Key Points
- AI adoption is still in its early phase, with business surveys showing that only 10% of companies report using AI to produce a good or service, leaving plenty of room to grow.
- Nevertheless, the AI narrative continues to be supported by strong corporate earnings linked to AI. Big tech firms have recorded substantial growth in their cloud businesses in recent quarters, accompanied by expanding operating margins.
- While hyperscalers have a sufficient balance sheet strength to support elevated CAPEX into 2026 and 2027, smaller ecosystem participants may face funding constraints. This reinforces our preference for companies with clearer monetisation pathways, diversified revenue streams and proven pricing power.
- We believe that the second wave of AI growth will be led by autonomous agents—AI systems capable of planning, reasoning, and performing complex multi-step tasks independently.
- Our stance emphasises selective exposure to structurally advantaged leaders rather than broad thematic buying, enabling investors to balance upside potential with prudent risk management.
AI will continue to lead the charge
Since OpenAI launched ChatGPT in late 2022, the AI theme has been the main driver of equity returns to this day. Despite concerns about overspending on AI, we believe that 2026 will be another year of AI advancement.
AI adoption is still in its early phase, with business surveys showing that only 10% of companies report using AI to produce a good or service, leaving plenty of room to grow. From a broader view, AI investment is currently around 1% of US GDP. In prior general purpose technology investment cycles (e.g., electricity, railroads, communications), investment peaked at 2%–5% of GDP. Therefore, despite such an impressive run during the AI investment boom, the historical precedent suggests that there’s still room for growth.
Figure 1: Large investment cycles have preceded prior general purpose technology productivity booms
Figure 2: AI adoption still in its infancy
The billion-dollar CAPEX whale – What’s the ROI
Companies continue to aggressively increase their bets on AI, without any signs of slowing down, with the largest tech companies (Microsoft, Meta, Amazon, Google) revising their CAPEX projections up to USD $370 billion for 2025, representing a 75% increase from 2024. We anticipate a further 35% rise in CAPEX in 2026 to USD 500 billion, as they continue investing in AI infrastructure to meet the demand from their customer backlogs. This has recently sparked significant concerns over the Return on Investment (ROI) of such massive CAPEX.
Nevertheless, the AI narrative continues to be supported by strong corporate earnings linked to AI. Big tech firms have recorded substantial growth in their cloud businesses in recent quarters, accompanied by expanding operating margins. These companies have also reported impressive order backlogs, justifying why they have chosen to revise CAPEX projections despite ongoing worries about an AI bubble. We continue to favour these big tech hyperscalers, as we remain positive on future cloud growth amid increased AI-related workloads (more tokens processed) and their ability to capture that growth due to the pricing models used in cloud offerings, which ensure scalability.
Table 1: AI monetisation meeting expectations; Remaining Performance Obligations (RPO) showing strong growth
Figure 3: Total AI tokens consumed per week over the past 1 year (billions)
Debt Binge. Investors are growing increasingly concerned about the level of leverage that Big Tech is taking on to build out its artificial intelligence infrastructure. The recent tug-of-war over AI-related CAPEX has shifted from being funded by healthy free cash flow to being financed through debt issuance — a move reminiscent of corporate behaviour during the Dot-com bubble of the 2000s. Among the five major AI spenders (Meta, Microsoft, Amazon, Oracle, and Alphabet), only Microsoft has not tapped the debt market recently. Hyperscaler debt issuance has surged to more than USD $108 billion this year, compared with an average of USD $28 billion over the past five years, signalling an alarming trend.
While the pace of spending may be unsettling to some, the current balance sheets of these companies could help sustain this level of investment. We expect around 80–90% of their planned capital expenditure to continue being funded by cash flows. Moreover, excluding Oracle, the top hyperscalers’ leverage ratios remain below 1x, indicating their total debt is still less than their earnings. These companies could absorb up to USD $700 billion in additional debt and still be viewed as financially sound, keeping leverage at levels below that of a typical A+ rated firm.
However, something that warrants concern is the expansion of the ecosystem to include companies with weaker balance sheets, such as Oracle and CoreWeave, alongside rising debt. We have also seen an increase in interlocking and circular revenue relationships, which represent a risk that needs to be monitored.
A key watchpoint is the cost of capital. Should credit markets tighten or yields rise, the economics of debt-funded AI CAPEX could shift materially.
While hyperscalers have a sufficient balance sheet strength to support elevated CAPEX into 2026 and 2027, smaller ecosystem participants may face funding constraints. This reinforces our preference for companies with clearer monetisation pathways, diversified revenue streams and proven pricing power.
Figure 4: Big Tech stocks binge on Debt
Figure 5: Balance sheet of hyperscalers remains resilient
More AI opportunities to tap into
The playbook for the past few years has been concentrated on the AI infrastructure space. We believe that the second wave of AI growth will be led by autonomous agents—AI systems capable of planning, reasoning, and performing complex multi-step tasks independently. These agents can operate across various software and databases, delivering complete outcomes rather than simple responses. The final wave involves deploying AI into the physical world through robots, drones, and autonomous systems performing real-world tasks.
Fortunately, the components within our recommended ETF (Invesco Nasdaq Internet ETF PNQI) have lower exposure to US consumers (50% aggregate exposure).
Table 2: Exposure to US
|
Sum of % Total Net Asset |
US Exposure |
|
|
Broadline Retail |
17.7% |
35.6% |
|
Capital Markets |
1.5% |
80.0% |
|
Entertainment |
13.4% |
59.7% |
|
Financial Services |
1.5% |
53.3% |
|
Ground Transportation |
4.7% |
46.8% |
|
Hotels, Restaurants & Leisure |
8.3% |
55.4% |
|
Interactive Media & Services |
18.8% |
51.6% |
|
IT Services |
6.1% |
68.9% |
|
Software |
17.0% |
56.5% |
|
Technology Hardware, Storage & Peripherals |
9.2% |
43.5% |
|
Bloomberg Finance L.P., iFAST compilations. Data as of 30 October 2025. |
||
Entertainment & Media – Advertising revenue for companies in this sector continues to show robust growth, but we expect performance ads (spending money to generate an immediate response) to outperform brand ads (aimed at building brand recognition and trust). Company-wise, Meta’s end-to-end AI-powered ad solutions, including the Advantage+ suite, have seen annual revenue run-rate surpass USD $60 billion, while the number of advertisers using at least one of Meta’s video-generation tools jumped 20% in Q2 2025. Meanwhile, revenue from Google’s advertising business rose 12.6%, well above expectations, with AI Overviews and AI Mode clearly resonating with users. AI-driven mobile app monetisation platform AppLovin reported a 92% y/y jump in quarterly earnings, powered by its AXON 2.0 AI engine.
Figure 6: Ads growth to continue see double digit expansion
Broadline Retail – Among broadline retailers, only Amazon has higher exposure to the US (59% of revenue derived from the US). While the company is likely to face headwinds from slowing consumer activity in the US, we believe the convenience of online shopping will continue to support its e-commerce business. Amazon should also continue to benefit from ongoing fulfilment network innovations, and the recently announced large-scale job cuts should help improve margins. Advancements in AI are also propelling the commerce business forward, with AI assistant Rufus already unlocking USD $10 billion in incremental revenue—more than most large retailers are forecast to deliver next year. On the recent Black Friday, the AI-driven traffic to US retail sites soared 805% compared to last year, when AI tools such as Walmart's Sparky or Amazon's Rufus had not yet been launched.
In addition, the global shift in consumer spending patterns from traditional brick-and-mortar stores to online shopping should remain a key catalyst for leading regional e-commerce players such as SEA Ltd, Coupang, and Alibaba.
Early signs of an AI bubble have emerged, but not worrisome (yet)
Recent developments, such as the massive CAPEX poured into AI and round-tipping activities (investing in another company that then buys back from you), are pointing towards a potential tech bubble.
Read more: Semiconductor: Are We in an AI Bubble?
While there are signs of a bubble forming, there are still differences compared with previous episodes. Notable historical examples include the railway boom in the 1840s and the internet boom in the late 1990s. These transformations did indeed change the world, but timing matters. From 1843 to 1853, railway mileage in the United Kingdom nearly quadrupled, yet railway revenue per mile remained flat or declined. By mid-2001, telecom companies had installed 39 million miles of fibre, but only 10% of those fibres were lit, and each lit fibre was utilising just 10% of the available wavelengths.
Both the railway and internet booms featured tremendous excess capacity—capacity that was not justified by concurrent consumer demand or unit economics. Today’s AI story certainly features the rhetoric and investment one would expect to see during a paradigm shift. However, we do not yet see excess capacity. Data centre vacancy rates are at a record low of 1.6%, and three-quarters of the data centre capacity under construction is already pre-leased. Across the computing, power, and data-centre value chain, components remain scarce relative to demand. Moreover, the latest earnings season confirms that AI usage is driving revenue growth for the largest companies.
However, this could still be a ticking time bomb. OpenAI has committed to multi-billion-dollar deals with major companies such as Nvidia, Oracle, AMD, and Microsoft without possessing the financial capacity to support such commitments. We therefore continue to question whether OpenAI can fulfil these massive contracts over the long term, as well as the potential overcapacity risk for cloud companies.
Although we see no excess capacity today, AI infrastructure cycles can pivot quickly if deployment lags enterprise adoption. Any slowdown in enterprise deployment would warrant reassessment.
Figure 7: ChatGPT adoption is much faster than the internet
Figure 8: CAPEX are backed by strong Remaining Performance Obligations (RPOs)
Earnings outlook for Internet companies remain robust; Macro catalysts are supportive
While valuations remain at the higher end, we believe that such valuations are justified by the companies’ underlying fundamentals and revenue potential. Moving into 2026, we expect strong earnings growth in the internet sector, supported by the broadline retail and entertainment industries. For the interactive media and software industries, we take a more selective approach, favouring larger-cap companies.
The macro environment remains supportive, with a favourable political backdrop, easing monetary policy, and beneficial tax rates under the One Big Beautiful Bill Act (OBBBA), all of which should continue to underpin company valuations.
Overall, earnings for the NASDAQ Internet Index are projected to grow by 11% in 2026 and 17% in 2027, following a 24% earnings increase in 2025.
Figure 9: Projected earnings growth remains robust
Downside risks to our call
a) Underestimation of depreciation
Recent debate about depreciation for GPU chips has been making headlines. Investors have begun to worry that GPUs have useful lives of only one or two years, while cloud providers depreciate these assets over five or six years. An accounting mismatch like this could lead to depreciation expenses being underestimated by around 30%. While nobody seems to have an exact answer on how the useful life of these chips should be recorded, we believe that even the worst case scenario (useful life reduces to 2years), the impact will not be a long lasting one as once analysts adjusted their models and forecast based on the new useful life, future earnings would not see a big downside surprise. As such, this would not derail the investment case of AI.
Table 3: Network/Compute Depreciation Useful Life (Years)
|
Company |
2020 |
2021 |
2022 |
2023 |
2024 |
2025 |
|
Meta |
3 |
4 |
4.5 |
4.5 |
4.5 |
5.5 |
|
|
3 |
4 |
4 |
6 |
6 |
6 |
|
Oracle |
5 |
5 |
5 |
5 |
6 |
6 |
|
Microsoft |
3 |
4 |
6 |
6 |
6 |
6 |
|
Amazon |
4 |
4 |
5 |
5 |
6 |
5 |
|
Source: Scion Asset Management. Data as of 30 September 2025. |
||||||
b) Power constraints
As utilities struggle to sort fact from fiction, the grid itself becomes a bottleneck. In the US, it is forecasted that data centres may consume 12% of US Electricity by 2028. Yet, the grid is not ready for this influx. Upgrading transmission infrastructure and securing alternative power sources that bypass the grid are among the urgent steps required. In the most recent earnings season, hyperscalers have raised their CAPEX guidance as “AI demand continues to outstrip supply.” However, if these data centres are built but cannot be brought online in time due to interconnection delays, this could affect the companies’ revenue.
Maintain Very Attractive star rating with a cautious tone
AI is entering a more volatile phase, echoing early-stage patterns seen in past technology revolutions.
Acknowledging concerns surrounding AI spending and a potential AI bubble, we believe the market is in the early, rational phase of the cycle, where elevated expectations remain supported by strong revenue momentum, expanding order books, and constrained infrastructure supply.
We continue to see upside for the internet sector while remaining mindful about over-optimism, particularly in relation to some hyperscalers’ RPOs and client financial clarity. As a result, we apply more conservative earnings forecast and avoid extrapolating peak demand scenarios.
Applying a fair P/E multiple of 30x translates into an upside potential of around 35% by 2027. At times like this, disciplined positioning (prioritising companies with strong monetisation pathways, balance-sheet strength, and lower risk of technological obsolescence) is key to avoid pockets of irrational exuberance.
While echoes of past bubbles persist, the greater risk today is under-exposure to this transformational technology.
In short, our stance emphasises selective exposure to structurally advantaged leaders rather than broad thematic buying, enabling investors to balance upside potential with prudent risk management.
We continue to recommend:
· ETF: Invesco NASDAQ Internet ETF (NASDAQ: PNQI)
· Unit Trust: Fidelity Global Technology A-ACC-USD and Eastspring Investments Unit Trusts – Global Technology SGD
Figure 10: Tech earnings have outpaced those of the global market
Table 4: Valuation for PNQI ETF
|
|
2024 |
2025 |
2026 |
2027 |
|
EPS |
48.7 |
60.5 |
65.9 |
76.5 |
|
EPS growth |
24.0% |
9.0% |
16.1% |
|
|
PE |
|
28.1 |
26.2 |
24.0 |
|
Upside |
|
|
|
35% |
|
Fair PE |
30 |
|
|
|
|
Target Price |
1699 (current) |
|
|
2295 |
|
ETF Target Price |
54 (current) |
|
|
73 |
|
Source: Bloomberg Finance L.P., iFAST compilations. Data as of 30 November 2025. |
||||
Declaration:
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.
All materials and contents found in this site are strictly for general circulation and informational purposes only and should not be considered as an offer, or solicitation, to deal in any of the funds or products found/identified in this site. While iFAST Financial Pte Ltd ("IFPL") has tried to provide accurate and timely information, there may be inadvertent delays, omissions, technical or factual inaccuracies and typographical errors. Any opinion or estimate contained in this report is made on a general basis and neither IFPL nor any of its servants or agents have given any consideration to nor have they or any of them made any investigation of the investment objective, financial situation or particular need of any user or reader, any specific person or group of persons. You should consider carefully if the products you are going to purchase are suitable for your investment objective, investment experience, risk tolerance and other personal circumstances. If you are uncertain about the suitability of the investment product, please seek advice from a financial adviser, before making a decision to purchase the investment product. Past performance is not indicative of future performance. The value of the investment products and the income from them may fall as well as rise. Opinions expressed herein are subject to change without notice. In respect of any matters arising from, or in connection with the said research analyses or research reports, recipients of the report are to contact IFPL at 10 Collyer Quay, #26-01 Ocean Financial Centre Building, Singapore 049315, or by telephone at +65 6557 2853. Where the report contains research analyses or research reports from a foreign research house and if the recipient of such research analyses or research reports is not an accredited investor, expert investor, institutional investor or an ex-accredited investor, IFPL accepts legal responsibility for the contents of such analyses or reports to such persons only to the extent as required by law. Please note that only certain security(ies) herein are available to all investors, while the rest are only available for certain persons to invest in, such as Accredited Investors (as defined in the Securities and Futures Act) or one who invests at least S$200,000 (or its equivalent currency) per transaction. To qualify as an Accredited Investor, one needs to submit a declaration form and certain relevant supporting documents, according to iFAST’s prevailing policies and procedures.
Please read our full disclaimers on the website at ( https://secure.fundsupermart.com/fsmone/policies/328125/investment-account-terms-&-conditions).
iFAST Financial Pte Ltd (IFPL) (registered address: 10 Collyer Quay #26-01 Ocean Financial Centre Singapore 049315, Telephone: 6557 2000) holds the Financial Advisers Licence issued by the Monetary Authority of Singapore ('MAS') to conduct regulated activities of advising on securities, marketing of collective investment schemes and arranging of any contract of insurance in respect of life policies, other than a contract of reinsurance and the Capital Markets Services Licence issued by the MAS to conduct regulated activities of dealing in securities and providing custodial services for securities. While IFPL has made every effort to ensure the independence of the report's contents, IFPL's nature of business is such that IFPL and its connected and associated entities together with their respective directors, officers and staff may be involved in providing dealing or investment-related services in the abovementioned securities, and have taken or may take positions in the securities mentioned in this report, and may also act as the principal for any buy or sell trades.