The Market Is Pricing the Future Before It Arrives: Tesla, Broadcom, Meta, Robinhood and the New Economics of AI, Autonomy and Tokenisation

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

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.  

From Cybercabs to AI Chips and Tokenised Finance: What Tesla, Broadcom, Meta and Robinhood Reveal About the Next Market Cycle

The Market Is Pricing the Future Before the Future Arrives: Tesla, Broadcom, AI, Oil, Meta and Robinhood

By Zion Zhao, 赵峻慷

The most interesting market story today is not that investors have stopped caring about risk.

It is that they are becoming increasingly selective about which risks matter most.

Tesla is launching purpose-built Cybercabs. Broadcom is producing extraordinary artificial intelligence growth. Oil remains elevated. Treasury yields are restrictive. Meta is expanding its artificial intelligence ecosystem. Robinhood is moving deeper into prediction markets and tokenised finance.

Yet equities continue to show remarkable resilience.

At first glance, that looks irrational.

It is not.

Markets do not simply respond to whether the news sounds good or bad. They price future cash flows, probabilities, interest rates, competitive positioning and expectations. Oil can rise while stocks rally. A company can report extraordinary earnings and still fall. A technologically transformative business can remain overvalued. A mature platform can appear boring while quietly building one of the strongest competitive positions in artificial intelligence.

The real investment question is therefore not whether the future is exciting.

It is how much of that future has already been priced.

Tesla’s Cybercab Is Real, but Scale Is the Test

Tesla’s Cybercab deployment matters because the company is moving beyond prototypes and into real-world operations.

Tesla began offering limited Cybercab rides in Austin using its two-seat autonomous vehicle without a steering wheel or pedals. Texas records showed 420 Tesla autonomous vehicles registered in the state, including 45 Cybercabs at the time of deployment (Kirkham et al., 2026).

Forty-five vehicles are simultaneously underwhelming and significant.

They are underwhelming relative to the scale implied by some of the most aggressive Tesla forecasts. They are significant because Tesla has moved another step from concept toward commercial autonomous transport.

The problem is that autonomous technology and autonomous economics are not the same thing.

Tesla itself states that Full Self-Driving Supervised does not make a vehicle autonomous. The driver remains responsible and must actively supervise the system (Tesla, 2026a).

That distinction matters because Elon Musk has historically been directionally ambitious but frequently premature on autonomy timelines.

In 2016, Tesla targeted a fully autonomous Los Angeles to New York demonstration by the end of 2017. That did not occur on schedule. Tesla later highlighted a verified cross-country Full Self-Driving Supervised journey in March 2026, but it remained a supervised trip rather than the fully autonomous demonstration originally envisioned (Tesla, 2026b).

This history should not be used to argue that Tesla cannot solve autonomy.

It should be used to distinguish technological progress from timeline certainty.

Tesla can eventually dominate autonomous transport and still take significantly longer than many investors expect.

Regulation, Liability and Insurance Cannot Be Vibe-Coded Away

The day after Tesla’s Cybercab deployment, the U.S. National Highway Traffic Safety Administration opened an Audit Query into Tesla’s self-certification of the Cybercab.

The inquiry focused on how Tesla determined that a vehicle without conventional human controls complied with applicable federal safety standards (National Highway Traffic Safety Administration, 2026).

This illustrates a larger truth.

Robotaxi deployment is not merely an engineering problem.

It involves federal vehicle standards, state regulation, municipal permitting, insurance rules, accident liability, cybersecurity, remote support, data reporting and consumer protection.

Academic research identifies safety, liability, cybersecurity and privacy as separate governance challenges for autonomous vehicles (Taeihagh & Lim, 2019). Research also suggests autonomous systems may reduce accidents under many conditions while remaining weaker in specific edge cases (Abdel-Aty & Ding, 2024).

The technology can therefore be impressive without the national deployment model being easy.

The hardest phase is not proving that a car can drive itself.

It is proving that thousands or millions of cars can operate safely, legally and profitably across different cities, weather conditions and legal jurisdictions.

Tesla Versus Uber Is Probably the Wrong Framework

One of the weakest assumptions in the robotaxi debate is that Tesla must destroy Uber.

Uber is not merely a driver network.

It is a transportation marketplace with dispatch, routing, payments, customer accounts, merchant relationships, delivery, logistics, insurance infrastructure and enormous user scale.

In the second quarter of 2026, Uber reported 3.9 billion trips, 208 million monthly active platform consumers, $58 billion of gross bookings and $2.8 billion of quarterly free cash flow (Uber Technologies, 2026a).

That is not a company waiting passively for extinction.

Platform economics also matters.

Ride-hailing benefits from network effects. More riders attract more vehicle supply. More supply reduces waiting times and improves convenience. Autonomous vehicles may eventually replace human drivers, but they do not necessarily eliminate the marketplace coordinating demand and supply (Castillo & Mathur, 2023).

Uber is already pursuing autonomous vehicle partnerships.

That makes a hybrid future more plausible than a winner-takes-all outcome.

Tesla may operate its own network. Uber may aggregate autonomous fleets. Professional operators may purchase vehicles and distribute them across multiple platforms. Different systems may dominate different geographies.

The future can be autonomous without being monopolistic.

Tesla’s Valuation Is Really a Portfolio of Options

Tesla’s valuation is difficult to interpret through traditional automotive metrics because investors are not merely paying for vehicle sales.

They are paying for future optionality.

That includes autonomous transport, humanoid robotics, energy storage, artificial intelligence, software and manufacturing automation.

For the quarter ended June 30, 2026, Tesla reported $28.24 billion of revenue but only $398 million of operating income (Tesla, 2026c).

That helps explain why the valuation debate is really a probability debate.

There is nothing irrational about assigning value to businesses that do not yet exist at scale.

The mistake is assigning certainty to them.

“Tesla may create an enormous global robotaxi business” is an investable scenario.

“Tesla will inevitably dominate global ride-hailing” is still a forecast.

The difference is crucial.

Why Did Stocks Rally With Oil Near $90?

Oil still matters.

Higher energy prices can raise transportation costs, pressure consumers, increase inflation expectations and influence monetary policy.

By September 7, West Texas Intermediate crude was around $92 per barrel while Brent traded near $97 amid renewed geopolitical concerns (Tan & Li, 2026).

Yet stocks rallied.

The reason is that the relationship between oil and equities is not mechanical.

Research by Kilian (2009) showed that the economic impact of oil depends on the source of the shock. Oil rising because global demand is strong differs materially from oil rising because supply is disrupted.

One source claim also required correction.

The discussion suggested that global oil consumption exceeded production by roughly six million barrels per day. Contemporary estimates were materially lower, generally around 1.65 million to 3.5 million barrels per day (Saha, 2026).

That distinction matters.

A structural six-million-barrel daily deficit would imply a considerably more severe market imbalance.

The Federal Reserve Helped the Rally

Equities were also responding to monetary policy.

Federal Reserve Governor Christopher Waller argued on September 3 that rates could remain unchanged if disinflation continued, although he left open the possibility of a hike if inflation reaccelerated (Waller, 2026).

Markets immediately reduced the implied probability of a September rate increase.

That matters because long-duration technology stocks are particularly sensitive to discount rates.

Lower expected rates increase the present value investors assign to future profits.

The market was therefore not saying that oil did not matter.

It was saying that, on that particular day, the combination of earnings strength and reduced rate-hike risk mattered more.

Broadcom’s Earnings Were Exceptional, but Expectations Were Higher

Broadcom’s fiscal third-quarter 2026 results were one of the strongest confirmations that artificial intelligence infrastructure spending remains extremely powerful.

One major numerical correction is necessary.

The underlying discussion attributed approximately $96 billion of revenue and $89 billion of data-centre revenue to Broadcom. Those were not Broadcom’s reported figures.

Broadcom actually generated approximately $29.6 billion of quarterly revenue, up 86 percent year over year. Artificial intelligence semiconductor revenue reached $16.7 billion, up 221 percent year over year (Broadcom Inc., 2026).

The company also guided to approximately $21.7 billion of AI semiconductor revenue in the following quarter.

Those numbers are extraordinary.

Yet Broadcom’s stock fell because its guidance was slightly below already elevated Wall Street expectations.

This illustrates one of the most important principles in markets.

Stocks trade expectations, not report cards.

A great company can produce outstanding earnings and still fall if investors expected even more.

Broadcom Is Becoming an AI Infrastructure Toll Road

Broadcom’s opportunity extends beyond conventional semiconductors.

The AI data centre increasingly requires custom accelerators, high-speed networking, switches, interconnect technology and specialised infrastructure.

Hyperscalers are increasingly designing chips around their own workloads.

That does not eliminate Nvidia.

It expands the AI infrastructure market.

Nvidia remains dominant in general-purpose accelerated computing. Broadcom increasingly benefits from custom silicon and networking.

Both can win.

The real risk is concentration.

If a significant portion of future Broadcom growth depends on a relatively small group of hyperscalers and frontier AI laboratories, investors must evaluate whether those customers ultimately earn adequate returns on their enormous capital expenditure.

AI Is Probably Early, but Early Does Not Mean Risk-Free

Artificial intelligence is increasingly embedded in real workflows.

Document summarisation, coding, translation, research, customer service and enterprise automation may sound less exciting than humanoid robots, but they generate real productivity value.

The fact that AI improves rapidly also matters.

Technology platforms rarely stop at their first generation.

Cloud computing, smartphones and microprocessors all improved through repeated iterations.

AI is likely to follow the same path.

But the existence of technological progress does not guarantee attractive investment returns.

Transformative technologies can still experience periods of overinvestment.

Railroads changed transportation.

The internet changed commerce.

Neither prevented investors from occasionally allocating capital at irrational valuations.

The next AI debate is therefore less about whether AI works and more about who captures the economics.

Chip designers?

Cloud companies?

Model laboratories?

Software developers?

Businesses using AI to reduce costs?

Consumers receiving lower prices?

The answer is likely some combination.

AI Financing Is Becoming More Complex

Another important correction concerns financing.

The AI infrastructure boom is not funded purely through equity.

Equity remains important. Anthropic raised enormous private financing rounds in 2026.

However, the broader ecosystem increasingly uses debt, project finance, guarantees, leases and revolving credit facilities.

Reuters reported that Nvidia agreed to provide up to $105 billion of guarantees relating to an OpenAI-linked data-centre project, while Anthropic was reportedly pursuing a multibillion-dollar revolving credit facility (Rooprai & Bautzer, 2026; Wang, 2026).

This matters.

Equity losses are absorbed by shareholders.

Debt creates contractual obligations.

The more leveraged the AI infrastructure ecosystem becomes, the more important utilisation rates, cash generation and customer economics become.

Meta’s Biggest AI Advantage May Be Distribution

Meta may have one of the strongest AI positions precisely because it does not need to build distribution from scratch.

Facebook, Instagram, WhatsApp and Messenger already reach billions of people.

In the second quarter of 2026, Meta generated $60.8 billion of revenue, with advertising representing the overwhelming majority (Meta Platforms, 2026).

That means Meta can monetise artificial intelligence indirectly.

It can improve ad conversion, recommendations, customer service, commerce, creator tools and business messaging without necessarily charging users directly for every AI interaction.

Its principal risk is cost.

Meta’s operating margin fell while expenses rose significantly, and Reality Labs continued generating multibillion-dollar quarterly losses.

The question is not whether Meta can afford the spending.

It is whether the incremental return justifies the capital.

Meta’s major youth-safety settlement has also reduced an important legal overhang, but legal risk has not disappeared entirely. Other litigation and regulatory proceedings remain active.

The correct conclusion is therefore not that Meta has become risk-free.

It is that its core monetisation engine remains extraordinarily powerful.

Robinhood Is Becoming More Than a Brokerage

Robinhood is evolving into a broader consumer financial platform.

Its ecosystem increasingly spans equities, options, cryptocurrencies, retirement accounts, prediction markets and tokenised assets.

In Q2 2026, Robinhood reported $1.31 billion of net revenue, while event-contract revenue reached $156 million (Robinhood Markets, 2026a).

That changes the strategic story.

Robinhood is no longer simply competing for stock trades.

It is competing for financial engagement.

Prediction markets increase frequency of interaction. Crypto expands product breadth. Robinhood Chain creates an onchain infrastructure layer.

Yet tokenisation requires careful legal interpretation.

A tokenised stock is not necessarily direct ownership of the underlying equity.

Robinhood itself discloses that some Stock Tokens represent tokenised debt securities providing economic exposure rather than legal ownership of the underlying shares (Robinhood Markets, 2026b).

The U.S. Securities and Exchange Commission has similarly emphasised that securities laws continue to apply regardless of whether an instrument is represented on a blockchain.

Technology changes infrastructure.

It does not abolish law.

The Real Investment Lesson

Tesla, Broadcom, Meta and Robinhood appear different, but investors are valuing them through a common framework.

They all contain optionality.

Tesla offers autonomy and robotics.

Broadcom offers AI infrastructure.

Meta offers AI monetisation through distribution.

Robinhood offers the convergence of brokerage, crypto, prediction markets and tokenisation.

Optionality has value because outcomes are uncertain.

The danger begins when possibility is mistaken for certainty.

Tesla has demonstrated that Cybercab deployment is real. It has not demonstrated national fleet economics.

Broadcom has demonstrated extraordinary AI demand. It has not proved that current spending growth continues indefinitely.

Meta has demonstrated unparalleled digital monetisation. It has not proved that every AI investment will earn superior returns.

Robinhood has demonstrated expanding engagement. It has not proved that every onchain or prediction-market product will survive unchanged across jurisdictions.

That is the central discipline investors need today.

Separate what is possible from what has already been demonstrated.

Conclusion

The market does care about oil.

It cares about bond yields.

It cares about inflation.

But it also cares about earnings, technological productivity, platform economics, distribution and future cash flows.

Tesla can revolutionise transportation without immediately destroying Uber.

Broadcom can produce spectacular AI growth while its share price falls because expectations were even higher.

Meta can remain exposed to regulatory and spending risks while possessing one of the strongest AI distribution platforms in the world.

Robinhood can help move finance onchain while tokenised products remain subject to conventional securities law.

Artificial intelligence can transform the global economy while parts of the AI investment cycle eventually become overbuilt.

These ideas are not contradictory.

They are what serious investing looks like.

The easiest narratives are usually the most dangerous.

The investor’s job is not to choose between optimism and pessimism.

It is to assign probabilities, understand valuation and determine whether tomorrow can justify today’s price.

That is where the real analysis begins.

Important Disclaimer

This article is provided solely for general information, education and commentary. It does not constitute financial, investment, trading, legal, tax, property or other professional advice, nor any offer, solicitation or recommendation to buy, sell or hold any security, digital asset or other investment. Views expressed are analytical opinions based on information believed to be reliable at the time of writing. Forecasts, valuations and forward-looking statements are inherently uncertain. Readers should conduct independent due diligence and consult suitably qualified professional advisers before making financial or investment decisions.

References

Abdel-Aty, M., & Ding, S. (2024). A matched case-control analysis of autonomous vs. human-driven vehicle accidents. Nature Communications, 15, 4931. https://doi.org/10.1038/s41467-024-48526-4

Broadcom Inc. (2026, September 2). Broadcom Inc. announces third quarter fiscal year 2026 financial results and quarterly dividend. Broadcom Investor Relations.

Castillo, J. C., & Mathur, S. (2023). Matching and network effects in ride-hailing. AEA Papers and Proceedings, 113, 244–247. https://doi.org/10.1257/pandp.20231069

Kilian, L. (2009). Not all oil price shocks are alike: Disentangling demand and supply shocks in the crude oil market. American Economic Review, 99(3), 1053–1069. https://doi.org/10.1257/aer.99.3.1053

Kirkham, C., Shirouzu, N., & Roy, A. (2026, September 3). Tesla starts Cybercab rides in Austin, drawing U.S. safety agency interest. Reuters.

Meta Platforms, Inc. (2026, July 29). Meta reports second quarter 2026 results. Meta Investor Relations.

National Highway Traffic Safety Administration. (2026, September 4). NHTSA opens investigation into Tesla Cybercab self-certification following Austin deployment. U.S. Department of Transportation.

Robinhood Markets, Inc. (2026a, July 29). Robinhood reports second quarter 2026 results. Robinhood Investor Relations.

Robinhood Markets, Inc. (2026b, July 1). Robinhood accelerates global expansion with Robinhood Chain mainnet, Stock Tokens, agentic trading and new suite of DeFi products.

Rooprai, A., & Bautzer, T. (2026, August 17). Nvidia to provide up to $105 billion guarantee for OpenAI's Ohio data center. Reuters.

Saha, S. (2026, August 31). Oil to hold above $80 a barrel as Middle East supply risks persist. Reuters.

Taeihagh, A., & Lim, H. S. M. (2019). Governing autonomous vehicles: Emerging responses for safety, liability, privacy, cybersecurity, and industry risks. Transport Reviews, 39(1), 103–128. https://doi.org/10.1080/01441647.2018.1494640

Tesla, Inc. (2026a). Full Self-Driving (Supervised). Tesla.

Tesla, Inc. (2026b, March 25). A cross country trip with Full Self-Driving (Supervised). Tesla.

Tesla, Inc. (2026c). Quarterly report for the period ended June 30, 2026 [Form 10-Q]. U.S. Securities and Exchange Commission.

Uber Technologies, Inc. (2026a). Uber announces results for second quarter 2026. Uber Investor Relations.

U.S. Securities and Exchange Commission. (2026). Statement on tokenized securities.

Waller, C. J. (2026, September 3). The economic outlook and some comments on my policy communication. Board of Governors of the Federal Reserve System.

Wang, E. (2026, September 4). Anthropic IPO launch shifts toward mid-October, sources say. Reuters.

Autonomy, AI and Onchain Finance Are Reshaping Markets, but Investors Still Need to Separate Possibility From Proven Economics

Tesla, Broadcom, Meta and Robinhood show that markets are increasingly pricing future optionality, not merely current earnings. Autonomy, artificial intelligence and tokenisation offer enormous potential, but regulation, capital intensity and execution remain decisive. The key discipline is separating technological possibility from proven economics, then asking whether tomorrow justifies today’s valuation.

If you are considering Singapore property for investment, relocation, education planning, family wealth structuring or long-term asset allocation, I believe the quality of your adviser matters beyond knowing projects, prices and transactions.

Real estate does not exist in isolation. Interest rates, global liquidity, geopolitics, equity markets, currencies, regulations and capital flows can all influence property demand, financing conditions and portfolio decisions.

As a Singapore real estate salesperson, I dedicate hours each day to studying macroeconomics, global affairs, equities, digital assets, asset allocation, Singapore property law and market developments, and to writing research-based essays such as this. I believe proper due diligence and a broader cross-asset perspective help clients make more informed, disciplined decisions.

For Family Offices, 13U/13O structures, GIP and VCC participants, UHNW individuals, institutional investors, international families, China-based clients, Southeast Asian investors,陪读家长 and overseas-study families, I provide Singapore property advisory focused on long-term suitability, risk awareness and strategic portfolio positioning.

Singapore real estate can, where appropriate, complement a diversified portfolio through relatively lower day-to-day price volatility, potential capital appreciation and rental income, but outcomes are never guaranteed.

If you value thoughtful analysis, professionalism and long-term perspective, I would be honoured to assist with your Singapore property journey.

特斯拉无人出租车真的来了，博通AI收入暴增221%：市场为什么敢无视高油价，继续为未来买单？

市场正在为尚未到来的未来定价：特斯拉、博通、人工智能、油价、Meta 与 Robinhood

作者：Zion Zhao 赵峻慷

今天市场上最值得关注的现象，并不是投资者已经不再关心风险，而是市场正在变得更加挑剔，开始重新判断：究竟哪些风险才真正重要。

特斯拉正在部署 Cybercab 无人驾驶出租车。博通 Broadcom 正在交出极其强劲的人工智能业务增长。油价仍处于高位，美国国债收益率依然具有压制性。Meta 正不断扩大人工智能生态系统，Robinhood 也正进一步深入预测市场 Prediction Markets 与资产代币化 Tokenisation。

然而，股票市场依然展现出相当强的韧性。

乍看之下，这似乎并不合理。

其实并非如此。

市场并不会简单地根据“今天的消息是好还是坏”来定价。市场真正折现的是未来现金流、利率、竞争格局以及各种结果发生的概率。油价可以上涨，而股市同时走高。一家公司可以公布非常优秀的业绩，股价却下跌。一家拥有颠覆性技术的公司，估值依然可能过高。一个看似已经成熟的平台，也可能在市场低估之下悄然建立起极其强大的结构性优势。

真正的问题从来不是未来是否令人兴奋。

真正的问题是：市场今天的价格，已经提前计入了多少未来？

特斯拉 Cybercab 已经成为现实，但真正的考验是规模化

特斯拉 Cybercab 的落地具有重要意义，因为这代表公司又向商业化自动驾驶交通迈进了一步。

特斯拉开始在美国得州奥斯汀 Austin 投入没有方向盘、没有踏板的专用 Cybercab 车辆。当时，得州登记的特斯拉自动驾驶车辆共有 420 辆，其中包括 45 辆 Cybercab（Kirkham et al., 2026）。

45 辆车既可以说令人失望，也可以说意义重大。

与市场上最激进的特斯拉长期预测相比，这个数字显然很小。但它的重要性在于，特斯拉已经进一步从概念展示走向真实世界部署。

关键在于区分两件事：

技术能力，和商业经济性。

特斯拉自己也明确表示，Full Self-Driving (Supervised)，即“完全自动驾驶能力监督版”，并不会使车辆真正成为完全自动驾驶汽车，驾驶者仍必须持续监督系统（Tesla, 2026a）。

这一点之所以重要，是因为 Elon Musk 埃隆·马斯克过去对于自动驾驶时间表的判断，往往过于激进。

2016 年，特斯拉曾计划在 2017 年底前完成从洛杉矶到纽约的完全自动驾驶展示，但并未按时实现。

这并不意味着特斯拉无法最终解决自动驾驶问题。

它意味着投资者必须清楚区分：

技术正在进步，和技术一定会按照原先承诺的时间表实现，并不是同一件事。

Cybercab 部署后的第二天，美国国家公路交通安全管理局 National Highway Traffic Safety Administration，简称 NHTSA，就针对特斯拉对 Cybercab 所进行的自我安全认证 Self-Certification 启动了 Audit Query 审查程序。

这进一步说明，Robotaxi 无人驾驶出租车并不仅仅是工程问题。

它同时涉及车辆安全标准、州级监管、保险、事故责任、网络安全 Cybersecurity，以及运营监管与持续监督（National Highway Traffic Safety Administration, 2026; Taeihagh & Lim, 2019）。

真正最困难的问题，并不是证明一辆车能够自动驾驶。

而是证明成千上万辆自动驾驶汽车，可以在不同城市、道路环境和法律辖区中，安全、合法并且具有经济效益地运行。

“特斯拉对决 Uber”可能从一开始就是错误的问题

Robotaxi 投资叙事中最薄弱的假设之一，就是认为特斯拉最终一定要消灭 Uber。

Uber 并不只是由一群人类司机组成的平台。

它是一个完整的交通出行市场基础设施，拥有调度 Dispatch、路线规划 Routing、支付体系、客户关系、外卖配送、物流网络以及巨大的用户规模。

2026 年第二季度，Uber 共完成 39 亿次出行，月度活跃平台消费者 Monthly Active Platform Consumers 达到 2.08 亿，Gross Bookings 总预订额达到 580 亿美元（Uber Technologies, 2026）。

平台经济学 Platform Economics 在这里非常重要。

更多乘客会吸引更多运力，而更多运力又会提高车辆可获得性并缩短等待时间。自动驾驶汽车未来也许会替代大量人类司机，但它并不一定会消灭负责连接需求与供给的平台（Castillo & Mathur, 2023）。

未来完全可能出现多种模式并存。

特斯拉可以建立自己的 Robotaxi 网络。

Uber 可以聚合多家自动驾驶车队。

专业车队运营商也可以同时参与多个平台。

因此，未来完全可以是高度自动驾驶的，但并不一定是由单一公司垄断的。

特斯拉的估值，本质上也反映了这种不确定性。

投资者今天购买的，不只是电动车业务。

他们同时在为自动驾驶、人形机器人 Humanoid Robotics、人工智能、能源业务与软件平台的未来选择权 Optionality 付费。

这种估值逻辑并不一定不合理。

但未来选择权不等于未来必然兑现。

为什么油价接近 90 美元，股市仍然上涨？

油价当然仍然重要。

更高的能源价格可能推高运输成本，压缩消费者可支配收入，加剧通胀预期，并影响中央银行的货币政策。

然而，油价和股票市场之间并不是机械式的线性关系。

Kilian（2009）的研究指出，油价上涨对经济的影响，在很大程度上取决于这轮油价上涨究竟来自什么。

如果油价上涨是因为全球需求强劲，其经济含义，与因为供应中断而导致油价暴涨，并不相同。

原始讨论中还有一个数据需要修正。

其中提到全球石油需求可能每天比供给高出约 600 万桶。

但当时更可靠的市场估计，大约在每天 165 万至 350 万桶之间（Saha, 2026）。

这是一个非常重要的区别。

每天 600 万桶的长期结构性缺口，意味着比当时实际市场状况严重得多的供需失衡。

股市之所以能够在高油价背景下继续上涨，还有另一个原因：

货币政策预期发生了改善。

美国联邦储备委员会理事 Christopher Waller 克里斯托弗·沃勒表示，如果通胀继续放缓，美联储可以选择维持利率不变，从而降低了市场对于近期再次加息的预期（Waller, 2026）。

所以，市场并不是在说“油价不重要”。

市场真正表达的是：

在那个交易日，降低的加息风险与强劲企业盈利，比高油价的负面影响更加重要。

博通证明，人工智能需求是真实存在的

Broadcom 博通 2026 财年第三季度业绩，再次证明人工智能基础设施支出仍然极其强劲。

这里同样需要做一个重要的数据修正。

原始讨论中一度将约 960 亿美元季度营收，以及约 890 亿美元数据中心收入归于博通。

这并不是博通实际公布的数据。

博通当季实际营收约为 296 亿美元，同比增长 86%。

其中人工智能半导体 AI Semiconductor Revenue 收入达到 167 亿美元，同比增长 221%（Broadcom Inc., 2026）。

这是非常惊人的增长速度。

但博通股价在财报后依然下跌。

为什么？

因为公司给出的未来指引，没有明显超过已经极度乐观的华尔街预期。

这揭示了金融市场最重要的规律之一：

股票交易的不是成绩单，而是预期差。

一家优秀公司可以公布非常优秀的财报，股价依然下跌，因为市场之前期待的是更加优秀。

博通目前在定制 AI 芯片 Custom AI Silicon、高速网络以及数据中心基础设施中的重要性也越来越高。

NVIDIA 英伟达依然在通用加速计算 Accelerated Computing 领域占据主导地位，而博通则受益于大型云计算与科技企业根据自身工作负载，设计更加专业化芯片的趋势。

两家公司完全可以同时受益。

真正需要关注的风险，是这些推动 AI 资本开支的大客户，最终能否从如此巨额的投入中获得足够高的资本回报率。

人工智能仍然处于早期，但“早期”并不等于没有风险

人工智能已经逐渐融入真实的工作流程。

代码开发、研究、翻译、客服、文档分析以及企业自动化，都正在成为实际应用。

这意味着 AI 已经具有真实的经济价值。

但一项技术足够伟大，并不代表围绕它进行的每一笔投资都是合理的。

铁路改变了交通运输。

互联网改变了商业。

但这并没有阻止投资者在某些阶段以过高价格投入过多资本。

所以，下一阶段真正重要的 AI 问题已经不再只是：

“AI 到底有没有用？”

而是：

谁最终能够获取其中的经济利润？

是芯片设计商？

云计算公司？

大模型实验室 Frontier AI Labs？

软件开发商？

利用 AI 降低成本的传统企业？

还是因为价格下降而受益的消费者？

答案很可能是以上多方共同分享。

与此同时，人工智能基础设施融资也正在变得更加复杂。

这个产业的资本开支已经不只是依靠股权融资 Equity Financing，也越来越多地采用债务融资 Debt、担保 Guarantees、租赁 Leasing 与项目融资 Project Finance。

这意味着，当整个系统变得更加资本密集 Capital Intensive 之后，算力利用率 Utilisation Rate、现金流以及客户自身的商业经济性，就会变得越来越重要。

Meta 和 Robinhood 说明了为什么“分发能力”如此重要

Meta 最大的人工智能优势，可能不是它拥有某一个最强的大模型，而是它已经拥有极其强大的分发网络 Distribution。

Facebook、Instagram、WhatsApp 和 Messenger 已经触达数十亿用户。

因此，Meta 不需要从零建立用户基础，就可以把 AI 部署到广告、推荐系统、即时通讯、创作者工具以及商业服务当中。

2026 年第二季度，Meta 实现约 608 亿美元营收，充分体现了其核心商业化引擎的强大能力（Meta Platforms, 2026）。

Meta 甚至不一定需要对每一次 AI 使用直接收费。

它可以通过提高广告转化率、提升推荐效率、加强用户参与度、改善企业沟通以及提高商业交易效率来间接变现。

真正的风险是成本。

AI 基础设施以及 Reality Labs 现实实验室仍然需要大量资本投入。

所以，真正的问题不是 Meta “有没有钱烧”。

而是：

下一美元投入，能否产生足够高的边际回报？

Meta 在青少年社交媒体相关案件中达成的大型和解，也降低了一个重要法律风险，但并不意味着公司已经完全摆脱诉讼与监管压力。

因此，更准确的判断应该是：

Meta 并没有变成一家没有风险的公司。

但它的核心商业化能力仍然极其强大。

Robinhood 则代表另一种未来选择权。

它正在从传统网络券商，逐渐发展成为更加完整的消费者金融平台。

其业务已经横跨股票、期权、加密资产、退休账户、预测市场和代币化资产。

2026 年第二季度，Robinhood 实现 13.1 亿美元净营收，其中 Event Contracts 事件合约业务贡献约 1.56 亿美元收入（Robinhood Markets, 2026）。

这意味着 Robinhood 已经不只是争夺股票交易用户。

它正在争夺用户的整体金融活动与金融注意力。

然而，资产代币化需要极高的法律精确性。

Tokenised Stock 代币化股票，并不一定等同于直接持有真实股票。

Robinhood 自己也披露，部分 Stock Tokens 股票代币实际上属于 Tokenised Debt Securities 代币化债务证券，为投资者提供与相关股票挂钩的经济敞口 Economic Exposure，但并不代表投资者直接拥有相应股票的法律所有权（Robinhood Markets, 2026）。

因此：

技术可以改变金融基础设施，但技术不会消灭证券法。

真正的投资启示

Tesla、Broadcom、Meta 和 Robinhood 看似属于不同产业，但市场对这几家公司进行估值时，其实使用了一个共同逻辑：

未来选择权 Optionality。

Tesla 提供自动驾驶与机器人选择权。

Broadcom 提供 AI 基础设施选择权。

Meta 提供通过超级分发网络进行 AI 商业化的选择权。

Robinhood 提供券商、加密资产、预测市场以及代币化金融融合的选择权。

之所以未来选择权有价值，正是因为结果并不确定。

真正的危险，是当市场把“可能发生”错误地理解成“一定发生”。

Tesla 已经证明 Cybercab 真实存在并开始部署。

但它还没有证明全国性 Robotaxi 车队的商业经济性。

Broadcom 已经证明 AI 需求非常强劲。

但它还没有证明当前这种资本开支增长速度可以无限持续。

Meta 已经证明其数字广告商业化能力极其强大。

但它还没有证明每一笔 AI 投资都能够产生超额回报。

Robinhood 已经证明其用户金融参与度正在不断扩大。

但它还没有证明每一个代币化金融产品和预测市场产品，都能在没有监管阻力的情况下大规模扩张。

因此，投资者真正需要做的，并不是简单选择“看多”或者“看空”。

而是：

为不同情景赋予概率，理解估值，并判断今天支付的价格，能否被未来真正兑现的盈利所证明。

这才是真正专业的投资分析开始的地方。

重要免责声明

本文仅用于一般信息、教育及市场评论用途，不构成任何形式的金融、投资、交易、法律、税务、房地产或其他专业意见，也不构成任何购买、出售、持有证券、数字资产或其他投资产品的要约、招揽或推荐。

文中观点属于基于截至撰写时被认为可靠资料所作出的分析性意见。所有预测、估值及前瞻性陈述均存在不确定性。历史表现不代表未来结果。投资者应根据自身目标、财务状况与风险承受能力进行独立尽职调查，并在需要时咨询具备相关资质的专业顾问。

参考文献 References

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

Castillo, J. C., & Mathur, S. (2023). Matching and network effects in ride-hailing. AEA Papers and Proceedings, 113, 244–247. https://doi.org/10.1257/pandp.20231069

Kilian, L. (2009). Not all oil price shocks are alike: Disentangling demand and supply shocks in the crude oil market. American Economic Review, 99(3), 1053–1069. https://doi.org/10.1257/aer.99.3.1053

Kirkham, C., Shirouzu, N., & Roy, A. (2026, September 3). Tesla starts Cybercab rides in Austin, drawing U.S. safety agency interest. Reuters.

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

National Highway Traffic Safety Administration. (2026). NHTSA opens investigation into Tesla Cybercab self-certification following Austin deployment. U.S. Department of Transportation.

Robinhood Markets, Inc. (2026). Robinhood reports second quarter 2026 results.

Saha, S. (2026, August 31). Oil to hold above $80 a barrel as Middle East supply risks persist. Reuters.

Taeihagh, A., & Lim, H. S. M. (2019). Governing autonomous vehicles: Emerging responses for safety, liability, privacy, cybersecurity, and industry risks. Transport Reviews, 39(1), 103–128. https://doi.org/10.1080/01441647.2018.1494640

Tesla, Inc. (2026). Full Self-Driving (Supervised). Tesla.

Uber Technologies, Inc. (2026). Uber announces results for second quarter 2026. Uber Investor Relations.

Waller, C. J. (2026, September 3). The economic outlook and some comments on my policy communication. Board of Governors of the Federal Reserve System.

AI、无人驾驶、资产代币化全面加速：特斯拉、博通、Meta、Robinhood正在重写下一轮财富逻辑

Tesla 特斯拉、Broadcom 博通、Meta 与 Robinhood 说明，市场越来越重视未来选择权 Optionality，而不仅是当前盈利。自动驾驶、人工智能与资产代币化潜力巨大，但监管、资本强度与执行能力仍是关键。真正的投资纪律，是区分技术可能性与已验证的商业经济性，并判断明天的盈利能否证明今天的估值。

如果您正在考虑通过新加坡房地产进行投资、移居、子女教育规划、家庭财富配置或长期资产配置，我始终认为，一位优秀的房地产顾问，不应只懂楼盘、价格与成交，更应该理解房地产背后的宏观环境与跨资产逻辑。

房地产从来不是一个孤立存在的资产类别。利率、全球流动性、国际地缘政治、股票市场、汇率、法规政策与跨境资本流动，都会影响房地产需求、融资成本与资产配置决策。

作为一名新加坡房地产销售人员，我每天都会投入大量时间研究宏观经济、国际事务、股票市场、数字资产、资产配置、新加坡房地产法律及市场动态，并持续撰写基于研究与事实分析的文章。我重视尽职调查，也相信，拥有跨资产视角，能够帮助客户作出更理性、更有纪律、更符合长期目标的决策。

无论您是 Family Office 家族办公室客户、13U/13O 架构参与者、GIP 全球投资者计划申请人、VCC 可变资本公司相关投资者、UHNW 超高净值人士、机构投资者、国际家庭、中国客户、东南亚投资者、陪读家长，或有留学与移居规划的家庭，我都希望以长期适配性、风险意识与资产配置视角，为您提供专业的新加坡房地产服务。

在适合自身财务状况与风险承受能力的前提下，新加坡房地产可作为多元化投资组合中的重要一环。相较于部分高波动资产，其日常价格波动通常较低，同时具备潜在资本增值与租金收入来源，但任何投资结果都不应被视为保证。

如果您重视专业分析、严谨尽调、长期视角与负责任的服务，我将很荣幸协助您规划属于自己的新加坡房地产之路。