NVIDIA Stock just delivered a record-breaking Q4 with $68.1 billion in revenue, 73% year-over-year growth, and earnings per share of $1.62 that crushed estimates. JPMorgan, among others, wasted no time raising its price target from $250 to $265.

Yet on February 26, the stock fell nearly 7% from its session high of $197 to under $185. The results are undeniable. But the price action, the money flow, and the institutional behavior tell a very different story. At least, for now.

The Numbers Look Bulletproof, Until You Look Closer

NVIDIA’s Q4 numbers speak for themselves. Revenue hit $68.1 billion, up 73% year-over-year. The data center segment alone pulled in $62.3 billion, making up 91% of total revenue. EPS (Earnings Per Share) of $1.62 beat the $1.53 consensus by nearly 6%.

And the Q1 FY2027 guidance of $78 billion blew past Wall Street’s $72.8 billion estimate — a figure that notably excludes any revenue from China.

JPMorgan analyst Harlan Sur responded by lifting the Nvidia price target from $250 to $265.

But here is what most analysts are not highlighting. NVIDIA’s quarter-over-quarter growth rate is quietly decelerating. Q3 grew 22% over Q2. Q4 grew 19.5% over Q3.

The Q1 guidance implies roughly 14.5% sequential growth. Revenue keeps hitting records, but the pace of acceleration is fading. For a stock priced on growth momentum, this distinction matters. Something big money might be watching.

BREAKING: Nvidia stock, $NVDA, extends gains to +4% after posting a record quarter, adding +$190 BILLION in market cap.

Earnings results include:

1. Record quarterly revenue of $68.1 billion
2. Q1 2026 revenue guidance of $79.6 billion
3. Data Center revenue up ~+1,200% since… pic.twitter.com/X0sCBnUYyQ

Advertisement

— The Kobeissi Letter (@KobeissiLetter) February 25, 2026

There is also the question of who is actually driving this revenue. Deepwater Asset Management’s Gene Munster estimates that roughly 70% of Nvidia’s revenue comes from just 8 companies.

I estimate roughly 70% of $NVDA revenue currently comes from just 8 companies. This concentration underpins investor concerns regarding the long-term sustainability of this growth.

— Gene Munster (@munster_gene) February 25, 2026

CFO Colette Kress confirmed that the top 5 hyperscalers (cloud computing providers) account for slightly over 50% of data center revenue. That level of customer concentration means that even a modest 10-15% reduction in AI capex from a few major buyers could translate into billions in lost quarterly revenue.

Advertisement

It is also worth noting that JPMorgan’s asset management division is itself a significant institutional holder of Nvidia.

This is standard on Wall Street, but it is a context that retail investors should be aware of when evaluating the bullishness behind a price target upgrade.

What Retail NVDA Investors See vs What Institutions Are Doing

On-Balance Volume (OBV), an indicator that tracks cumulative buying and selling pressure by adding volume on up days and subtracting it on down days, tells a positive story on the surface.

OBV has maintained higher highs throughout Nvidia’s 3-month consolidation, suggesting retail-driven buying pressure remains consistently positive. However, it still needs to break past its ascending trendline resistance to confirm genuine broad-based strength.

The most recent 13F filings (quarterly reports large investors must file with the SEC revealing their positions) for Q4 2025 show a dramatic shift in institutional sentiment.

Net institutional money flow surged to approximately $149 billion in purchases against $36 billion in sales — a net inflow of roughly $113 billion. That is a massive improvement from Q3, where institutions bought $38 billion and sold $34 billion, leaving a net inflow of just $4 billion.

Advertisement

Yet despite this wall of institutional money entering NVDA in Q4, the stock barely moved — trading sideways for most of the period. That suggests institutions were accumulating, but supply from insiders and earlier holders absorbed the demand. NVIDIA director Mark Stevens sold approximately $40 million in shares in December.

Bank of America, while slightly increasing its equity stake, closed out both its call and put options positions entirely — neutralizing its directional bets.

Institutions are clearly positioned. But the hedging and the flat price despite massive inflows suggest they are bracing for something. The next section explores what that might be.

The Risk Hiding in the Charts

The Chaikin Money Flow (CMF), an indicator that measures whether money is flowing into or out of a stock based on where the price closes within its daily range weighted by volume, reveals what the earnings headline does not.

Since February 5, as the right shoulder of Nvidia’s inverse head and shoulders pattern formed, CMF climbed steadily alongside the price. It rose all the way into the February 25 earnings breakout when Nvidia briefly touched $197.

Advertisement

Then on February 26, as the stock reversed sharply to $185, CMF plunged.

That sudden collapse suggests the money flowing in during the rally was speculative positioning — not committed institutional capital — and it evaporated the moment the breakout failed. And based on what we discussed earlier, revenue deceleration could be a reason.

The monthly VWAP (Volume Weighted Average Price, which approximates where institutions have built their positions) reinforces this. NVIDIA had been trading above its monthly VWAP since breaking out on February 17.

The last time Nvidia broke below the monthly VWAP was on January 30, which led to a correction of approximately 8.5% by early February.

Advertisement

As of February 26, the stock has once again fallen below this line. This means recent institutional buyers are now underwater, which historically triggers further selling as stop losses unwind.

The technical breakdown has context. Michael Burry flagged today that Nvidia’s supply commitments have ballooned to levels that mirror Cisco before the dot-com bust — a company that wrote down billions when demand didn’t meet expectations.

CFO Kress acknowledged Nvidia has locked in inventory “further out in time than usual.” Bulls like BofA’s Vivek Arya argue this secures Nvidia’s dominance. But CMF collapsing and VWAP breaking on the same day suggests the market isn’t waiting to find out who’s right.

The NVIDIA Stock Price Levels That Decide What Happens Next

The charts, the money flow, and the institutional positioning all point to the same conclusion — $195 is where conviction gets tested, a level highlighted later on the chart. But first, the risk.

On the daily chart, a hidden bearish divergence has formed between November 10 and February 25. During this period, the NVIDIA stock price made a lower high while the Relative Strength Index (RSI), a momentum indicator, made a higher high

Advertisement

It is a signal that upward momentum is quietly fading even as the stock appears to hold its range.

Since that November divergence started developing, Nvidia has been locked between $169 and $199. It couldn’t break out of this consolidation despite multiple attempts — including the inverse head-and-shoulders breakout on February 25, which failed within 24 hours.

The Fibonacci extension levels from the pattern now frame what comes next. On the downside, $183 at the 0.5 level is the immediate support. Below that, $180 at the 0.382 level becomes critical — a break there exposes $170, the right shoulder low, and $169, the head. Those levels would invalidate the pattern entirely.

On the upside, the neckline at $195 remains the key resistance and the conviction tester. A clean daily close above it, which the NVIDIA stock failed to do yesterday, is needed to reactivate the pattern.

That could push it towards the projected target at $226, the full head-to-neckline measurement.

The next extension at $235 brings it closer to JPMorgan’s $265 target. The path exists on paper.

Advertisement

But as the money flow, the hidden bearish divergence, and today’s 7% rejection all confirm, this is a market that’s not buying it yet.

Licensing Expansion to Cover Dealers and Custodians

The government will introduce a bill this year to license digital asset dealers and custodians. The proposal will expand regulation beyond trading platforms and bring more service providers under formal supervision. As a result, authorities aim to close regulatory gaps and strengthen operational standards.

Officials structured the reforms under Hong Kong’s second digital asset policy statement. The framework seeks to balance innovation with clear compliance obligations across the market. At the same time, regulators intend to reinforce market integrity and financial stability.

The Securities and Futures Commission will oversee key parts of the expanded regime. It plans to broaden approved products and services for professional participants. In addition, the regulator will launch an accelerator program to support compliant financial technology development.

Stablecoin Licensing and Market Liquidity Measures

Authorities have confirmed that Hong Kong has implemented a licensing system for fiat-referenced stablecoin issuers. The first batch of licenses will be granted next month under the new framework. Consequently, the city will move from regulatory planning to live market authorization.

Regulators will work with approved issuers to develop controlled and compliant use cases. Officials aim to integrate stablecoins into payment and settlement activities within clear risk parameters. Meanwhile, authorities will monitor issuance structures and reserve management standards.

The Securities and Futures Commission will also take steps to deepen digital asset market liquidity. It will expand the scope of eligible instruments and services available to professional market participants. Therefore, policymakers expect stronger capital flows and improved price discovery across platforms.

Hong Kong’s broader strategy reflects rising global competition among financial centers. Several jurisdictions have advanced stablecoin and tokenization rules in recent years. In response, Hong Kong has accelerated its regulatory timetable to maintain regional leadership.

Tokenized Bonds and OECD Reporting Framework

Tokenization forms another core pillar of the government’s digital asset strategy. Authorities will issue guidance allowing debenture holder registers to operate on distributed ledger systems. This clarification will support legal certainty for tokenized bond structures.

Officials will also explore electronic signatures for bond issuance documents. In parallel, authorities will examine the digitalization of bearer bonds within existing legal boundaries. These measures aim to modernize debt markets while preserving regulatory oversight.

Hong Kong has already experimented with tokenized green bond issuance in recent years. Those pilot projects demonstrated operational feasibility and settlement efficiency. Building on that experience, policymakers now seek broader institutional adoption.

At the same time, the government will amend the Inland Revenue Ordinance. The changes will implement the OECD Crypto-Asset Reporting Framework and the updated Common Reporting Standard. A bill is expected in the first half of this year.

The new reporting rules will strengthen cross-border tax transparency for digital asset transactions. Authorities intend to align Hong Kong with global standards on financial disclosure. Therefore, the reforms will address tax compliance while supporting market credibility.

Together, the licensing expansion, stablecoin approvals, and tokenization guidance mark a coordinated policy push. The measures integrate regulation, innovation, and tax reporting into a unified framework. As implementation begins, Hong Kong positions itself as a structured and competitive global digital asset hub.

The AI-driven data-center expansion is increasingly financed through debt, and lenders are weighing risk and opportunity in the AI-infrastructure and crypto-mining nexus. TheEnergyMag’s latest newsletter tracks roughly $33 billion in long-term senior notes raised over the past 12 months, excluding convertible debt, underscoring how traditional lenders view capture risk and growth potential in this space. In parallel, debt markets show widening spreads: AI- and crypto-linked issuers typically pay 7%–9% coupons, versus 4%–5% for regulated utilities. The momentum comes as Nvidia reports robust AI demand, while Bitcoin miners map a path toward dozens of gigawatts of new power capacity to support AI workloads.

Key takeaways

• AI data-center issuers have raised about $33 billion in long-term senior notes over the past year, excluding convertible debt, illustrating the scale of capital chasing AI compute capacity tied to crypto operations.
• Debt pricing shows a notable spread: AI/crypto-linked papers are typically priced around 7%–9% coupon, compared with 4%–5% for traditional regulated utilities.
• Recent placements include CoreWeave at 9.25% in May 2025 and 9% in July 2025, Applied Digital at 9.2% in November 2025, TeraWulf at 7.75%, and Cipher Mining at 7.125% and 6.125% as part of diversified AI-infrastructure financing.
• Nvidia’s fourth-quarter results underline sustained AI demand as a macro driver for data-center investments, with net income at about $43 billion and revenue near $68.1 billion, up sharply year over year.
• Bitcoin miners are targeting roughly 30 gigawatts of new power capacity to run AI workloads, a figure that would nearly triple current capacity and signal a coordinated push into AI-centric compute.

Tickers mentioned: $BTC

Sentiment: Neutral

Market context: The move to finance AI infrastructure via high-yield debt sits at the intersection of AI demand, crypto mining expansion, and a debt market that increasingly values long-dated, growth-oriented assets with offtake risk. As lenders price risk, capital flows reveal how investors are balancing the prospect of AI-driven compute with the volatility and energy-intense nature of crypto operations.

Why it matters

The current financing environment highlights a broader redefinition of what counts as infrastructure in the digital era. Projects that blend AI compute with crypto mining—whether repurposed data centers or greenfield AI data-hub builds—are increasingly treated as growth credits rather than traditional utility-style assets. This shift matters for developers and investors because it widens the pool of potential capital, but at a higher financing cost reflective of perceived tail risks, project complexity, and energy demand. The elevated coupons imply lenders are pricing in uncertainties around offtake arrangements, energy supply contracts, and regulatory risk, even as long-term demand for AI workloads remains a tailwind for data-center-heavy businesses.

The Nvidia earnings backdrop reinforces how AI compute can catalyze investment waves across adjacent sectors. Nvidia’s fourth-quarter performance—net income of about $43 billion and revenue of $68.1 billion, with year-over-year profit growth approaching the mid-to-high double digits—signals robust demand for AI accelerators and the compute capacity that data centers must deliver. While Nvidia is not a crypto-specific company, its results illuminate the demand side of AI infrastructure that, in turn, informs how lenders price risk for related projects. In parallel, Bitcoin miners’ plans to pursue roughly 30 gigawatts of new power capacity for AI workloads suggest a deliberate alignment between hash-rate economics and AI compute needs, potentially shaping energy markets and grid usage for years to come.

The financing narrative also underscores why some observers view the AI-infrastructure supercycle as broader than crypto alone. The sector’s access to capital hinges on how easily developers can secure long-duration debt with credible offtake, and how regulators and utilities respond to aggregate energy demand. The mix of blue-chip AI demand signals and crypto-driven compute pipelines paints a picture of a market that is increasingly comfortable funding ambitious buildouts—yet only under terms that reflect the complexity and risk of these multi-use facilities.

For readers tracking the intersection of AI, crypto, and infrastructure finance, the core takeaway is clarity: lenders are increasingly differentiating between steady, regulated load and growth-oriented, asset-light models that rely on AI-driven demand. That distinction translates into a bifurcated debt market where some projects on the frontier of AI infrastructure can access capital at high yields, while others with less certain offtake or regulatory clarity may see more muted appetite. The practical implication is a potential deceleration in some buildouts if the cadence of funding slows or if risk pricing tightens further, even as marquee projects with visible AI demand and confirmed long-term offtake can attract funding dollars more readily. The convergence of AI compute, crypto mining, and energy capacity decisions therefore remains a critical lens for investors navigating 2026 funding cycles.

Links and references from the reporting track the contours of this evolution. For instance, recent bonds tied to AI infrastructure were highlighted by TheEnergyMag’s analysis, which cites deals ranging into the 7%–9% coupon band. The same narrative is echoed in a presentation from Janus Henderson Investors, drawing on research from BofA Global Research, that underscores selective issuance in the high-yield space for 2026. At the project level, public disclosures and industry reporting have highlighted strategic moves by miners and AI infrastructure players, including stakes and capacity expansions in U.S. sites and AI-driven data-center deployments, which you can corroborate through industry updates linked below.

Related coverage includes a Canaan-led expansion in Texas mining sites and a Google-backed stake in Cipher Mining as part of broader AI-deal strategies that tie mining assets to compute demand. These developments illustrate how the collateral base for crypto-related data centers is expanding beyond traditional power contracts to include AI workloads and software-defined infrastructure. The broader takeaway is that the convergence of AI and crypto compute is reshaping both the risk-return profile and the capital allocation frameworks for data-center projects across the sector.

For readers seeking the underlying documents and official statements shaping these conclusions, the linked materials offer direct insight into issuer terms, credit ratings, and the strategic narratives driving these financing choices. The discussion remains dynamic: as AI adoption accelerates, lenders will recalibrate risk premia, and developers will adapt by locking in offtake commitments, hedging energy costs, and exploring hybrid models that blend traditional infrastructure with growth-oriented, AI-enabled compute.

What to watch next

• Upcoming bond issuances by AI-infrastructure developers and crypto-mining operators, including pricing, term sheets, and offtake arrangements.
• Regulatory developments affecting data-center expansions, energy usage, and crypto mining operations that could influence debt pricing and project viability.
• Updates on AI workload adoption by mining-centric or multi-use data centers, with potential implications for energy demand and grid resilience.
• Further commentary from chipmakers and AI platforms on demand trajectories and capital expenditure plans that could influence future risk pricing.

Sources & verification

• TheEnergyMag newsletter tracking about $33 billion in long-term senior notes tied to AI data-center and related projects: https://www.minerweekly.com/p/33-billion-bonds-ai-arms-race?
• Janus Henderson Investors article on high-yield bonds outlook citing BofA Global Research: https://www.janushenderson.com/en-ch/investor/article/high-yield-bonds-outlook-increasing-selectivity-in-2026/
• Canaan’s stake expansion in Texas mining sites: https://cointelegraph.com/news/canaan-buys-49-stake-texas-bitcoin-mining-sites-40m
• Google’s stake in Cipher Mining as part of an AI deal: https://cointelegraph.com/news/google-acquires-5-4-stake-in-bitcoin-mining-company-cipher-mining-in-ai-deal

AI infrastructure financing reshapes risk in crypto data centers

The 2026 US midterm elections are increasingly viewed as a potential catalyst tied to liquidity cycles and broader crypto market recovery.

The US midterm elections scheduled for Q4 2026 are increasingly being discussed as a potential macro catalyst for financial markets.

This includes crypto, amid expectations of changing liquidity conditions.

Asset Prices, Not Politics

According to a macro thesis by market participant ‘Egrag Crypto,’ early signals from betting markets point to relative Republican weakness, which could raise incentives for market-friendly economic conditions heading into the election window.

The framework outlines a three-phase timeline, which begins with a broader market correction in early 2026, during which criticism is expected to intensify toward Federal Reserve Chair Jerome Powell.

This is followed by mid-2026 pressure for a change in monetary stance, which could potentially result in liquidity easing as policymakers respond to economic and political constraints. Under this scenario, markets could enter a recovery phase in the second half of 2026, aligning with the election period.

The thesis argues that rising asset prices tend to improve public sentiment rapidly, supported by factors such as dividend income, potential tax relief for small businesses, and broader “feel-good” economic conditions. They further suggest that the Federal Reserve often becomes a focal point for blame during downturns, which, in turn, allows political narratives to shift as liquidity conditions improve.

As such, the view validates the idea that market structure and liquidity trends may play a leading role in shaping political outcomes, rather than political developments acting as the primary driver of markets.

You may also like:

“Structure first. Politics later. Markets always lead.”

2024 Flashback

In 2024, the cryptocurrency market saw significant price rallies following Donald Trump’s election victory. Bitcoin rose to record highs on investor optimism about a potentially more crypto-friendly regulatory environment and pro-crypto lawmakers in Congress.

However, by early 2026, much of the post-election upside had been eroded. Bitcoin, for one, retreated toward $60,000, and broader crypto sentiment cooled amid macro pressures and fading Trump-driven euphoria.

Bitcoin (BTC) has rebounded 7.45% over the past two days after dropping to $62,400 on Tuesday, below a key onchain price support. Despite the bounce, holders who bought between six months and two years ago remain at an average cost of $74,500, a level that now stands as a potential inflection level.

As BTC moves higher, the concentration of supply around $74,500 stands as a key test for the current trend; a decisive reclaim of that level may signal demand and a shift in short-term market structure.

Why $74,500 matters to Bitcoin bulls

Bitcoin’s realized price tracks the average onchain acquisition cost for a given UTXO age band. For coins aged 18 to 24 months, that level stands near $64,200.

Crypto analyst Anıl noted that Bitcoin tested this threshold and reclaimed it by the daily close on Tuesday, keeping the zone intact for now.

Cost basis levels act as psychological pivots and when the price trades below them, investors face unrealized losses and the risk of distribution increases. A sustained position above the band tends to reduce investor stress and encourages BTC re-accumulation. 

Expanding the lens to BTC UTXOs aged six months to two years captures investors from the prior cycle’s consolidation and breakout phases. The realized price for these cohorts is near $74,500, which is well above the current price.

The cohort’s MVRV ratio, which compares market value to realized value, now sits at 0.88. A reading below 1 signals that the group is, on average, holding at a loss.

As Bitcoin fell below $74,500, investors who bought between six months and two years ago moved into unrealized losses, turning that level into an important profitability threshold.

A sustained move back above $74,500 places much of this group back in aggregate profit, which may ease sell-side pressure from holders looking to exit near their breakeven price.

BTC long-term supply climbs to 3-month high

Onchain supply data from CryptoQuant shows that the long-term holder balance is back near 14 million BTC (13.96 million) after falling to a multi-year low on November 21, 2025. The recovery in the aged supply points to continued coin dormancy despite recent volatility.

If investors who bought between six months and 2 years ago choose to hold and absorb selling near their average entry price, the supply sitting between $74,500 and $100,000 may thin out more quickly.

A sustained rally above $74,500 may push a large portion of these coins back into profit, potentially shifting focus toward liquidity near $100,000. 

Related: GD Culture Group board authorizes Bitcoin treasury sales

BTC realized cap and capital flows remain flat

An uptick in BTC’s realized cap, which measures the aggregate value of coins based on their last onchain movement price, may also signal a trend shift.

The metric is holding near cycle highs, though its rate of expansion has slowed. The realized cap net position change has compressed toward neutral or 0%, signaling that capital inflows are negligible.

While the realized cap remains near all-time highs, it is trending lower, indicating a slowing pace of new capital entering at the higher cost basis levels.

Historically, late bear market phases tend to show flat, or contracting realized cap, while early recoveries begin with stabilization before acceleration. A renewed expansion in the net position change back toward the 2–4% range may provide clearer confirmation that fresh capital is re-entering and that accumulation is on the rise.

Related: Bitcoin’s upcoming $10.5B options expiry may end bear market: Here’s how