Five years ago, such a concentration of value would have been difficult to envision. Today, it reflects a deeper transformation in how value is created, captured and sustained in the global economy and equity market. At the centre of this lies the transformation of AI.
Big just got bigger
Companies like Nvidia, now exceeding US$5 trillion in market capitalisation, are not merely large — they are central connections in this new economic architecture.
Its dominance reflects control over a critical bottleneck: high-performance compute. This centrality is key to understanding why scale has not become a constraint but rather an advantage in the current cycle.
This exclusive club of fifteen companies with market capitalisation of more than US$1 trillion includes Nvidia (which topped the list), Apple, Alphabet, Microsoft, Amazon, Broadcom, Taiwan Semiconductor Manufacturing Company (TSMC), Meta Platforms, Samsung Electronics, Micron Technology and SK Hynix. AI-related companies accounted for eleven of these fifteen companies.
See also: Is the AI of Things sustainable?
To contextualise the magnitude, Nvidia’s market capitalisation exceeds Germany’s GDP, the world’s third-largest economy. This highlights the sheer scale and critical importance of this sector to the global economy. It is the single largest investment theme, attracting the bulk of current investments and funding.
The key question is not whether AI will continue to drive growth, because it clearly will, but whether the scale and valuation of these companies can continue expanding.
Historically, large companies face diminishing returns, regulatory scrutiny and competitive erosion. However, the AI cycle exhibits characteristics that challenge this framework.
See also: Asian stocks rally as Trump signals Iran deal, oil slips
Scale begets scale
Unlike previous technology waves, AI is defined by increasing returns to scale. Larger datasets improve model performance, greater compute accelerates capability, and wider deployment creates feedback loops that reinforce incumbency. This creates a compounding dynamic in which scale begets more scale.
This dynamic is reinforced by the vertical integration of the AI stack. The leading players increasingly control multiple layers of the ecosystem: semiconductor design, chip fabrication, equipment manufacturing, cloud infrastructure and application-level deployment.
Nvidia dominates compute, TSMC controls advanced manufacturing nodes, and ASML represents a near-total monopoly in extreme ultraviolet lithography. On top of this infrastructure sit hyperscalers such as Microsoft and Amazon, which distribute AI capabilities via cloud platforms, while companies like Alphabet and Meta control user interfaces and data ecosystems. This vertical integration compresses margins in peripheral segments while concentrating economic value at the top of the stack.
Capital intensity further entrenches this concentration. The AI arms race requires unprecedented levels of investment, with hyperscalers committing tens of billions annually to data centre expansion and semiconductor fabrication plants costing upwards of US$20–US$40 billion each. These costs are prohibitive for new entrants, effectively limiting meaningful competition to a handful of players with sufficient financial and technical resources.
ASML provides a textbook example: its Extreme Ultraviolet (EUV) machines are indispensable for producing advanced chips. Through decades of accumulated expertise and a highly complex supply chain, this makes replication virtually impossible. It has a complete and total monopoly of the EUV market. Control of such bottlenecks translates directly into pricing power and long-term earnings visibility.
Policy dynamics are reinforcing, rather than disrupting, this concentration. AI has become a strategic priority for governments globally, leading to industrial policies that support domestic champions. Subsidies for semiconductor manufacturing, restrictions on technology exports and national AI strategies all serve to deepen existing competitive moats.
While this introduces risks, it also raises barriers to entry and protects incumbents within key markets. In effect, policy is acting as both a shield and a catalyst for the largest players.
Given these structural advantages, the argument that “big can get bigger” holds substantial weight.
Sink your teeth into in-depth insights from our contributors, and dive into financial and economic trends
Can expensive stay expensive?
The more pressing question is whether elevated valuations can remain. High-growth companies have often enjoyed higher price-earnings multiples, justifiable because of high earnings potential.
The current AI-driven rally is underpinned by tangible financial performance, as shown from first-quarter 2026 earnings for the S&P 500 companies, which beat market expectations, with AI-linked companies delivering outsized contributions. Importantly, earnings revisions have been positive, suggesting that the market is underestimating the pace of growth.
This distinction is critical. In previous bubbles, valuations expanded ahead of fundamentals, often leading to sharp corrections when expectations were not met. In the current cycle, earnings are rising alongside valuations, providing a degree of support for elevated multiples.
Scarcity also plays a role. Not all AI exposure is created equal, and the market is increasingly differentiating between companies based on their strategic positioning.
Firms that control critical technologies or infrastructure — such as Nvidia, ASML or TSMC — command a premium because they are effectively irreplaceable within the ecosystem. This scarcity translates into pricing power and better margins. In contrast, companies operating in more commoditised segments may struggle to sustain similar valuations.
Second-order winners
Beyond semiconductor AI beneficiaries, second- and third-order beneficiaries have also experienced extraordinary share price gains due to potential spillover benefits.
For example, SanDisk Corp posted an exceptional gain of 623% year-to-date or in less than six months. Micron Technology surged 273% for the same period, while ARM Holdings went up by 268%.
In May alone, several big-cap AI-linked stocks also displayed strong double-digit monthly gains. This included Micron Technology (+88% in May 2026), Murata Manufacturing (+87%) and SK Hynix (+82%).
One group that stood out in particular was cybersecurity. This included CrowdStrike Holdings (+64%), Fortinet, Inc (+64%) and Palo Alto Networks (+57%).
Cybersecurity is a key second-order beneficiary that has emerged from the AI boom. The proliferation of AI systems increases the complexity and vulnerability of digital infrastructure, creating a structural need for advanced security solutions.
As enterprises deploy AI at scale, they face rising risks related to data breaches, model manipulation and system integrity. This will support demand for cybersecurity services and products.
Companies in this sector are benefiting from several favourable characteristics. Cybersecurity spending is largely non-discretionary, particularly in highly regulated industries where compliance requirements are stringent.
Palo Alto Networks exemplifies this trend. The company has delivered strong revenue growth and significant expansion in its next-generation services segment, indicating successful monetisation of AI-enabled offerings.
Its platform strategy, which consolidates multiple security functions into a unified system, reflects a broader industry trend. As AI adoption accelerates, cybersecurity will become an indispensable layer of the technology stack, ensuring its position as a durable growth area.
Following Palo Alto’s results on June 3, there was also a flurry of fair value upgrades. This ranged from a low of US$285 to a high of US$375 with an average 12-month target of US$317 (based on Bloomberg consensus). This compares to the current share price of US$280, which is up 52% year-to-date.
Stay focused on earnings
Despite the positive outlook for the AI sector, several risks could challenge the sustainability of current valuations and market concentration. Geopolitical tensions remain a key concern, particularly in the context of technology export controls and supply chain disruptions.
Companies with significant exposure to a single market could potentially face revenue losses if restrictions tighten further. Supply-side constraints also pose challenges, as shortages of semiconductors, power infrastructure and data centre capacity could limit the pace of growth even in the presence of strong demand.
While adoption is accelerating, monetisation is still evolving. Companies are investing heavily in AI capabilities, but the productivity gains and cost efficiencies are not always immediate.
If these returns fall short of expectations, there is a risk that capital expenditure could slow, potentially impacting demand for AI infrastructure. A high degree of market concentration increases vulnerability to volatility, as a small number of companies account for a disproportionate share of index performance.
Looking ahead, the likely scenario is one in which the AI super-cycle continues, but with greater differentiation among participants. We continue to favour companies with robust earnings, supported by ongoing investment and expanding revenue.
The largest companies are likely to continue to consolidate their positions, benefiting from scale, integration and scarcity. High valuations are likely to persist, though further multiple expansion may be limited. In this case, price upside will likely be driven by earnings growth rather than valuation re-rating.
Potential future drivers will come from faster-than-expected monetisation of AI applications, which could drive broader adoption and unlock new revenue streams.
This would broaden market participation beyond the current leaders. Conversely, a downside scenario could emerge if returns on AI investment disappoint, leading to a slowdown in spending and a subsequent re-rating of the sector.
The AI trade is evolving. The initial phase rewarded broad exposure to the theme, as capital flowed into anything associated with AI. The next phase will require greater precision. Investors should focus on companies that control key bottlenecks and exhibit clear pathways to monetisation.
While the broader AI space will continue to benefit from strong earnings and structural tailwinds, at current prices, investors need to be more selective and gravitate towards companies with clear monetisation pathways, visible earnings growth and defensible competitive advantages.
Carmen Lee is head of equity research at OCBC
