Digital Economy at a crossroads: Navigating the 2026 AI infrastructure boom and software shakeout

Key Points

The rapid rise of generative AI following the debut of ChatGPT in late 2022 has been a major catalyst for the digital economy sector. For much of 2023–2025, companies investing aggressively in AI infrastructure were rewarded by the market, with capital expenditure (CAPEX) announcements interpreted as a signal of technological leadership.

That narrative has now shifted.

Shares of Big Tech giants have faced heavy selloffs due to concerns over the massive scale of AI CAPEX and its immediate impact on free cash flow (FCF). At the same time, the broader software and cybersecurity sectors have seen significant de-ratings, fuelled by fears that AI-native applications could displace established incumbents.

In this article, we bring together our views on three key pillars of the digital economy – Big Tech, Software, and Cybersecurity – to help investors navigate the evolving AI landscape.

Are Big Tech firms overstretching themselves?

The recent sell-off in Big Tech was not triggered by poor earnings, but by the "beat and raise" nature of CAPEX projections. Total CAPEX for Meta, Alphabet, Microsoft, and Amazon is now expected to reach USD 650 billion in 2026, representing a 56% increase from 2025 —a massive leap from our previous estimate of USD 500 billion. Furthermore, spending is increasingly supplemented by debt issuance. For instance, Google recently announced a 100-year bond issue. This AI arms race has fundamentally altered the FCF profile of these giants, moving them from high-margin stability to a period of heavy, front-loaded investment.

Figure 1: Big Tech firms beat and raise CAPEX projections

Investors are rightly concerned about rising depreciation costs and whether the return on investment (ROI) will materialise quickly enough to justify the surge in spending. Such concerns are not unusual during the early stages of major technology cycles. However, historical precedent suggests that periods of aggressive infrastructure investment often appear excessive before clear monetisation emerges. The early build-out of the internet, for example, resulted in a capacity glut before demand and profitability eventually caught up.

Importantly, the CAPEX outlays of hyperscalers today are supported by genuine demand rather than speculative overbuilding. Data centre vacancy rates remain extremely tight at less than 1%, indicating that existing infrastructure is already operating close to full utilisation. Cloud demand backlogs for Microsoft and Alphabet recently grew by more than 100% year-on-year in their latest earnings quarter, suggesting that much of the capacity currently being built is already effectively pre-committed. Given the robust underlying fundamentals of Big Tech firms, we expect this FCF erosion to be temporary and cash flow to recover as future CAPEX eventually moderates.

Moreover, the evolution from simple chatbot interactions to Agentic AI is driving a sharp increase in AI inference demand. These agents execute multi-step reasoning processes, consuming far more tokens than traditional AI queries. As the primary providers of the cloud infrastructure required to run these workloads, Big Tech hyperscalers stand to be the direct beneficiaries of this surge in compute demand.

Overall, we maintain an attractive view on Big Tech, given that earnings visibility remains relatively strong and fundamentals remain healthy.

Table 1: Latest Big Tech earnings and outlook

Related article:  AI: Gift or Disruptor? (Big Tech)

Software faces AI disruption, but opportunities remain

While Big Tech grapples with questions of overspending, the software sector faces a different challenge: the fear of obsolescence.

Software stocks have experienced sharp selloffs this year, driven by fears that AI-native agents could replace traditional Software-as-a-Service (SaaS) models. Investors are particularly concerned about moat erosion, as AI threatens to break down the high barriers to entry and switching costs that have historically protected software profits.

In the past, the need for a massive developer workforce served as a natural barrier to competition. Today, AI coding agents like GitHub Copilot and OpenAI’s Codex allow developers to write code exponentially faster, reducing development friction and enabling new AI-native competitors to enter the market with far less capital than previously required. This intensifies competitive pressure across the sector.

At the same time, AI tools like Claude’s Coworker can autonomously handle complex tasks—from drafting legal documents to managing customer support tickets—at significantly lower marginal cost, threatening the pricing power of premium SaaS subscriptions designed to enhance human productivity.

Despite these structural challenges, the market seems to have overreacted to the risks facing industry leaders. For enterprise customers, switching from established platforms like Salesforce or ServiceNow to standalone AI models remains incredibly difficult. These systems are deeply embedded in corporate workflows, complex data architectures, and strict compliance and governance frameworks that have been customized over many years. Replacing them would require massive data migration, security validation, and regulatory approval, all while risking significant operational disruption.

Rather than being eliminated, the software sector is being reshaped as incumbents “turn foes into friends” by integrating AI into their core platforms. Leading firms are already successfully monetising this shift:

• ServiceNow’s Now Assist has already surpassed USD 600 million in net new annual contract value and is on track for a USD 1 billion run rate by 2026

• The Annual Recurring Revenue (ARR) of Salesforce’s Agentforce saw a 169% year-on-year increase

After the YTD -17% selloff, the software sector (as measured by the iShares Expanded Tech-Software Sector ETF) is currently trading at 22x forward PE, a meaningful discount to its historical average of 38x. With double digit earnings growth expected through 2028, we project an attractive upside potential of approximately 62%.

Nonetheless, we maintain our Neutral view on the software sector, as valuations may remain under pressure amid rising competition and potential margin compression. Investors should therefore remain selective, focusing on large-cap platforms with durable workflow and data moats, as well as vendors that can successfully embed agentic AI into their offerings without materially eroding pricing power.

Table 2: Projections for iShares Expanded Tech-Software Sector ETF

Related article: SaaSpocalypse: Software stocks crash on AI disruption fears – But is the market overreacting?

Cybersecurity remains a critical, resilient sector in the AI era

Within the broader software landscape, cybersecurity stands out as particularly insulated from AI disruption. The technical complexity and "mission-critical" nature of security make it nearly impossible for enterprises to replace established vendors with DIY AI code. In fact, AI is expanding the "attack surface" for hackers, creating a structural need for more security.  As businesses deploy agentic AI, they face new risks like "prompt injection" and data leakage. Adding to this pressure, rising geopolitical tensions, including heightened US-Iran hostilities, have led to state-sponsored attacks targeting critical infrastructure, corporations, and financial systems. Taken together, these factors have reinforced cybersecurity as a strategic imperative, not merely a discretionary IT expense.

That said, while the sector is structurally well-positioned, near-term growth may be moderated by corporate budget constraints. Security spending is expected to expand only modestly in 2026, with 56% of organisations surveyed by UBS Evidence Lab anticipating a mere 1–5% growth, 23% expecting increases of 6% or more, and 18% projecting no change. As budgets come under pressure, investors must be selective and focus on companies with exposure to higher-priority spending areas. According to the UBS survey, Chief Information Security Officers (CISOs) rank cloud security and identity security as their top priorities for additional spending.

Figure 2: Budget expectations by segment

Source: UBS Evidence Lab. Data as of November 2025.

Cloud security is currently the fastest-growing subsegment, with Gartner projecting 30% growth in 2026 as organisations migrate massive AI workloads and data to the cloud. This migration creates an expansive risk surface that requires robust protection.

At the same time, the explosion of machine identities—AI bots and automated systems that now outnumber human identities by 82 to 1—is driving urgent demand for identity solutions capable of governing privileged access and preventing sophisticated spoofing attacks.

Beyond these segments, cybersecurity providers are turning AI from a threat into a powerful tool to address the global shortage of 4.8 million unfilled security roles. AI-powered solutions, such as CrowdStrike’s Charlotte AI, are being used to automate security operations (SecOps) and accelerate threat detection, reinforcing the market position of established leaders rather than undermining them.

Another key trend in the cybersecurity sector is vendor consolidation, or platformisation. Enterprises are moving away from "best-of-breed" tools toward integrated platforms to lower complexity and cost. These platforms can also provide better security outcomes by unifying data across an organisation into a single environment. We favour platform leaders such as Palo Alto Networks and CrowdStrike, which have strong exposure to high-priority areas like cloud and identity security. This positions them well to benefit from the platformisation trend and capture a larger share of constrained IT budgets

For investors seeking a more diversified and stable return profile, a broad-based cybersecurity ETF such as the Global X Cybersecurity ETF (NASDAQ: BUG) is worth considering. BUG holds 30 cybersecurity companies, providing exposure to platform leaders while mitigating the volatility associated with individual stocks. Based on projected 2028 earnings, we estimate a target price of USD 58 for the ETF, implying upside potential of approximately 120% from current levels.

Table 3: Projections for the Global X Cybersecurity ETF

Related article: Cybersecurity leaders poised for big gains after recent sell-off

The digital economy remains an attractive investment theme

Despite recent volatility, the digital economy remains one of the most compelling long-term investment themes in global equities.

Big Tech: We reaffirm our positive view of these giants as their massive capital outlays are underpinned by substantial cloud demand and order backlogs, providing strong long-term revenue visibility.

Software: We maintain a neutral and selective stance. While AI-driven moat erosion poses structural challenges to margins and valuations may remain compressed, the recent sell-off has created opportunities to selectively accumulate high-quality leaders at attractive discounts.

Cybersecurity: We remain positive on this sector over the long term, viewing it as a strategic priority for both companies and governments, and relatively insulated from AI disruption. We particularly favour platform vendors with strong exposure to cloud, identity, and AI security.

In light of the above, we reaffirm our 3.5 Stars “Attractive” Rating for the digital economy internet sector, with target price of USD 72 and a projected upside of approximately 56% for the Invesco Nasdaq Internet ETF (NASDAQ: PNQI).

Table 4: Projections for the Invesco Nasdaq Internet ETF

Figure 3: Share prices are driven by earnings growth in the long run

Table 5: Recommended products for the digital economy (internet)

Declaration:

This research report was prepared with the assistance of artificial intelligence (AI) tools. iFAST Financial Pte Ltd does not rely exclusively on AI for content generation; the content of this report – including all investment theses, ratings, price targets and conclusions – has been independently reviewed and verified by the research analyst(s) to ensure accuracy and professional integrity.

For specific disclosure, at the time of publication of this report, IFPL (via its connected and associated entities) and the analyst who produced this report hold a NIL position in the abovementioned securities.