To The Moon

The Gambler’s Fallacy

The $S&P 500(.SPX)$ delivered gains of 26% in 2023, 25% in 2024, and 18% in 2025. After such a strong run, it is natural for investors to expect weaker returns in 2026. This mindset reflects the classic gambler’s fallacy—the belief that strong past performance must be followed by mean reversion.

At this stage of the cycle, bearish arguments also tend to sound more persuasive, and investor sentiment often turns cautious. However, the bull market has continued to show resilience. The ongoing AI-driven transformation remains a powerful structural tailwind and is likely to continue providing support into 2026.

2026 Roadmap

Our base-case scenario sees the $S&P 500(.SPX)$ ending 2026 at 7,546, representing roughly 10% upside. This assumes modest valuation normalization, with the P/E ratio compressing from 25.4x to 24.5x, offset by approximately 14% EPS growth in 2026.

That said, a downside scenario cannot be ruled out. In a more adverse outcome, a sharper PE reset to around 20x could drive the $S&P 500(.SPX)$ down by about 21% in 2026.

Historical seasonality for the S&P 500 during US presidential Year 2 (since 1929) suggests a familiar pattern: some upside in Q1, a correction during Q2 and Q3, followed by a recovery and rally into Q4—assuming historical seasonality repeats.

Separately, the price action in 2025 shows notable similarities to the 1980s period. Under a bearish scenario, this raises the risk that the $S&P 500(.SPX)$ could enter a prolonged consolidation phase lasting up to seven months starting in January 2026.

More broadly, investors typically approach 2026 with caution due to the historical dynamics of the US midterm election cycle. Historically, presidential Year 2 has delivered the weakest market outcomes among the four-year cycle, with the S&P 500 posting the lowest average return (around 7%), the largest maximum drawdown (approximately -19%), and the lowest probability of a positive year (about 58%) compared with Year 1, 3, and 4.

Source: Bloomberg, Tiger Brokers

However, a more relevant seasonal pattern for investors to note is the $S&P 500(.SPX)$‘s performance during the second year of presidents serving a second term. Historically, this subset of presidential cycles has delivered relatively strong outcomes, with an average return of around 24% and a maximum drawdown of approximately -10%, which is notably more constructive than the broader midterm-year average.

Why Markets Can Climb the Wall of Worries in 2026

Markets may continue to move higher in 2026 because many widely cited risks are either manageable or less severe than feared. At times, bad news can be good news—as weaker data may prompt the Fed or policymakers to step in and provide support.

Bearish Arguments — and Why They May Be Overstated

• Weakening labour market A cooling labour market could strengthen the case for rate cuts, which is generally supportive for equity valuations rather than a negative catalyst.

• GPU depreciation concerns Extending GPU depreciation periods has recently been mischaracterised as accounting manipulation. While the physical lifespan of GPUs used for intensive training may be only 1–3 years, their economic lifespan can extend beyond five years, as older chips remain highly effective for inference workloads.

• Circular funding concerns The so-called “circular funding” issue does not represent an immediate systemic risk. It only becomes problematic if long-term AI demand fails to materialise, leaving highly leveraged players unable to service their debt. As long as real AI adoption and usage continue to expand, this risk remains contained.

• Potential Supreme Court loss on tariffs If the Supreme Court rules against Trump’s tariffs, replacement tariff measures or rebate mechanisms could be introduced. While heightened policy uncertainty may temporarily weigh on markets, it is unlikely to derail the broader equity trend.

• Tight liquidity With bank reserves nearing perceived “scarcity” levels, the Fed ended QT on December 1 and began technical buying of T-bills (approximately $40 billion per month) to manage liquidity. While officials distinguish this from stimulus, investors increasingly expect a return to quantitative easing in 2026 should liquidity stress persist. Moreover, tighter liquidity conditions could ultimately force the Fed to deliver larger-than-expected rate cuts.

• Expensive tech valuations Rapid advances in AI help justify higher valuations because they are turning hype into real products, real revenue, and real productivity gains.

• Tariff escalation Tariffs today are smaller in scale compared with the broad, global tariff actions seen in early 2025. While tariffs may contribute to higher inflation and limit the pace of rate cuts, inflation has so far remained milder than expected.

What Needs to Fall into Place to Lift the Stock Market in 2026?

Several key conditions would need to align to support further upside in equities in 2026.

1. Fiscal support Government policy could play a meaningful role through higher subsidies and tax cuts. For example, many provisions under the One Big Beautiful Bill (OBBB) are set to take effect on January 1, 2026. Trump has also proposed a $2,000 tariff “dividend” for households, excluding high-income earners. In addition, restoring 100% immediate equipment expensing under the OBBB would reduce investment uncertainty, encouraging corporate spending and supporting economic growth from 2026 onward.

2. Fed support Rate cuts and the potential return of quantitative easing would help ease financial conditions and provide support for risk assets.

3. Bank deregulation Easing regulatory constraints could improve credit availability and market liquidity. For instance, a potential change to the supplementary leverage ratio (SLR) would reduce balance-sheet constraints, allowing banks to hold more Treasuries and, in turn, enhance overall market liquidity.

4. AI: From narrative-driven growth to monetisation AI companies need to demonstrate that AI adoption is translating into real revenue growth and materially higher profit margins, rather than remaining a purely narrative-driven theme.

The “AI Contractor” Dilemma

Some AI players, such as Oracle, are currently experiencing a liquidity squeeze driven by success rather than fundamental deterioration.

Many AI-focused cloud providers—such as CoreWeave, Nebius, and Oracle—are seeing explosive demand, reflected in rapidly growing backlogs of signed contracts. However, converting this demand into revenue requires substantial upfront investment in data centres, GPUs, and power infrastructure. As a result, capital expenditure runs far ahead of revenue recognition, creating a temporary cash-flow gap and higher leverage.

This situation is further exacerbated by ongoing bottlenecks—such as constrained supply of HBM, power infrastructure, and CoWoS capacity—which delay the completion and ramp-up of data centres. The core risk here is timing, not demand: cash inflows only accelerate once facilities are fully built and operational. Beyond that point, the business model can become highly profitable.

What If AI Demand Falls Short?

In a worst-case scenario where AI demand fails to materialise, most signed contracts would likely be renegotiated rather than abruptly cancelled. Terms could be delayed or scaled down, as neither customers nor suppliers would want to trigger defaults. Technology companies have historically responded to failed innovation cycles by cutting capital expenditure and reducing headcount to stabilise cash flows.

Equity prices may initially sell off sharply, but investors often respond more positively once cost discipline is restored and further margin deterioration is avoided.

From Low-Capex to High-Capex Tech

Tech stocks have corrected in recent months as markets realised that many previously asset-light AI leaders have become capital-intensive businesses, driven by the need to build AI infrastructure and data centres. The implications are significant.

We are increasingly seeing lower—or even negative—free cash flow due to rising capex, reducing companies’ ability to pay dividends or expand share-buyback programmes as they did in the past. Higher capex also pressures profit margins and near-term EPS growth. To fund these investments, companies are taking on more debt, reflected in rising debt-to-equity ratios.

Source: Bloomberg, Tiger Brokers

Companies may instead turn to equity issuance; however, with valuations already elevated across much of the AI sector, such moves may be poorly received by the market. Investors tend to view new share issuance at high valuations as dilution, often triggering sharp share-price declines. As a result, debt financing may be the preferred option for funding capex.

The market has also become far more pragmatic. Investors are losing patience with the traditional “burn cash for growth” narrative and are now demanding a clearer path to monetisation.

The key question is whether the old model—keeping capex low to maximise free cash flow for buybacks—remains superior, or whether higher capex aimed at boosting long-term productivity and earnings growth is the better approach. I would argue the latter: higher capex, if well executed, is an investment in securing future cash flows and earnings power.

The real risk is whether companies have enough time to execute this transition without facing financial stress. Large tech companies such as the Magnificent Seven are likely to have sufficient runway, supported by strong cash generation from their core businesses. Smaller players—such as Nebius, IREN, CoreWeave, and Oracle—face higher risks of credit stress given thinner balance sheets and heavier reliance on external funding.

Too big to fail?

Nvidia and OpenAI have aggressively partnered with a broad network of cloud providers, hardware suppliers, and investment partners to build a deeply interconnected AI ecosystem spanning large-scale data centres and advanced AI development. Over time, this growing interdependence increases systemic importance and begins to resemble a “too big to fail” structure.

Put simply, the AI ecosystem is becoming so interconnected that problems at one major player could spill over to the rest of the system. The real risk is not that a single weak company fails on its own, but that even a single messy default could shake confidence across the wider credit market.

If a systemically important player were unable to service its debt, lenders could quickly become more cautious toward the entire sector. This would likely lead to higher borrowing costs, tighter credit conditions, and reduced liquidity, creating a knock-on effect across the AI financing ecosystem.

AI Boom or Doom in 2026?

We are firmly in an AI-led investment cycle. While many investors acknowledge bubble-like characteristics, this does not necessarily imply an imminent collapse. Even if an AI bubble exists, it appears to have further room to run, supported by several structural factors:

• A large and expanding total addressable market

• Valuations that remain well below dot-com era extremes

• AI spending becoming an increasingly important driver of economic growth and corporate earnings

Why This AI Bull Market Is Different from the Dot-Com Bubble

Many investors draw parallels between the 2000 dot-com bubble and the current AI rally, arguing that valuations are elevated in both periods. However, there are important differences. At the peak of the dot-com bubble, the Nasdaq 100 traded at a P/E ratio of approximately 175x, compared with around 38x today. While a 38x multiple is still high relative to recent history, strong future cash-flow and EPS growth can gradually bring current valuations down over time.

The most critical distinction is that many dot-com era companies were built on compelling narratives but generated little to no revenue. In contrast, today’s AI leaders are seeing real demand, measurable adoption, and growing revenue contributions from AI-related products and services.

Why the AI Boom Resembles 2008 More Than 2000?

That said, the current AI bull market arguably shares more similarities with the 2008 financial crisis than with the dot-com bubble. The 2008 crisis began when weaker borrowers—particularly subprime mortgage holders—defaulted, triggering a chain reaction through complex financial products such as MBS and CDOs. These losses spread across the global financial system, froze credit markets, and ultimately led to major institutional failures.

In the current AI cycle, early signs of stress are beginning to appear among weaker borrowers. These companies typically carry higher debt-to-equity ratios and sustain elevated levels of capital expenditure relative to both revenue and free cash flow. Over time, such firms may face challenges in servicing interest payments or refinancing principal at maturity.

For now, however, credit risks remain contained. There have been no widespread credit-rating downgrades or material failures to meet interest obligations. In fact, AI-related debt could be repackaged into structured products such as CLOs or CBOs, distributing risk across a broader investor base and potentially extending the AI-led credit and investment cycle.

The key risk that could ultimately disrupt this AI boom is an unexpected rise in interest rates. Higher rates would tighten financial conditions, raise funding costs, and expose leverage across the AI ecosystem—potentially acting as the catalyst that brings the cycle to an end.

Conclusion

• The $S&P 500(.SPX)$ enters 2026 well positioned for another year of solid performance, supported by earnings momentum, favourable liquidity conditions, and powerful tailwinds from the AI supercycle driving capital spending, productivity gains, and multi-sector earnings growth.

• While concerns are rising around stretched valuations, supply-chain bottlenecks, debt-funded expansion, policy uncertainty, and shifting tariff expectations under a Trump administration, these appear to be manageable risks rather than immediate deal-breakers.

• High valuations alone are unlikely to trigger a sharp market correction. Valuations can normalise over time through earnings and cash-flow growth, provided fundamentals continue to improve.

• The more meaningful risk lies in AI-related credit stress. Movements in investment-grade and high-yield option-adjusted spreads (IG & HY OAS) are among the earliest indicators of rising systemic stress.

• AI is still far from achieving artificial general intelligence (AGI). A true speculative peak would likely occur only when markets begin pricing in AGI as imminent or already achieved.

For SG users only, Welcome to open a CBA today and enjoy access to a trading limit of up to SGD 20,000 with unlimited trading on SG, HK, and US stocks, as well as ETFs.

🎉Cash Boost Account Now Supports 35,000+ Stocks & ETFs – Greater Flexibility Now

Find out more here.

Complete your first Cash Boost Account trade with a trade amount of ≥ SGD1000* to get SGD 688 stock vouchers*! The trade can be executed using any payment type available under the Cash Boost Account: Cash, CPF, SRS, or CDP.

Click to access the activity

Other helpful links:

• 💰Join the TB Contra Telegram Group to Get $10 Trading Vouchers Now🎉

• How to open a CBA. How to link your CDP account. Other FAQs on CBA. Cash Boost Account Website.