本文以实证方式检验2026年标志着技术全面压倒资本的历史分水岭这一命题。定义近代人类历史的技术与资本之间的权力博弈，自蒸汽机时代以来一直呈现一致模式——技术创造价值，资本收割价值。然而截至2026年，AI技术的进化速度超过资本反应时间十倍以上，中产阶级信用基础的结构性侵蚀同步推进，这一长期秩序正在逆转。本研究基于实时事件和数据——包括阿里通义千问核心团队的解体、SaaS市值蒸发2万亿美元、贝莱德私人信贷基金赎回限制、美以打击伊朗及由此引发的油价暴涨、黄金现货价格创历史新高——通过四重战争框架分析技术压倒资本的机制：速度战、多模态机动战、斩首打击和围困战。
引言：技术与资本的永恒对抗
贯穿近代人类历史最根本的张力是技术与资本之间的权力博弈。当蒸汽机侵蚀封建贵族的土地财富时，工业资本家迅速吸收了技术的破坏力量来构建新的权力结构。当电力重构传统制造业时，金融资本通过股票市场和公司治理控制了技术的方向。当互联网消灭了流通中间商时，风险资本通过选择性支持和控制技术企业家来维持资本的霸权。
在每一个案例中，一种一致的模式都占据主导：技术创造价值，资本在初期混乱之后收割价值。熊彼特的”创造性破坏”总是伴随着资本的重组，而破坏后的秩序无一例外地由既有资本权力重建。
然而在2026年，AI技术正在从根本上颠覆这一结构。AI以天为周期进化，而资本以季度为周期反应。这种速度不对称使资本理解和控制技术方向的前提本身失效。与此同时，AI直接取代中产阶级白领就业——资本据以发放信贷的根基——资本体系的存在基础正在被侵蚀。
本研究采用四种方法论。第一，事件驱动分析考察阿里千问团队解体、美国私募基金赎回限制、SaaS行业崩盘和军事行动案例。第二，实时金融数据分析审视黄金价格、美国科技股和油价的走势。第三，比较文化分析识别美国与东亚创新生态系统的结构性差异。第四，军事行动案例分析推导经济战场与物理战场的统一逻辑。
微观案例：阿里通义千问核心团队的解体
2.1 事件经过
On the afternoon of March 3, 2026, Zhou Jingren, CTO of Alibaba Cloud, conveyed a team restructuring plan to Qwen’s technical lead Lin Junyang (age 32). The key change involved transitioning the Qwen team from its existing vertically integrated structure to a horizontally segmented system divided into pre-training, post-training, text, image, and audio units. This reorganization effectively reduced Lin Junyang’s scope of authority.
That same afternoon, at the farewell gathering for Yu Bowen, head of Qwen’s post-training division (on his last working day), team members were informed for the first time that Zhou Hao — a former Google DeepMind senior researcher who had joined Alibaba in January 2026 — would be managing the post-training team. Zhou Hao was a core contributor to the Gemini 3.0 model and had led multi-stage reinforcement learning research.
At 12:11 AM on March 4, Lin Junyang posted a brief message on X (formerly Twitter). The news sent shockwaves through the entire AI industry. What made the situation particularly striking was that just the previous evening, the Qwen team had released the Qwen 3.5 small model series, earning direct praise from Elon Musk himself. Less than 24 hours after receiving the highest level of industry recognition, the core technical leader had departed — a dramatic turn of events.
核心人员离职
（2026年1-3月）
千问团队规模
（单一决策者结构）
阿里对大模型的
直接投资
千问累计
开源模型数
2.2 资本压制技术的三重机制
机制①：通过组织重构稀释权力
The transition from vertical integration to horizontal segmentation ostensibly claims “efficiency” as its justification, but the essence is the dispersal of the technology leader’s unified decision-making authority. Lin Junyang had overseen the entire pipeline from pre-training to post-training, from text to multimodal — but after the restructuring, each direction would have a separate lead. This destroys the coherence of the technical vision.
机制②：空降外部人才
Recruiting Zhou Hao from Google DeepMind was not merely talent reinforcement, but the construction of a new reporting line that bypassed the existing technology leader. Zhou Hao was set to report directly to CTO Zhou Jingren rather than to Lin Junyang. This is a classic corporate power tactic that effectively neutralizes a technology leader’s influence within the organization.
机制③：叙事重构——”不能神化个人”
At an emergency all-hands meeting on the afternoon of March 4, Alibaba’s Chief People Officer Jiang Fang stated: “We must not deify individuals, and we cannot retain people through unconditional, irrational means.” The irony is that Jiang Fang herself is one of Alibaba’s 18 founding members — a direct beneficiary of the Jack Ma mythology. Past entrepreneurial heroes, having become the establishment, now use the framework of “individual deification” to block the emergence of new technical heroes — a modern reenactment of the Chinese proverb “once the rabbits are caught, the hunting dogs are cooked.”
2.3 事件的结构性意义
The Lin Junyang incident is a quintessential case of the technology-capital confrontation manifesting as an internal organizational power struggle. The key insight: when a technologist’s independent influence exceeds the organization’s controllable boundaries, the organization’s instinctive response is not gratitude but fear. The Qwen team produced world-class open-source models with approximately 100 people and limited computing resources — a fraction of what competitors deployed. This very achievement was perceived as a threat by the organization.
According to reporting by 36Kr, the Qwen team numbered roughly 100, and the entire Tongyi Lab supporting it comprised only several hundred. By contrast, ByteDance’s foundational model training team Seed alone approaches 1,000 people, and the absolute headcount Alibaba invested in each direction was a mere fraction of its competitors’. This extreme efficiency was heavily dependent on Lin Junyang’s individual leadership — and it was precisely that dependency that organizational managers perceived as risk.
结构分析：东亚创新生态系统的极限
3.1 美国与东亚的根本差异
In the American ecosystem, a structure in which technologists lead capital is possible. Dual-class share structures institutionally guarantee this. Google’s Larry Page, Meta’s Mark Zuckerberg, and Tesla’s Elon Musk hold absolute voting power regardless of their equity stakes, thereby protecting their technological vision from the short-term pressures of capital markets. The OpenAI case — where Sam Altman, temporarily removed by the board, was reinstated through collective pressure from employees — further demonstrates how powerful the bargaining position of technical talent is in the US.
In East Asia (China, Korea, Japan), the social system itself does not permit technologists to lead capital markets. In China, the government can directly intervene in technology companies’ strategic direction, and indirectly controls the trajectory of innovation through computing power reviews, data compliance regulations, and licensing systems. In Korea, the chaebol system pursues short-term profits and structurally neglects foundational research investment.
3.2 韩国案例：硬件强国，软件荒漠
Korea provides the clearest illustration of the structural limits of the East Asian innovation ecosystem. The Coupang data breach at the end of 2025 encapsulates these limitations.
Beginning in June 2025, a former developer used internal credentials to extract customer data, but Coupang’s security team failed to detect this for five months. The anomaly was not discovered until November 18, and ultimately 33.7 million individuals’ personal information was confirmed compromised. This represents nearly two-thirds of Korea’s entire population, and the leaked data included delivery addresses, phone numbers, and even apartment door-lock passwords. CEO Park Dae-jun resigned on December 10, police raided Coupang’s headquarters, and class-action securities lawsuits were filed in the United States, erasing more than $8 billion in market capitalization.
SK Telecom displayed similar structural problems. In a 2024 SIM data breach, approximately 23.24 million users were affected, and subsequent investigations confirmed that malware had remained dormant in the system for an extended period. Peter Kim of KB Securities noted in a CNBC interview that Korean corporations, in their relentless pursuit of cost efficiency, have structurally neglected areas such as cybersecurity.
美国独角兽企业
中国独角兽企业
韩国独角兽企业
韩国初创企业3年
存活率（OECD最低）
3.3 政府收编：东亚创新的终极天花板
East Asian innovators face a triple ceiling. The first layer is corporate control. As the Lin Junyang case shows, when a technology leader exceeds the organization’s controllable boundaries, immediate checks follow. The second layer is capital control — the absence of dual-class share structures, evaluation based on short-term profit metrics, and an investment culture that favors applications over fundamental research. The third layer is government co-optation.
Government co-optation operates through three modalities. Direct control involves intervening in resource allocation, research direction, and personnel decisions under the banner of national security. Indirect taming uses computing power reviews, data compliance requirements, and licensing systems to steer technology’s direction. Talent co-optation absorbs technology leaders into the system — through academician appointments, political consultative committee positions, and similar mechanisms that transform innovators into advisors. Under this triple-ceiling structure, genuine creative destruction in East Asia is structurally impossible.
AI的本质：全信息链的效率革命
4.1 重新定义AI
AI is not a single technology but an efficiency revolution spanning humanity’s entire chain of information collection, organization, processing, analysis, and generation. Starting from this definition, the targets AI eliminates become clear: the inefficient work of information intermediaries once known as “white-collar workers.”
Legal assistants, junior analysts, translators, auditors, entry-level programmers, customer service agents, administrative staff — what they share in common is that they are fundamentally intermediate nodes in information distribution. AI’s role is to short-circuit these nodes, enabling information to flow directly from producer to consumer.
4.2 ‘SaaS末日’的实证证据
The SaaS stock collapse of January–February 2026 is the most dramatic empirical validation of this logic. On January 30, 2026, Anthropic released 11 open-source plugins for Claude Cowork on GitHub, demonstrating that a general-purpose AI system could handle tasks such as contract review, compliance verification, sales preparation, legal intake, and internal investigations.
The market’s response was immediate. Between January 15 and February 14, 2026, approximately $2 trillion in market capitalization evaporated from the software sector. The iShares Expanded Tech-Software ETF (IGV) fell 22% year-to-date, its largest decline since the rate hikes of 2022. Individual stocks suffered even more severely — Atlassian fell 35%, Salesforce dropped 28%, and Intuit plummeted more than 34%. ServiceNow, despite nine consecutive quarters of earnings beats, fell 11% in a single day.
According to Bain & Company’s analysis, investors are now pricing in the concern that AI can replicate core functionalities and erode installed bases. With net revenue retention stagnating and seat growth no longer serving as a meaningful growth driver, the market has effectively entered a state of paralysis. Goldman Sachs strategist Ben Snider compared software’s future to the newspaper industry, warning of long-term downside risk.
4.3 核心洞察：人类必须回归物理世界
Over the past several decades, human society has undertaken a massive migration from the physical world to the information world. However, once AI takes over the efficiency of the information world, humanity’s unique value resides in the physical world. AI can write code, but it cannot build a house, care for the elderly, or repair a broken machine. Plumbers, chefs, nurses, mechanics — the value of physical-world skills will undergo a fundamental reappraisal.
金融市场证据：从虚拟到实体的历史性转向
5.1 黄金价格创历史新高
Gold is currently experiencing the most dramatic rally in human history. On March 3, 2026, gold spot prices reached an all-time high of $5,417 per ounce, directly triggered by surging safe-haven demand amid escalating Middle East conflict. Gold had first broken through $3,000 per ounce in March 2025, surpassed $4,000 by October 2025, and entered the $5,000 range in January 2026. In just 12 months, the price had risen by more than 100%.
黄金现货价格/盎司
（截至2026.03.02）
黄金12个月
涨幅
2026年季度投资者+
央行需求估算
Midas基金3月
价格目标
J.P. Morgan Global Research projected that investor and central bank gold demand would reach approximately 585 tons per quarter in 2026, with central bank purchases accounting for 190 tons per quarter and gold bar and coin demand at 330 tons per quarter. According to the World Gold Council, gold prices rose nearly 5% in the two days immediately following the outbreak of the Middle East conflict, and the downward trend in the US Dollar Index (DXY) was identified as a major upward driver for gold prices.
The essence of this phenomenon is not simple safe-haven preference. It reflects a historic transformation: the structural weakening of US dollar credit, global central banks’ de-dollarization and asset diversification, and a fundamental revaluation of physical assets relative to financial assets.
5.2 美国科技股的大规模抛售
US technology stocks have faced unprecedented selling pressure since the beginning of 2026. In February, the Nasdaq fell 3.4% — its largest monthly decline since March 2025 — while the S&P 500 dropped 0.9%. The technology sector recorded consecutive periods as the market’s worst performer, with software stocks hit particularly hard.
Big Tech’s massive AI infrastructure investments paradoxically triggered stock price declines. When Alphabet announced annual capital expenditures of $175–185 billion for 2026, far exceeding the market consensus of $119.5 billion, its stock fell more than 4%. Microsoft likewise showed above-expectation capital expenditures and slowing cloud revenue growth in its quarterly results, erasing $360 billion in market capitalization in a single day.
UBS Global Wealth Management noted that Anthropic’s new AI automation tools had triggered concerns about the business model outlook for certain companies. The market now rewards AI spending only when accompanied by strong revenue growth, and after three years of a powerful AI rally, investors have become far more demanding.
5.3 私募基金赎回限制：信用体系的裂缝
On March 6, 2026, BlackRock restricted redemptions from its $26 billion HPS Corporate Lending Fund (HLEND). Investors had requested approximately $1.2 billion in redemptions — 9.3% of net asset value — but management capped the redemption rate at 5%, actually paying out only approximately $620 million. This was the first time since HLEND’s inception that the quarterly redemption limit had been exceeded.
The event sent shockwaves through the entire $1.8 trillion private credit industry. BlackRock shares fell as much as 8.3% on the day, and other alternative asset managers — KKR, Ares Management, Apollo, and Blue Owl Capital — also dropped 5–6%. Blackstone likewise faced record redemption requests the same week, having to raise its standard 5% redemption cap to 7% and invest $400 million of firm and employee capital to meet all requests.
The liquidity crisis in private credit funds is directly connected to AI. As AI erodes the business models of SaaS companies, the underlying asset values of mid-sized technology firms to which private funds have lent are deteriorating. Investors are shifting capital to safe-haven assets amid the convergence of AI-driven industrial restructuring, economic slowdown fears, and market uncertainty from the Middle East conflict.
5.4 油价暴涨：物理世界的回归
On February 28, 2026, the United States and Israel launched a massive attack on Iran. Iran’s nuclear facilities, ballistic missile bases, and command-and-control nodes were struck, and Iran’s supreme leader was killed. Iran retaliated with hundreds of drones and missiles targeting Israel and US military bases in the Persian Gulf.
When markets opened on Monday, March 2, Brent crude surged 9% to approximately $79 per barrel, and WTI crude rose more than 7.5%. Within one week, US crude had climbed 36% to $90.90 per barrel, while Brent gained 27%. With tanker traffic through the Strait of Hormuz effectively halted, approximately 20 million barrels per day — roughly 20% of global daily oil consumption — was at risk.
Iran launched drone strikes on the US Embassy in Saudi Arabia and hit key Saudi refineries and Qatari LNG facilities, disrupting approximately 20% of global LNG supply. Energy analysts estimated that roughly 9 million barrels per day of oil had been removed from the market through facility damage or precautionary measures, warning of an “extreme deficit” scenario.
WTI原油周涨幅
（截至2026.03.07）
WTI原油结算价
（2026.03.07）
从市场移除的石油
（桶/天）
经霍尔木兹海峡
的全球石油占比
5.5 定价范式的四重逆转
What the financial market data above reveals is not the impact of a single event, but a structural reversal of the pricing paradigm itself.
逆转#1：软资产 → 硬资产
Capital is migrating toward physically scarce resources — gold, non-ferrous metals, energy. All-time high gold prices, surging oil, and rising base metal prices provide empirical proof.
逆转#2：虚拟经济 → 实体经济
$2 trillion evaporated from software stocks versus rising prices in energy, commodities, and defense stocks. Physical manufacturing capability is emerging as a core national security asset.
逆转#3：美元信用 → 实物信用
Accelerating central bank gold purchases, a downward trend in the US Dollar Index, and movements by foreign investors to hedge or exit US assets are being observed.
逆转#4：信息-虚拟价值 → 物理世界价值
The more AI advances, the more demand surges for physical infrastructure — electricity, chips, data centers, cooling systems. Big Tech’s hundreds of billions in capital expenditures prove this.
核心命题：技术对资本的全面压制——四重战争框架
6.1 速度战：时间不对称的压制
According to John Boyd’s OODA Loop theory, the side whose Observe → Orient → Decide → Act cycle is faster than the enemy’s dominates the engagement. AI’s OODA cycle operates on a scale of hours to days, whereas capital’s cycle runs on a scale of weeks to months. When this speed differential exceeds a factor of ten, the slower side loses the capacity for command itself.
The case of approximately $285 billion in software sector market capitalization evaporating within days of Anthropic’s January 30 release of the Claude Cowork plugins provides empirical evidence. In a capital market structure that makes decisions based on quarterly earnings reports and analyst notes, an environment in which disruptive new capabilities are released daily is fundamentally impossible to defend against.
6.2 多模态机动战：全维度同步打击
AI advances simultaneously across all modalities — text, code, image, video, audio, reasoning, law, medicine, finance. Capital’s defense systems are organized by industry, sector, and asset class, but AI ignores these boundaries. The chain reaction in which the collapse of SaaS companies cascades into private credit deterioration, which in turn expands into a broader credit crunch, demonstrates that capital’s sector-by-sector defenses are powerless against AI’s all-dimensional assault.
6.3 精确斩首打击：消灭商业模式的存在理由
AI’s decapitation strikes directly negate the core premises of existing business models.
| 商业模式 | 核心前提 | AI的否定 |
|---|---|---|
| SaaS SaaS订阅模式 | “企业必须购买专业软件” | 通用AI执行专业软件的功能 |
| Credit Private credit | “高增长科技公司的未来收入可预测” | AI使科技公司的收入预测本身失效 |
| Jobs White-collar employment | “信息处理需要受过高等教育的人类” | AI短路信息处理的中间节点 |
| Seats Seat-based pricing | “价值随用户数量扩展” | AI Agent将人类席位减少90% |
Goldman Sachs’ Ben Snider comparing software’s future to the newspaper industry precisely captures the nature of this decapitation strike: not weakening, but the complete annihilation of the reason for existence. The newspaper industry was replaced by digital media not because newspaper quality declined, but because the newspaper’s very raison d’être — as an intermediary medium delivering information to readers — ceased to exist.
6.4 围困战：摧毁信用扩张的基础
Three core infrastructural pillars of the capital system are under simultaneous attack.
Destruction of the command center — The price discovery mechanism of the capital system is premised on analysts being able to keep pace with change. AI’s daily evolution invalidates this premise. While analysts are parsing quarterly reports, AI has already transformed the fundamental structure of the next industry.
Severing the supply lines — In the capital supply chain — middle-class stable income → consumer credit → corporate revenue → corporate lending → financial products → leverage — when AI attacks the first node (middle-class employment), the entire chain weakens in cascade. Workday’s announcement of an 8.5% workforce reduction, citing AI-driven efficiency gains, is a concrete example of this supply-line severance.
Disrupting the communications system — The capital market’s credit assessment system is based on historical data and linear extrapolation. In an environment where AI non-linearly transforms industrial structures, credit assessments based on historical data systematically produce errors. The BlackRock HPS fund redemption crisis is a direct consequence of this communications disruption.
技术在物理战争中的压制：军事行动与经济战场的统一逻辑
7.1 伊朗-以色列冲突（2026年3月）
The Iran-Israel conflict of late February to early March 2026 exemplified the modern integration of military and economic battlefield logic. The US “Operation Epic Fury” and Israel’s “Operation Roaring Lion” were launched simultaneously, striking Iran’s nuclear facilities, missile bases, and command-and-control centers.
The defining characteristic of this conflict was multi-domain simultaneous execution. Operations were conducted simultaneously across six domains: cyber, space, air, maritime, ground, and electronic warfare. Iran’s retaliation likewise targeted multiple objectives simultaneously — the Israeli mainland, US bases in the Persian Gulf, the US Embassy in Saudi Arabia, Saudi refineries, and Qatari LNG facilities. Nearly every country in the Middle East sustained missile or drone strikes.
7.2 军事与经济战场的结构同构性
| 战争类型 | 军事战场 | 经济战场 |
|---|---|---|
| Speed | Command issued → target struck: hours | AI plugin released → $285B market cap erased: days |
| Maneuver | Simultaneous operations across cyber, space, air, maritime, ground, and electronic domains | Simultaneous advance across text, code, image, law, medicine, finance, and all modalities |
| Decapitation | Direct elimination of supreme leader, paralyzing command structure | Annihilation of the very reason a business model exists |
| Siege | Blockade of refineries, LNG facilities, straits → strategic paralysis | Attack on middle-class employment, price discovery, credit assessment → destruction of capital system foundations |
7.3 与1991年海湾战争的比较：范式转移的历史先例
After electronic warfare paralyzed Iraq’s air defense systems within hours during Operation Desert Storm in 1991, global military development aligned around electronic warfare capabilities. The 2026 Iran operation signals the beginning of a new paradigm in which AI dominates the entire operational chain — intelligence collection, target selection, attack planning, execution, and damage assessment.
There is a critical difference: in the electronic warfare era, the capability gap could be narrowed by purchasing hardware like radar and jamming equipment. But AI’s core capabilities — data, algorithms, and talent — cannot be simply bought. This means the technology gap will be wider, more persistent, and more self-reinforcing.
结论：2026年作为历史分水岭的意义
8.1 技术压倒资本的三大论证
论证①：速度不对称
AI’s daily evolution versus capital’s quarterly response. This speed differential of tenfold or more causes capital to lose the capacity for price discovery itself. In a reality where a single AI plugin release can evaporate hundreds of billions of dollars in market capitalization within days, the quarterly earnings-based price discovery mechanism is structurally impotent.
论证②：碎片化破坏的不可预测性
AI’s destruction operates in a state of quantum superposition. Which industry, at what point, and at what scale it will be destroyed — the market cannot estimate in advance. The $2 trillion evaporation in SaaS, the private credit redemption crisis, the erosion of professional services in law, accounting, and translation — all proceed asynchronously yet simultaneously.
论证③：信用基础的侵蚀
What AI attacks is precisely the middle-class white-collar employment that forms the basis of capital’s credit expansion. Workday’s 8.5% workforce reduction, Atlassian’s first-ever decline in enterprise seats, and widespread seat compression across software companies are early signals of supply-line severance.
8.2 在历史脉络中的定位
| 技术革命 | 破坏对象 | 资本的回应 | 结果 |
|---|---|---|---|
| 蒸汽机 | 封建地主贵族 | 转化为工业资本 | 资本控制重建 |
| 电力 | 传统制造业 | 通过金融市场控制 | 资本控制维持 |
| 互联网 | 中间分销 | 通过风险资本选择性支持 | 资本控制维持 |
| AI | 资本权力结构本身 | 逃向黄金/实物资产，赎回限制 | 资本失去控制？ |
In every previous technological revolution, capital reconstructed its control after an initial period of disruption. However, in the AI era, two decisively different factors exist. First, the marginal cost of technological creation converges to zero. A single developer wielding AI tools can build software in hours that would have required an entire team half a year just months ago. Second, technologists’ dependence on capital is plummeting. Whereas large-scale capital investment was once indispensable for technological realization, the democratization of AI tools is causing startup costs and development expenses to fall precipitously.
8.3 新权力结构
In the former power dynamic, technology petitioned capital: “Please invest in me.” In the current dynamic, capital petitions technology: “What should I invest in so I don’t become obsolete?” The fact that the Big Four tech companies are investing more than $660 billion in AI infrastructure in 2026 is not a voluntary choice but the result of structural compulsion — they cannot exit the AI race.
8.4 未解之问
This paper has analyzed the mechanisms by which technology overwhelms capital, but subsequent questions remain open. After technology overwhelms capital, what new equilibrium will form between technology and political power? As the Lin Junyang case illustrates, in East Asia a triple ceiling of corporate power, capital control, and government co-optation structurally constrains technologists’ autonomy. In the United States, how will AI companies’ power coexist with national security and democratic governance? This is a question of an entirely different dimension, requiring separate in-depth research.
参考事件时间线
参考文献与数据来源
[1] 36氪, “林俊旸出走，阿里千问告别英雄时代”, 2026.03.05
[2] 36氪 / 智能涌现, “千问模型负责人林俊旸提出离职，阿里高管紧急答疑”, 2026.03.06
[3] 晚点LatePost, “林俊旸离职风波始末”, 2026.03.06
[4] 观察者网, “深夜地震！阿里千问大模型负责人突曝离职”, 2026.03.04
[5] CBS News, “What is the price of gold today: March 2, 2026?”, 2026.03.02
[6] J.P. Morgan Global Research, “Gold price predictions”, 2026
[7] World Gold Council, “Gold Market Commentary: February 2026”, 2026.03.05
[8] Bloomberg / Reuters, “BlackRock $26 Billion Private Credit Fund Limits Withdrawals”, 2026.03.06
[9] InvestmentNews, “BlackRock curbs redemptions at HPS private credit fund”, 2026.03.06
[10] Al Jazeera, “Oil prices rise sharply after US, Israeli attacks on Iran”, 2026.03.02
[11] CNN, “Oil surges and stock futures sink as war in Iran threatens crude supply”, 2026.03.01
[12] CNBC, “Middle East conflict poses fresh test to central banks”, 2026.03.04
[13] PBS / AP, “Oil and gas prices rise rapidly as Iran war escalates”, 2026.03.07
[14] SaaStr, “The 2026 SaaS Crash: It’s Not What You Think”, 2026.01.30
[15] Bain & Company, “Why SaaS Stocks Have Dropped”, 2026.02
[16] CNBC, “AI fears pummel software stocks”, 2026.02.06
[17] Serenities AI, “SaaSpocalypse: AI Tools Triggered $285B Software Crash”, 2026.02.09
[18] Digital Applied, “The SaaSpocalypse: AI Agents Disrupting Software Industry”, 2026.02
[19] CNBC, “Coupang CEO resigns over data breach”, 2025.12.10
[20] Fortune, “Coupang CEO resigns over historic South Korean data breach”, 2025.12.10
[21] UBS Global, “Tech sell-off highlights need for diversification”, 2026.02.04
[22] CNN, “Nasdaq and S&P just had their worst month since March”, 2026.02.27
[23] Nasdaq.com, “Stocks Retreat as Tech Stocks Fall and the US Labor Market Weakens”, 2026.02