Has the AI bubble burst? Will the 2000 tech sector crash repeat itself?

EX. IO Research date: July 16, 2026

The AI bubble narrative reached unprecedented intensity in July 2026—not just in financial media, but simultaneously issuing warnings from sovereign funds, central banks, legendary investors, and on-chain data. But this is not a prediction article about whether the AI bubble will burst. This is a hypothetical scenario simulation: If a bubble bursts, how will capital move? As $2-3T of capital spills over from the AI narrative searches for the next destination, the RWA/tokenization track—coinciding with DTCC’s launch of production-grade tokenization transactions—may be one of the most structurally advantageous alternative narratives. For Hong Kong’s virtual asset ecosystem, this represents a potential inflection window for narrative leadership shifting from AI to crypto/RWA—but this is a scenario that requires ongoing validation, not deterministic judgment.

1. Signal Density Has Reached Critical Mass: Why July 2026 Is Different from Previous Years

“AI bubbles” are not a new topic. Since ChatGPT ignited the market in 2023, there has been a round of discussions about the “bubble bursting” every 3-4 months. But the difference in July 2026 is that the warning signals come from completely different sources and are independently verified by each other.

On-chain signal: Smart Money is voting with its feet

On July 14, on-chain data on Hyperliquid showed that Smart Money had established a $13 million short position on the NVDA perpetual contract and is still continuously adding positions(Net position change +$29.2K, indicating increased position rather than holding onto the coin from a wait-and-see perspective)[1]. It should be noted that $13M is very small compared to NVDA’s roughly $3T market cap—but the signal value of on-chain perpetual contract data lies in its leadership and transparency (real-time, immutable, 24/7) rather than its volume. Short selling of similar scale in traditional markets requires regulatory disclosure to be observed, while on-chain data provides near-instant sentiment readings.

What deserves more attention is the position structure:

• – NVDA perpetual contract open interest reached $153.8 million—an extremely high level of activity for on-chain derivatives of a single stock
• – Long-short ratio 8:13, bears have the advantage
• – This data comes in a key context: Hyperliquid is a crypto-native perpetual contract platform, and its NVDA short positions represent a collective judgment by crypto traders on AI stocks—not traditional fund managers reading research reports on Bloomberg Terminal, but smart money making decisions based on on-chain data

Traditional Financial Signals: From “Discussion” to “Pricing”

Signal source	Action	Date	Signal strength
Michael Burry (the prototype for ‘The Big Short’)	Short Nvidia and Palantir	2026	⚡⚡⚡
Norway Sovereign Wealth Fund ($2.2T)	Publicly modeling AI crash scenarios	2026 Q2	⚡⚡⚡
Peter Thiel + Softbank	Betting on a pullback in the AI sector	2026	⚡⚡
Paul Tudor Jones	Warning of a possible 35% market plunge	2026 Q2	⚡⚡
Bank of England (BoE)	The official report warns of AI bubble risks	2025 Q4	⚡⚡
Sam Altman（OpenAI CEO）	Publicly acknowledges that “there is a bubble in the AI field”	August 2025	⚡⚡
S&P 500 internal rotation	$3.2T shifted from chip stocks to Mag7	July 2026	⚡⚡
Nasdaq 100 futures	A 30-day consecutive downward trend	July 2026	⚡

On July 16, Yahoo Finance headlined: “A $3.2 trillion rotation from chips to the Magnificent 7 has left the S&P 500 going nowhere” [3]. This $3.2T rotation is not about exiting the market—it’s about shifting from AI chips to more defensive tech giants within the market. This is typical cyclical tail behavior.

Opposing viewpoint (must be presented honestly)

On July 16 (today), TSMC raised its capex and full-year revenue guidance, citing continued strong demand for AI chips [4]. The Japanese government is procuring Nvidia Rubin chips to build sovereign AI infrastructure. Some analysts point out that current AI investments are backed by real corporate revenues, which differs from the “click-through value” of the dot-com era.

Our judgment: The opposing view points to the structural reasons why the AI bubble will not burst tomorrow. But bubble bursts are often not because “AI has no value,” but because “AI’s valuation cannot match its revenue growth”—which is exactly what the market is currently pricing in. TSMC’s upward guidance is a lagging signal on the supply side; Short selling on Hyperliquid is a leading signal on the demand side.

2. If the AI bubble really bursts, where will the money go? — Three paths of capital migration

This is not a theoretical issue. Norway’s sovereign wealth fund is already modeling this issue. When $2-3T exits the AI sector, there are only three ways to spend that money:

Path 1: Defensive cash/Treasury bonds (most certain in the short term)

This is a historical template. After a bubble bursts, capital’s first conditioned reflex is always “safe haven”—government bonds, money market funds, gold. This has a negative short-term impact on Bitcoin and crypto: BTC remains highly correlated with Nasdaq, and panic selling in the AI sector will be transmitted to the crypto market through risk appetite channels.

Signal support: On July 16, BTC pulled back from a monthly high of $65,500 to $64,000, attributed by CoinDesk to profit-taking + the combined effect of Iran’s attack on a US military base [2]. BTC’s 30-day implied volatility is at 38%—historical data shows that volatility readings below 40% often signal upcoming market turmoil [2].

Path Two: Searching for the Next “True Income” Narrative (Mid-Stage—Core Argument of This Article)

After the dot-com bubble burst in 2000, the survivors were not the best storytellers but companies with real revenue and cash flow (Amazon’s e-commerce, Google’s advertising). Similarly, when AI narratives tire out, capital will look for alternative narratives supported by real income.

RWA/tokenization is currently the only track that simultaneously meets the following conditions:

• – Has real underlying assets (government bonds, stocks, real estate)
• – Verifiable revenue streams (revenue tiering, transaction fees)
• – Mainstream traditional financial infrastructure is currently launching (DTCC launched production-level tokenized trading this week)
• – Clear regulatory frameworks (Hong Kong SFC, Dubai VARA, Singapore MAS)
• – Natural bridges with traditional finance (no need to “educate the market” on tokenization)

Path Three: Crypto-native Narratives (Survivors of DeFi and DeAI)

If funds withdraw from AI stocks but remain in the risk asset category, the crypto-native sector will see some inflows—but this is selective benefit, not a broad rally.

Beneficiaries:

• – Bitcoin (Digital Gold Narrative)
• – Ethereum/DeFi blue chips (with real protocol revenue)
• – RWA tokens (ONDO, Morpho, Maple, etc.)

Victims:

• – AI agent tokens (directly related to AI narratives, the first to be hit when the bubble bursts)
• – Purely narrative-driven meme coins
• – No product-market-adapted L1/L2

Data Verification: On July 16, DeFi lending protocol MORPHO bucked the trend and rose 3.5%, testing the $2.20 resistance level—while most altcoins were declining [2]. MORPHO, as a decentralized lending protocol (not an AI token), saw its rise coinciding with Galaxy’s launch of an institutional stablecoin yield product based on Morpho [5]. This trend aligns with the assumption that “protocols backed by real revenue outperform narrative-driven tokens in risk adjustment,” but single-day data is insufficient to conclude — a pattern that still needs ongoing verification.

3. Why is RWA/Tokenization the “Right Next Narrative”?

Coincidence of timing is not a coincidence: DTCC production-level tokenized trading + AI bubble narrative emerges simultaneously

On July 15, 2026—the same week when the AI bubble narrative was at its peak—Depository Trust & Clearing Corporation (DTCC) processed its first production-grade tokenized securities transaction [6].

How big is this a big deal? DTCC is the backbone of the U.S. securities clearing system, custody over $114 trillion in securities. Every day, it records and processes the ownership and settlement of stocks, bonds, and other securities. The following organizations participated in this pilot:

• – Banks: JPMorgan Chase, Goldman Sachs
• – Asset Management: BlackRock, Vanguard
• – Specific transactions: JPMorgan tokenizes the Invesco QQQ Trust ETF → as CME central counterparty margin→ the SPDR S&P 500 ETF is tokenized→ treasury tokenized trading → repurchases and collateral transfers
• – Underlying chain: Hyperledger Besu + Canton Network
• – Timeline: DTCC plans to officially launch tokenization services in October 2026

DTCC President Frank La Salla’s statement is noteworthy: this is not about “creating new digital assets,” but about “converting existing securities into digital twins on the blockchain, retaining the same legal ownership, dividend rights, and governance rights” [6]. This is completely different from the crypto-native “synthetic stock” model—tokenization is done within the existing legal framework.

The second signal: Cantor + Securitize collaborated on blockchain IPO

On July 15, Cantor Fitzgerald and Securitize announced a partnership to launch a blockchain IPO channel, creating a pathway for listed companies to raise funds directly on-chain and issue tokenized securities [7]. Cantor has deep roots in global capital markets; Securitize is the technical standard setter for tokenized securities (infrastructure provider for BlackRock BUIDL).

Third signal: Alpaca raised $135 million—real money invested in tokenized equity infrastructure

On July 16, crypto brokerage Alpaca completed a $135 million funding round dedicated to building tokenized equity infrastructure [9]. Alpaca is a licensed broker in the United States, with API clients covering thousands of fintech and trading platforms—this funding means the tokenized infrastructure of the brokerage layer is moving from ‘do or not’ to the execution stage of ‘financing-build-launch.’

Previously, in January, Alpaca had already partially allocated tokenization from its $150 million funding round, and this $135 million special financing round further concretizes its roadmap. The Alpaca CEO positioned this financing as “a next-generation trading infrastructure competing with Interactive Brokers.”

Additional signal: RWA perpetual contract trading volume hits a record high

In June, RWA perpetual contract trading volume soared to a record $311 billion; CEX spot trading volume rose for the first time in five months, growing 15.3% to $1.11 trillion [2]. This set of data shows that RWA trading demand is not self-emerging from crypto-native territory, but is supported by real market liquidity.

Why are these signals grouped together?

Signal	Levels	Meaning
DTCC production-grade tokenization	Settlement layer	The “licensing stamp” of traditional financial infrastructure—not “can it be done,” but “when it will be fully rolled out.”
Cantor + Securitize IPO channel	Distribution Layer	The gateway to the capital market is opening—not a crypto exchange listing, but a real IPO
Alpaca raised $135M	Brokerage layer	Real money voting—brokers betting on tokenized stocks is the next wave of trading infrastructure
RWA perpetual contract $311B	Trading layer	Demand already exists—not waiting for liquidity—but liquidity is already pouring in

From settlement → issuance→ brokerage→ trading, all four layers simultaneously released intensive signals on July 15-16. This is no coincidence—it represents a critical moment for the RWA/tokenization sector to “launch infrastructure synchronously.” When $2-3T flows out of the AI sector, is there a track that offers both real revenue support and a complete infrastructure stack to take over? The answer is RWA/tokenization—and just as July 2026, this track will move from “proof-of-concept” to “full-stack production-ready.”

4. Crypto AI Exposure: Which Assets Are Most Vulnerable During Bubble Bursts?

⚠️ The following is a structural exposure analysis and does not constitute any asset rating or trading advice.

Asset classes	AI bubble exposure	Source of resilience
**AI Agent Tokens** (FET, AGIX, WLD, etc.)	Extremely high	None—completely relying on AI narratives
**GPU Compute Tokens** (RENDER, IO, etc.)	High	Some have real computing power demands, but valuations are highly correlated with AI narratives
**DeFAI / DeAI Middleware**	Middle	If survivors after the bubble burst can prove the product’s value, it could become the next cycle leader
**RWA Tokens** (ONDO, MORPHO, MPL, etc.)	Low	The underlying assets are unrelated to AI; the returns come from government bonds and borrowing
**BTC**	Low	Digital gold narratives are independent of AI
**ETH / DeFi Blue Chips**	Mid-to-low	In the short term, risk appetite may drag down; in the medium term, narrative rotation may benefit

Key takeaway: MORPHO’s counter-trend rise is no coincidence

On July 16, while most altcoins followed BTC downward, MORPHO bucked the trend and rose 3.5% [2]. MORPHO is a decentralized lending protocol whose token value is backed by actual protocol usage and revenue—not AI narratives. At the same time, Galaxy Digital announced the launch of a Morpho-based institutional stablecoin yield product targeting 2,400 institutional clients of Fireblocks [5].

The core lesson of this case: when the market panics, funds do not withdraw evenly—they internally reallocate from high-narrative, low-yield assets to those with real income. MORPHO is one of the beneficiaries of this “internal rotation.”

5. 🇭🇰 What does this mean for Hong Kong’s virtual asset ecosystem?

Macro Opportunities: The “Licensing Effect” of Regulatory Infrastructure

Hong Kong is not a global leader in the field of AI. But at the intersection of tokenization + compliant digital assets, Hong Kong boasts one of the most comprehensive regulatory infrastructures in the world

• – SFC’s Virtual Asset Trading Platform Licensing Regime (VATP)
• – HKMA’s stablecoin sandbox
• – Precedent of government green bond tokenization (HKD 80 billion in 2023)
• – Project Ensemble (wCBDC + tokenized deposits)
• – Digital Hong Kong Dollar (e-HKD) pilot

In the hypothetical scenario of market narrative rotation, Hong Kong’s tokenization regulatory infrastructure may become a structural advantage in attracting institutional capital—not by relaxing regulations to attract crypto-native venture capital, but by ensuring compliance to attract institutional allocation funds withdrawing from the AI sector and seeking a “regulated alternative narrative.”

Competitive landscape: South Korea is accelerating

On July 15, South Korea announced amendments to its 76-year-old law, classifying cryptocurrencies as “national assets,” and confirmed that next year it will pilot tokenized government bonds and explore tokenized state-owned real estate [8]. South Korea’s move in the Asian tokenization race puts direct competitive pressure on Hong Kong—especially in the RWA sector, where the regional leader position remains undecided.

Implications for Hong Kong’s licensed platforms

• – DTCC model can be referenced: DTCC’s tokenization framework (digital twin, retained legal rights) provides Hong Kong licensed platforms with a clear paradigm for RWA product design—tokenizing within existing legal frameworks rather than creating new asset classes
• – Narrative Window Period: DTCC plans to officially launch tokenization services in October—from now until October, licensed platforms have a window to build market awareness and product readiness
• – Regional Positioning: In the potential narrative rotation of AI → RWAs, Hong Kong-licensed platforms have the opportunity to position themselves as “compliant RWA infrastructure” rather than “crypto exchanges.”

Risk warning

• – AI bubbles do not show significant adjustments, or the order of adjustments differs from the projected scenario (first impact on crypto, then on AI)
• – The escalation of Middle East geopolitics (Iran has begun attacking U.S. military bases) could overwhelm all market narratives
• – Regulatory arbitrage may divert capital to Dubai or Singapore rather than Hong Kong—the outcome of regional competition depends on execution speed

6. Scenario simulation: Three-stage monitoring indicators

The following are hypothetical scenario simulations describing possible development paths under different market conditions and are not recommendations for action.

Stage One: Bubble Pressure Accumulation Period (Current Stage)

Key monitoring indicators (by priority):

1. NVDA drops >8% in a single day—a landmark event that could trigger narrative acceleration

2. Norway’s sovereign wealth fund publicly announces reduction of its AI holdings—a turning point in the “signal > action.”

3. Nasdaq 100 pulls back >20% from its high—confirmation of a technical bear market

4. BTC/ETH decoupling from AI stocks—a breakdown in positive correlation = a confirmation signal for the start of rotation

5. RWA TVL is accelerating growth—evidence of actual capital inflows

6. Divergence between AI tokens and RWA tokens—quantitative indicators of internal differentiation

Stage Two: Decoupling Confirmation Period

If ≥ of the above indicators are triggered simultaneously, the following scenarios may unfold:

• – RWA/DeFi blue-chip stocks may be the first to attract liquidity withdrawing from the AI narrative
• – Traditional financial institutions may accelerate RWA product deployment after DTCC tokenization goes live (in October).
• – The compliance advantages under the Hong Kong SFC framework may become apparent at this stage

Stage Three: Narrative Rotation Establishment Period

If the first two stages occur in sequence, the following patterns may form:

• – The narrative leadership transition of AI → RWA has become a market consensus
• – Tokenized products from DTCC participating institutions (JPMorgan, BlackRock, etc.) have entered scale
• – Regional regulatory race acceleration (Hong Kong/Dubai/Singapore/South Korea)

7. Risks and Uncertainties

• – Risk of false alerts: The AI bubble may last another 6-12 months. TSMC is still increasing capex, and the Japanese government is procuring Rubin chips—physical evidence on the demand side remains
• – Geopolitics Dominates: The escalation of the Iran-U.S. conflict (the July 16 Iran attack on a U.S. military base) could shift the narrative across all asset classes to safe havens
• – Risk of “simultaneous breakdown”: Crypto and AI may fall simultaneously, rather than a rotation of one rising and one falling—short-term correlation may be higher than expected
• – Regulatory enforcement risk: If DTCC’s October tokenization service launch is delayed or restricted by the SEC, the RWA narrative may be delayed

Conclusion: Looking back at it in 7 days, is this article still important?

We judge: important.

Not because we can accurately predict when the AI bubble will burst—no one can do it. Rather, it’s because the structural changes outlined in this article (DTCC tokenization launch, RWA $311B trading volume, AI bubble signal density) will define the capital allocation direction for the next 12-18 months no matter when the bubble bursts.

The core judgment can be summed up in one sentence: if the AI bubble undergoes significant adjustments, it does not necessarily mean the end of crypto—it could accelerate the shift from “narrative-driven” to “revenue-driven” under certain conditions. RWA/tokenization is currently one of the most structurally advantageous alternative narratives, but realizing this scenario depends on multiple variables forming simultaneously (how the bubble bursts, timing window, regulatory rhythm), and is a hypothesis that requires continuous validation rather than deterministic judgment.

Source

[1] “Smart Money Short Nvidia $13 Million on Hyperliquid — Is the AI Bubble About to Burst?” — HackerNoon / The Chain Audit, July 14, 2026. https://hackernoon.com/smart-money-short-nvidia-$13-million-on-hyperliquid-is-the-ai-bubble-about-to-burst

[2] “Bitcoin Pulls Back to $64,000 After Hitting Monthly High as Bears Take Control” — CoinDesk, July 16, 2026. https://www.coindesk.com/markets/2026/07/16/bitcoin-pulls-back-to-usd64-000-after-hitting-monthly-high-as-bears-take-control

[3] “A $3.2 trillion rotation from chips to the Magnificent 7 has left the S&P 500 going nowhere — Chart of the Day” — Yahoo Finance, July 16, 2026. https://finance.yahoo.com/markets/article/32-trillion-rotation-from-chips-to-the-magnificent-7-has-left-the-sp-500-going-nowhere-chart-of-the-day-100000821.html

[4] “TSMC raises capex and revenue forecast, highlighting growing AI chip demand” — Yahoo Finance, July 16, 2026. https://finance.yahoo.com/markets/article/tsmc-raises-capex-and-revenue-forecast-highlighting-growing-ai-chip-demand-113101950.html

[5] “Galaxy targets institutional stablecoin yield with new DeFi vaults” — CoinDesk, July 16, 2026. https://www.coindesk.com/tech/2026/07/16/galaxy-targets-institutional-stablecoin-yield-with-new-defi-vaults

[6] “DTCC moves tokenized securities into live trading, marking a milestone for Wall Street’s blockchain push” — CoinDesk, July 15, 2026. https://www.coindesk.com/business/2026/07/15/dtcc-moves-tokenized-securities-into-live-trading-marking-a-milestone-for-wall-street-s-blockchain-push

[7] “Cantor and Securitize collaborate on blockchain-based IPOs” — CoinDesk, July 15, 2026. https://www.coindesk.com/business/2026/07/15/cantor-and-securitize-collaborate-on-blockchain-based-ipos

[8] “South Korea to modify 76-year-old law to classify cryptocurrencies as national assets” — CoinDesk, July 15, 2026. https://www.coindesk.com/policy/2026/07/15/south-korea-to-modify-76-year-old-law-to-classify-cryptocurrencies-as-national-assets

[9] “Crypto brokerage firm Alpaca raises $135 million for tokenized stock infrastructure” — CoinDesk, July 16, 2026.

[10] “Global stock markets fall sharply over AI bubble fears” — The Guardian, various dates.

[10] “One AI bubble has already burst. The next one—a ‘rare’ kind—is still growing, economist warns” — Fortune, March 29, 2026.

[11] “Bank of England warns of risks from AI bubble” — BBC News, December 2025.

[12] “Paul Tudor Jones warns Trump-era markets could crash 35%” — Economic Times, 2026.

[13] “U.S. Senate unanimously opposes clemency for FTX founder Sam Bankman-Fried” — CoinDesk, July 16, 2026.

[14] “Two groups of crypto investors are selling bitcoin to cap its latest price recovery” — CoinDesk, July 16, 2026.

Disclaimer

This material is for general informational and research reference purposes only and does not constitute any investment, financial, legal, or tax advice, nor does it constitute any solicitation, offer, or recommendation. The materials may contain third-party information or viewpoints, and do not represent the official position of any organization or individual. Prices of virtual assets and related products may be highly volatile. Investors should make independent judgments based on their financial situation, investment objectives, and risk tolerance, and conduct their own research (DYOR) and consult independent professional advisors before making any investment decisions. The relevant parties shall not be liable for any losses arising from the use or reliance on this material, except where applicable law cannot exclude or limit them.