particularly at the early career level. Tasks such as data processing, routine analysis and administrative support, once the foundation of many junior roles, are increasingly being automated – at least to some degree. This shift raises questions about how the next generation of workers will enter and progress through the labour force.

Recent labour‑market data already highlights some pressure points. SignalFire, a venture capital firm that tracks US workforce trends, reports that tech firms reduced graduate hiring by 25% in 2024 compared with the previous year. At the same time, the unemployment rate for recent college graduates rose to 5.7% in the fourth quarter of 2025, up from 5.3% the prior quarter, according to the Federal Reserve Bank of New York. Historically, the unemployment rate for recent college graduates has been lower than the overall unemployment rate.

Despite the weakening data for younger workers and recent graduates, key labour-market indicators we monitor have yet to signal a meaningful deterioration in conditions for the employment market as a whole. From a historical standpoint, the US unemployment rate remains low, and the economy continues to add jobs on a monthly basis (note: the declines recorded in late 2025 and early 2026 appear to be distorted by temporary factors, including adverse weather events and a large-scale labour strike). Consistent with this view, jobless claims – the number of individuals filing for unemployment benefits – have not shown a sustained or material increase.

Zeroing in on the technology sector, we believe the recent moderation in hiring reflects, in part, a correction following the covid-era hiring surge, rather than a broader deterioration in labour demand. Furthermore, when you step back, the tech sector made up just 5-6% of the total US labour market in 2025, according to the Computing Technology Industry Association (CompTIA), an American trade association, that issues certifications for companies that are recognised for validating IT skills.

It’s important to remember that, historically, labour markets have evolved alongside technological progress, and workers have consistently adapted with them. As we see it, the near‑term outlook for AI is no different: most disruption will come not from AI replacing people outright, but from reshaping how work is done. As the IBM Institute noted in 2023, “AI won’t replace people – but people who use AI will replace people who don’t,” a framing that captures the likely shift toward augmentation rather than displacement.

One significant consequence of the meteoric rise in AI‑related stocks has been a reshaping of major US equity benchmarks – most notably the S&P 500. The ’Magnificent Seven’ mega‑cap companies (Apple, Microsoft, NVIDIA, Alphabet, Amazon, Meta, and Tesla), many of which sit at the centre of the AI revolution, now make up roughly one‑third of the index’s total market capitalisation.

Because benchmarks like the S&P 500 are market‑cap weighted, such heavy concentration means that the performance of just a handful of companies can dominate the index’s overall direction. This dynamic can obscure market breadth, which measures how many companies are participating in a broader market move.

When market breadth is strong, a large number of stocks across sectors are rising together – an encouraging sign of broad‑based economic and earnings strength. However, when breadth is narrow, gains are being driven by only a small cluster of large companies. In this scenario, the headline performance of an index can give a misleading impression of underlying market health.

As a result, the S&P 500 may appear to be performing well on the surface, while the majority of its constituents are lagging. This divergence raises the risk that a market rally is masking underlying fragility, since the index becomes more vulnerable to any weakness in the few companies responsible for most of its gains.

The global software sector has been under pressure in early 2026 as rapid advances in AI have raised questions about the durability of traditional SaaS (Software‑as‑a‑Service) business models. In particular, the growing perception that AI can increasingly automate end‑to‑end workflows has challenged assumptions around recurring subscription revenues and long‑term pricing power.

These concerns intensified following the release of Anthropic’s Claude Cowork platform. For some investors, Cowork marked a shift from ’AI assistance’ toward ’AI delegation’, with certain plugins potentially reducing the need for dedicated software subscriptions across parts of the work cycle. While the long‑term commercial implications remain uncertain, the announcement weighed heavily on sentiment toward the sector.

As a result, the Software & Services Index declined by approximately 25% over the first two months of 2026. The index includes several large‑cap software companies such as Microsoft, Adobe, Salesforce, Workday and ServiceNow.

At a portfolio level, our International Share Fund delivered positive relative performance over this period, reflecting an existing underweight exposure to the software sector. At a company level, this includes underweight positions in Microsoft, Oracle, Palantir and Salesforce.

Hyperscalers are facing scrutiny that the ability to monetise their massive AI investments is becoming more challenging as the cost of advanced infrastructure – specialised GPUs, power‑hungry data centres, and intensive training workloads – continues to rise faster than the revenue these models currently generate. This creates a widening gap between capex commitments and commercial returns, raising the risk that AI capacity expands faster than real, profitable adoption. 

If deployment and monetisation fail to keep pace, hyperscalers could find themselves in an overbuild cycle reminiscent of the 1990s fibre‑optic boom, when telecom companies deployed far more cable than demand required. The result was a glut of underutilised assets that eventually triggered a collapse in valuations and pushed major players such as WorldCom into bankruptcy. 

Regardless of how the AI race ultimately plays out, dispersion among participants is inevitable – some will emerge as winners, others will not. Investors diversified across multiple stages of the AI value chain are therefore likely to be best positioned. As a active manager, this creates an opportunity for ANZ Investments to capitalise on structural change in the sector without relying on a single, concentrated bet on how the future of AI unfolds.

AI has the potential to deliver major economic and social gains, yet its risks are growing just as quickly. With the rapid evolution of AI tools and large language models (LLMs), effective regulation is becoming essential.

So far, global regulation has been piecemeal. The European Union (EU) has taken the lead with its AI Act, passed in March 2024, which classifies AI applications by risk – from unacceptable to minimal – and sets corresponding obligations.

The United States is taking a fragmented and increasingly contested approach to AI regulation, shaped by contrasting federal and state priorities. At the federal level, the Trump Administration has adopted a strongly deregulatory stance, revoking the Biden Administration’s earlier safety‑focused framework.

However, there are growing signs that lawmakers across the political spectrum are taking a more active role in shaping AI regulation, particularly when it comes to preventing government misuse of AI, strengthening safety oversight, and addressing broad social impacts such as employment and civil liberties. These areas represent meaningful points of overlap between progressive and conservative factions, driven by shared concerns about transparency, accountability, and public protection. Recent bipartisan analyses and public opinion data reinforce this trend, showing that Americans widely support targeted forms of AI oversight that balance innovation with safeguards.

New Zealand is taking a flexible, guidance‑driven approach to AI regulation rather than creating sweeping new laws. In 2025, the Ministry of Business, Innovation & Employment (MBIE) released New Zealand’s Strategy for Artificial Intelligence: Investing with Confidence. The framework seeks to accelerate private sector AI-adoption to support economic growth, while safeguarding against basic rights. Overall, the MBIE argues New Zealand should be a ‘fast follower’ aiming to benefit from global progress without overregulating.

At ANZ Investments, we believe disciplined, in‑depth analysis is critical to navigating this fast‑moving and transformational technology. Our focus is on identifying genuine AI‑driven opportunities, avoiding market hype, and building resilient portfolios capable of withstanding what is likely to remain a dynamic and sometimes volatile theme in the years ahead.

As an active manager, we expect the rapid evolution of AI to create market inefficiencies as the industry matures. These shifts provide opportunities to invest in businesses with real, durable value rather than speculative momentum. Our global partnerships remain central to our approach. By leveraging their expertise, we can continue to improve the outcome for our investors over the long term.