AI Meets the Cost of Capital: My Q4 2026 Playbook for Trading, Investing and Portfolio Risk

Author’s Note and Disclaimer:
Zion Zhao Real Estate | 88844623 | 狮家社小赵 | wa.me/6588844623 |  https://linktr.ee/zionzhao | https://www.huttonsgroup.com/vcard/R071443Z

This post is for general information, education, and market literacy only. It does not constitute financial, investment, trading, legal, tax, accounting, or other professional advice, and is not an offer, solicitation, recommendation, or endorsement. Views expressed are personal, general in nature, and subject to change without notice. While reasonable care is taken, no representation or warranty is given as to accuracy, completeness, or reliability. Readers should conduct independent due diligence and seek professional advice. To the fullest extent permitted by law, no liability is accepted for any loss arising from reliance on this material.  

AI Rally Meets Rate Reality as Investors Reprice the Cost of Growth

AI Leadership Meets the Rate Shock: My Q4 2026 Trading and Investment Playbook

The defining investment question heading into the fourth quarter of 2026 is no longer whether artificial intelligence is generating real economic activity. The evidence increasingly says that it is. The more important question is whether extraordinary AI-driven earnings growth can continue to outrun an increasingly expensive cost of capital.

That tension now defines the market.

Microsoft, Alphabet, Amazon, Nvidia, AMD, Broadcom, Taiwan Semiconductor Manufacturing Company, Meta and Palantir are participating in one of the most aggressive infrastructure and software investment cycles in modern technology history. Cloud capacity remains constrained in parts of the ecosystem. AI accelerators, networking equipment, advanced semiconductor manufacturing, high-bandwidth memory, data centres and enterprise AI software are absorbing enormous amounts of capital.

At the same time, the Federal Reserve has resumed tightening, long-term Treasury yields remain elevated, and investors are once again being forced to price duration, capital intensity and valuation risk seriously. In this environment, a company can deliver excellent earnings and still experience multiple compression. A technically powerful stock can fall sharply if expectations become excessive. Equally, a weak chart does not automatically mean the underlying business is deteriorating.

My framework is therefore neither mechanically bullish nor reflexively bearish. I view the current market as constructively bullish, but increasingly conditional.

The correct response is not prediction. It is disciplined positioning.

Technical analysis should help determine when to enter, where a thesis is invalidated and how much capital should be risked. Fundamental analysis should determine whether a business deserves to remain in the portfolio after the trade is over. Academic research provides some support for momentum and systematically defined technical patterns, but it also shows that technical strategies are highly dependent on methodology, market regime and execution costs (Lo et al., 2000; Moskowitz et al., 2012; Park & Irwin, 2007).

In other words, price action is evidence, not certainty.

The broader market remains the first gatekeeper. The technical framework from the source material identifies SPY support around 763.52 and QQQ support around 734.58, with QQQ facing a major resistance zone around 748.65. As long as both indices remain above their support zones, I am comfortable maintaining constructive long exposure. A decisive QQQ breakout above resistance would strengthen the case for another technology-led expansion. Conversely, simultaneous failures below SPY and QQQ support would be a reason to reduce short-term gross exposure rather than argue with the tape.

Among the mega-cap technology names, Microsoft remains one of the strongest combinations of quality, monetization and technical clarity. Its Azure business continues to demonstrate that AI demand is translating into commercial cloud revenue, not merely speculative capital expenditure. Microsoft reported fiscal fourth-quarter 2026 revenue of $90.0 billion, with Azure and other cloud-services revenue rising 43% (Microsoft, 2026). Technically, my research identifies approximately $489.30 as meaningful support and $526.73 as the major breakout threshold, with a longer-term channel objective near $599.21. I would rather buy Microsoft on a controlled pullback or after a confirmed breakout than chase it directly beneath resistance.

Alphabet may offer one of the cleanest risk-defined opportunities. The company continues to deliver strong operating growth, with second-quarter revenue increasing 24% and Google Cloud revenue surging 82% (Alphabet, 2026). Yet the stock is sitting near a major technical decision area. My research places critical support around $343.68, followed by a deeper structural pivot near $319.04. Holding that first level keeps the pathway toward approximately $408.61 and potentially $452.30 intact. Losing both would materially change the medium-term setup. This is the kind of opportunity I prefer: an excellent business, identifiable support and clearly defined invalidation.

Amazon illustrates the distinction between a good company and a good trade. The business is fundamentally strong. Second-quarter sales increased 20%, AWS revenue accelerated and operating income rose sharply (Amazon, 2026). However, the stock remains below my analysis of approximately $259.42 technical pivot. Below that level, the chart allows for a retreat toward roughly $232.14, while a recovery above $259.42 would reopen the path toward approximately $294 to $302. As a trader, I would wait. As a long-term investor, I would view substantial weakness differently if the fundamental thesis remained intact.

Meta represents another variation of the same problem. Its advertising engine remains exceptionally powerful, but its AI ambitions require enormous investment. Second-quarter revenue increased 28%, while capital expenditures reached more than $31 billion in the quarter, with full-year capital expenditure expected to remain extraordinarily high (Meta Platforms, 2026). Technically, the stock remains constructive above approximately $697.16, with upside zones near $836.20 and potentially $931.14. The key long-term question is no longer whether Meta can spend aggressively. It is whether those investments can earn attractive returns through higher engagement, better advertising economics, agents and new AI products.

The semiconductor complex remains the clearest expression of the AI infrastructure boom.

Nvidia continues to operate at extraordinary scale. Fiscal second-quarter 2027 revenue reached $96.2 billion, with Data Center revenue of $89.0 billion, up 117% year over year (NVIDIA, 2026). The short-term technical framework identifies approximately $221.09 as immediate support, $225.80 as a bullish confirmation level and $208 as more important secondary support. Long term, Nvidia remains a core AI infrastructure holding in my framework, but no company should become so large that one valuation shock dominates portfolio risk.

AMD is increasingly credible as an alternative AI-compute platform. Its second-quarter revenue reached $11.5 billion, while Data Center revenue rose 107% to $6.7 billion (AMD, 2026). Technically, approximately $556.48 is the major structural support identified in the primary analysis, while roughly $737 represents the next major channel objective. Given how far the stock has already advanced, I would prefer buying pullbacks rather than pursuing vertical momentum.

TSMC remains one of the most strategically valuable businesses in the technology supply chain because it benefits from multiple competing AI platforms rather than having to select a single winner. Its second-quarter revenue reached approximately US$40.2 billion with a gross margin of 67.7% (TSMC, 2026). The technical framework remains constructive above approximately $405.20, with intermediate resistance near $479 and a longer objective around $525.31. The principal long-term risks are capital intensity, semiconductor cyclicality, customer concentration and geopolitics around Taiwan.

Broadcom may offer the most interesting disagreement between fundamentals and technicals. Its business is exceptionally strong, with fiscal third-quarter 2026 revenue increasing 86% and free cash flow reaching approximately $13.7 billion (Broadcom, 2026). Yet the technical picture remains weaker below approximately $372.59. A recovery above that area would improve the setup materially and place approximately $419.61 back in focus. This is precisely why I separate investing from trading. Weak price action should influence execution, but it should not automatically be confused with weak business economics.

Palantir sits somewhere between phenomenal operating momentum and extraordinary expectations. Second-quarter revenue increased 93% year over year, while U.S. commercial revenue increased 149% (Palantir, 2026). The short-term chart provides a remarkably clean decision zone between roughly $188.37 support and $194.80 resistance. A breakout above the upper boundary opens the way toward approximately $207.52 and potentially the longer-term $248.35 channel target. A failure below support increases the probability of a deeper retracement. I would not chase Palantir in the middle of that range.

SpaceX also deserves special treatment because it is now a publicly traded company under SPCX rather than simply a private-market reference. Its post-IPO chart remains young, which means technical conclusions should be treated with additional caution. The source analysis identifies approximately $139.10 to $139.20 as the major pivot, with upside potential toward approximately $179.49 to $185 if support holds. A weekly failure below that pivot would materially weaken the structure. I would size SpaceX below mature mega-cap technology companies because both valuation uncertainty and public-market price discovery remain substantial.

Intel should be treated as a turnaround rather than a core compounder. Revenue growth has improved, but profitability and manufacturing economics still require confirmation. My technical analysis identifies approximately $113.07 as major support, $139.57 as the next resistance area and roughly $173.61 as the subsequent objective if the turnaround continues technically. I would demand multiple quarters of execution before materially increasing a long-term position.

Tesla remains technically constructive but fundamentally more difficult to value. Its market capitalization increasingly discounts autonomy, robotics, energy storage and artificial intelligence rather than conventional automotive earnings alone. That produces enormous optionality if execution succeeds, but also substantial valuation risk if timelines disappoint. I therefore treat Tesla as a smaller speculative-growth allocation rather than the equivalent of Microsoft, Alphabet or TSMC.

Micron is attractive because AI memory and high-bandwidth memory demand remain structurally important, but memory is still cyclical. With earnings immediately ahead, I would prefer waiting for the post-earnings reaction rather than converting a technical trade into an uncontrolled binary event.

Strategy should be analysed differently again. It is effectively a capital-markets vehicle whose economic exposure is dominated by Bitcoin. Its approximately $144 technical support and roughly $173.61 upside reference may be useful tactically, but long-term portfolio sizing should be considered part of cryptocurrency risk rather than traditional software exposure.

The most important element of my short-term framework is position sizing. I would generally risk approximately 0.35% to 0.75% of portfolio net asset value on a conventional swing trade and less on exceptionally volatile securities. The size of a position should be determined by the distance from entry to invalidation, not by confidence. If a portfolio is worth $200,000 and the maximum planned loss is 0.5%, the total acceptable loss is $1,000. A trade with $10 of risk per share therefore supports roughly 100 shares. A trade with $25 of risk supports approximately 40.

That calculation is far more important than whether a stock appears cheap or expensive in absolute dollar terms.

For long-term investing, my preference is different. I would concentrate the core of an AI and technology sleeve in Microsoft, Alphabet, Amazon, TSMC, Nvidia and Broadcom. Meta, AMD, Palantir and SpaceX belong in a higher-growth satellite allocation. Micron, Intel, Tesla and Strategy belong in smaller cyclical, turnaround or speculative positions.

I would also deploy long-term capital in stages rather than all at once. A first tranche establishes exposure. A second can be added during a conventional correction. A third can follow earnings confirmation. A final tranche remains available for either a major drawdown or a successful breakout and retest.

The objective is not to perfectly identify the bottom. It is to build exposure to businesses whose long-term earnings power can compound faster than the investor's cost of capital without taking catastrophic concentration risk.

The broader conclusion is therefore simple.

Artificial intelligence increasingly looks like infrastructure rather than experimentation. Real money is being spent, real revenue is being generated and real competitive advantages are being created. But the market is now operating in a world of higher discount rates, enormous capital requirements and elevated expectations.

That environment rewards selectivity.

I want to own the strongest businesses for the long term, trade around clearly defined technical pivots in the short term, preserve capital when the market invalidates a setup, and never confuse a powerful technological revolution with permission to ignore valuation.

AI may remain one of the defining investment themes of this decade.

The portfolio that ultimately wins, however, may not be the one that owns the most AI.

It is more likely to be the one that understands where economic value is being captured, what price is being paid for that value, and how much capital should prudently be placed at risk when the future remains uncertain.

References

Advanced Micro Devices, Inc. (2026). AMD reports second quarter 2026 financial results. AMD Investor Relations.

Alphabet Inc. (2026). Alphabet announces second quarter 2026 results. Alphabet Investor Relations.

Amazon.com, Inc. (2026). Amazon.com announces second quarter results. Amazon Investor Relations.

Broadcom Inc. (2026). Broadcom Inc. announces third quarter fiscal year 2026 financial results and quarterly dividend. Broadcom Investor Relations.

Federal Reserve Board. (2026, September 16). Federal Reserve issues FOMC statement. Board of Governors of the Federal Reserve System.

Intel Corporation. (2026). Intel reports second-quarter 2026 financial results. Intel Investor Relations.

Lo, A. W., Mamaysky, H., & Wang, J. (2000). Foundations of technical analysis: Computational algorithms, statistical inference, and empirical implementation. The Journal of Finance, 55(4), 1705–1765. https://doi.org/10.1111/0022-1082.00265

Meta Platforms, Inc. (2026). Meta reports second quarter 2026 results. Meta Investor Relations.

Microsoft Corporation. (2026). Microsoft Cloud and AI strength fuels fourth quarter results. Microsoft Investor Relations.

Moskowitz, T. J., Ooi, Y. H., & Pedersen, L. H. (2012). Time series momentum. Journal of Financial Economics, 104(2), 228–250. https://doi.org/10.1016/j.jfineco.2011.11.003

NVIDIA Corporation. (2026). NVIDIA announces financial results for second quarter fiscal 2027. NVIDIA Investor Relations.

Palantir Technologies Inc. (2026). Q2 2026 business update. Palantir Investor Relations.

Park, C.-H., & Irwin, S. H. (2007). What do we know about the profitability of technical analysis? Journal of Economic Surveys, 21(4), 786–826. https://doi.org/10.1111/j.1467-6419.2007.00519.x

Taiwan Semiconductor Manufacturing Company Limited. (2026). 2026 Q2 quarterly results. TSMC Investor Relations.

AI牛市没结束，但投资逻辑已经变了

AI龙头迎战利率冲击：我的2026年第四季度交易与投资框架

进入2026年第四季度，市场最关键的投资问题，已经不再是人工智能是否正在创造真实的经济价值。越来越多的证据显示，答案是肯定的。真正更困难的问题是，由AI驱动的超高速盈利增长，能否继续跑赢不断上升的资本成本。

微软 Microsoft、Alphabet、亚马逊 Amazon、英伟达 NVIDIA、AMD、博通 Broadcom、台积电 TSMC、Meta 和 Palantir，正在参与现代科技产业史上规模最大的基础设施投资周期之一。云计算资源、AI加速器、先进制程半导体制造、网络设备、高带宽存储器 High-Bandwidth Memory，简称 HBM，以及企业级人工智能软件的需求依然非常强劲。

与此同时，美联储 Federal Reserve 已重新进入加息周期，美国国债收益率仍处于高位，投资者再次必须认真评估久期风险 Duration Risk、资本密集度 Capital Intensity 以及估值风险 Valuation Risk。

因此，我对当前市场的判断是：

结构性偏多，但越来越依赖条件确认。

技术分析 Technical Analysis 不应该被视为预测未来的工具。它更适合作为执行、风险管理以及判断交易逻辑何时失效的框架。学术研究显示，动量 Momentum 以及部分系统化技术形态确实可能包含一定的信息价值，但其有效性会随着市场环境、研究方法及执行方式而显著变化（Lo et al., 2000; Moskowitz et al., 2012; Park & Irwin, 2007）。

在组合层面，SPY与QQQ依然是最重要的市场总开关。只要两者能够守住主要支撑区，我愿意继续维持选择性的多头敞口。若QQQ能够确认突破关键阻力，将进一步强化科技股进入新一轮上涨扩张阶段的可能性。相反，如果SPY与QQQ同时跌破主要支撑，则应考虑主动降低短期风险敞口。

从个股来看，微软仍然是商业质量、AI商业化能力与技术结构结合得最好的公司之一。Alphabet如果能够守住主要支撑，则提供相对清晰的风险收益结构。亚马逊的基本面依然强劲，但技术结构较弱，这正说明了为什么一家优秀公司并不代表在任何价格下都是优秀的交易机会。

半导体产业依然是当前AI基础设施繁荣最直接的表达方式。英伟达的数据中心业务仍在高速增长。AMD正在成为越来越具竞争力的AI计算替代平台。台积电拥有重要战略地位，因为它能够从多个相互竞争的芯片生态系统中受益。博通则是一个典型案例，其AI网络与定制芯片业务基本面极强，但短期技术结构相对偏弱，再次说明企业质量与股价技术质量必须分开分析。

Meta依然具有吸引力，但资本投入正在显著上升。Palantir继续保持异常强劲的增长，但其估值已经留下较小的执行失误空间。SpaceX拥有长期科技可选性 Long-Duration Technological Optionality，但由于上市后的公开市场价格发现 Price Discovery 仍处于较早阶段，仓位理应更小。

英特尔 Intel 更适合被视为企业转型 Turnaround，而不是已经被证明的长期复利型公司。特斯拉 Tesla 在自动驾驶 Autonomy、机器人 Robotics 和能源业务方面拥有巨大可选性，但其估值同样高度依赖未来执行。美光 Micron受益于AI存储与HBM需求，但存储产业仍然具有明显周期性。Strategy则更应该被视为具有资本结构放大效应的比特币敞口，而不是传统软件投资。

对于短期交易而言，仓位管理比主观信心更重要。我通常更倾向于每一笔交易只承担投资组合中较小比例的资本风险，并根据进场价与失效点之间的距离来决定仓位规模。半导体及AI相关股票之间的相关性较高，因此风险也应该从组合层面整体管理，而不是逐只股票孤立判断。

对于长期投资而言，我会把AI与科技配置的核心集中在微软、Alphabet、亚马逊、台积电、英伟达和博通。Meta、AMD、Palantir以及SpaceX则更适合作为高增长的卫星仓位 Satellite Positions。美光、英特尔、特斯拉和Strategy则应属于规模更小的周期性、转型型或高风险配置。

更大的结论其实非常明确。

AI正在越来越像一种经济基础设施，而不仅仅是一项实验性技术。但市场已经不再处于那个宽松、低利率且对估值高度包容的环境。

因此，真正有效的投资策略，并不是持有最多的AI股票。

真正重要的是理解：

经济价值究竟由谁获取，投资者为这些价值支付了什么价格，以及当未来仍然存在不确定性时，我们究竟应该让多少资本承担风险。

参考文献 References

Lo, A. W., Mamaysky, H., & Wang, J. (2000). Foundations of technical analysis: Computational algorithms, statistical inference, and empirical implementation. The Journal of Finance, 55(4), 1705–1765.

Moskowitz, T. J., Ooi, Y. H., & Pedersen, L. H. (2012). Time series momentum. Journal of Financial Economics, 104(2), 228–250.

Park, C.-H., & Irwin, S. H. (2007). What do we know about the profitability of technical analysis? Journal of Economic Surveys, 21(4), 786–826.

AI牛市遇上利率冲击：Q4怎么投？

进入2026年第四季度，我更关注的已经不是“AI还能不能涨”，而是：AI带来的盈利增长，能否继续跑赢越来越高的资本成本。

微软、Alphabet、亚马逊、英伟达、AMD、博通、台积电、Meta、Palantir，正在推动一轮史无前例的AI基础设施周期。云计算、AI芯片、先进制程、HBM高带宽存储器、数据中心与企业级AI软件，仍然有真实需求支撑。

但另一边，美联储重新收紧政策，美国国债收益率维持高位。资本不再便宜，市场开始重新计算估值、现金流、回报率与风险溢价。

所以我对当前市场的判断是：结构性偏多，但必须有条件地看多。

短线交易上，我不会只看“故事”，而更重视技术确认、支撑阻力、仓位控制与止损纪律。SPY和QQQ仍然是市场总开关，指数守住关键支撑，个股机会才更值得参与。

长期投资上，我更偏向把微软、Alphabet、亚马逊、台积电、英伟达、博通视为核心AI基础设施资产；Meta、AMD、Palantir、SpaceX属于更高成长、也更高波动的卫星仓位；Micron、Intel、Tesla与Strategy则更适合较小规模配置。

最重要的一点是：好公司，不等于任何价格都是好买点。

这也是我做新加坡房地产多年后越来越重视的一套逻辑。股票、房地产、债券，本质上都逃不开几个核心变量：增长、现金流、利率、估值、杠杆与风险。

利率低的时候，市场愿意为未来付高价；利率高的时候，现金流和资本回报率重新变得重要。

所以，无论是买股票还是买房，我都不会只问“还能不能涨”，而更会问：

这个资产真正创造什么价值？现在的价格贵不贵？如果判断错了，我愿意承担多少风险？

AI可能会定义未来十年。

但最终决定投资结果的，往往不是谁追得最快，而是谁更懂得资产配置、风险管理与耐心等待。

赵峻慷 Zion Zhao
新加坡房地产｜宏观经济｜资产配置｜投资研究

仅供市场研究与交流，不构成任何投资建议。