The stock market had a rough week. The Nasdaq entered correction territory on Thursday, down more than 10% from its October 2025 peak. Brent crude is sitting near $111 a barrel on the back of the Iran conflict. The Dow followed the Nasdaq into correction on Friday. And NVIDIA's stock slid another 3.7% in a single session, capping a 13% decline since March 4.

NVIDIA was already struggling before the geopolitical chaos gave investors cover to sell. The stock fell more than 5% on February 26, the same day it reported a 73% surge in quarterly revenue and beat estimates on every single line. Read that again. The company delivered what may be the strongest earnings print of any $4 trillion business in history, and the market sold it. That is not a macro story. That is a very specific signal about investor psychology around AI.

The real question is not: "Is the 13% drop because of the market?" The real question is: "What was already cracking underneath, and is this selloff revealing it or just accelerating it?" The answer is more uncomfortable than the Iran war headlines. That is what we are getting into today.

The Fool highlights Ares Capital, Energy Transfer, and Starwood Property Trust as high-yield income plays, with yields around 10.7%, 6.9%, and 11% respectively. The core pitch is that these are not just flashy yield traps, since all three have long records of stable or growing payouts and business models built to support cash distributions. Still, the real takeaway for me is that ultra-high yield only works if the underlying cash flow holds up, so investors need to focus on payout durability, not just the headline number.

This MarketWatch piece makes a pretty uncomfortable point: even if the S&P 500 itself has not fully cracked, a lot of the damage is already happening underneath the surface. More than half of the index’s 25 industry groups were already down at least 10% from recent highs, with software, autos and parts, commercial services, and real estate management among the hardest-hit areas, while energy was the only major S&P sector up in March as oil surged. In plain English, the market’s internals already look a lot weaker than the headline index suggests.

Eli Lilly has struck a major deal with Insilico Medicine worth up to $2.75 billion, including $115 million upfront, to license and develop AI-designed oral drug candidates for global markets. The agreement gives Lilly exclusive rights to develop, manufacture, and commercialize the preclinical candidates in selected disease areas, showing that large drugmakers are getting more serious about using AI to shorten the long and expensive drug-discovery process. The broader message is simple: AI in pharma is moving beyond hype and deeper into real commercial deals.

That's the AI paradox hiding in your CX stack. Tickets close. Customers leave. And most teams don't see it coming because they're measuring the wrong things.

Efficiency metrics look great on paper. Handle time down. Containment rate up. But customer loyalty? That's a different story — and it's one your current dashboards probably aren't telling you.

Gladly's 2026 Customer Expectations Report surveyed thousands of real consumers to find out exactly where AI-powered service breaks trust, and what separates the platforms that drive retention from the ones that quietly erode it.

If you're architecting the CX stack, this is the data you need to build it right. Not just fast. Not just cheap. Built to last.

See the data

A market repricing specific AI and software assumptions is a very different animal from a market in free fall but that distinction only holds until it does not. Right now, the damage is concentrated. Energy, materials, and industrials are actually up strongly year-to-date. The S&P 500 itself finished last week at its lowest level in over seven months, sitting 8.74% off its January 27 all-time high, below both its 50-day and 200-day moving averages. Meanwhile, 16 of the S&P 500's industry groups are at least 10% off their 52-week highs, and fou, real-estate management, software, autos and parts, and commercial services, have crossed into bear-market territory.

That asymmetry is the context. Now here is what is happening inside it.

NVIDIA reported a 73% jump in quarterly revenue on February 26. It beat estimates. It gave strong forward guidance. CEO Jensen Huang spoke bullishly about Blackwell chip demand. Then the stock fell more than 5% that day. It kept sliding. By last Thursday it closed at $171.24, down 4.16% in a single session, with volume surging 20 million shares above the daily average. That volume spike matters. Sellers are not panicking out. They are methodically reducing.

The easy narrative: Iran war, oil shock, tech selloff, macro noise. That is real. But the trickier, more honest story is that NVIDIA's stock had already started breaking before geopolitics gave institutional investors a clean excuse to exit the world's most crowded long trade.

The underlying business is not the problem. In fiscal year 2026, NVIDIA's revenue reached $215.94 billion, up 65.47% from the prior year, with earnings of $120.07 billion, up 64.75%. Gross margins sit at 71.31% and net margins at 55.60%. NVIDIA keeps 56 cents of profit for every dollar of revenue it earns. That is not a software company number. That is a toll-booth number.

According to 38 analysts, the average 12-month price target for NVDA is $265.97 - nearly 59% upside from current levels. Goldman Sachs has a $250 target. Jefferies sits at $275. Microsoft, Amazon, Google, and Meta are still spending tens of billions on its chips. The business case is intact. Not even close to broken.

Here is where it gets uncomfortable.

First, Microsoft - one of NVIDIA's largest customers - quietly cancelled data centre leases in February. CEO Satya Nadella has publicly suggested AI is not yet creating meaningful value. If one of the industry's lead buyers says this openly, other major clients may be feeling the same way behind the scenes, just more quietly.

Second, Google published research last week on a new algorithm designed to reduce AI memory usage. The market reacted immediately - chip stocks tumbled on March 26, with Micron falling 7% and NVIDIA dropping 3%. If AI workloads become more memory-efficient, the demand trajectory for high-end GPUs changes in ways no marketing keynote can offset. This is not a hypothetical. It is a published paper. The direction has been named.

Third, the China overhang. NVIDIA's guidance does not include potential revenue from China. Export controls remain a persistent drag on the addressable market, and any tightening hits the revenue model directly. This is a closed door that Wall Street has not fully priced.

The fear is not that NVIDIA's business collapses.

The fear is simpler and harder to model: the stock was priced for a trajectory that required everything - spending, demand, margins, geopolitics - to stay elevated, forever. The moment any of those variables wobble, multiple compression can be swift and severe.

Even a deceleration from 65% growth to 30% growth - still extraordinary by any industry standard, re-prices this stock significantly. That is the real tension. Not collapse. Deceleration. And deceleration is very hard to see coming until it is already in the numbers.

NVIDIA's 52-week high is $212.19, reached on October 29, 2025. The 52-week low is $86.62. The stock closed at $171.24 on March 26, and both short- and long-term moving averages are now generating sell signals - the long-term average sitting above the short-term, a classic bearish alignment. A sell signal was issued from a pivot top on February 25, 2026, and the stock has fallen 12.44% since that point.

Volume rising on falling prices is a signal worth taking seriously. This is not panic selling. This is distribution.

The key technical level to watch: the $167 range - the lower trend floor. A sustained close below that opens the door to further selling. Recovery above the $176–$183 resistance band, where accumulated volume sits, would be the first meaningful signal that stabilisation is underway. Until the stock reclaims those levels with conviction, every bounce is a patience test for long-term holders and a potential shorting opportunity for traders.

🏗️ Hyperscaler Capex Guidance: Read The Transcripts, Not The Summaries

The next major earnings cycle for Microsoft, Amazon, Google, and Meta will be the single most important data release for NVIDIA bulls and bears alike. These four companies represent roughly half of NVIDIA's data centre revenue.

But here is the nuance most coverage will miss: the number itself is less important than the language around it. A CEO who maintains $50 billion in capex while describing it as "rationalising our compute footprint" is sending a very different signal than one describing "accelerating infrastructure deployment." The specific phrase to listen for is "compute efficiency" or "inference optimisation." If two or more hyperscalers use that language on Q1 calls, treat it as a structural shift, not a one-quarter data point, and reprice NVIDIA's demand trajectory accordingly. If none of them use it, the bull case gets a meaningful reprieve.

Trump extended his deadline for Iran negotiations by 10 days to April 6. At current energy prices driven by the conflict, the cost of powering the AI buildout has risen dramatically from pre-conflict projections. Goldman Sachs estimated AI data centres could represent 8% of total US electricity consumption by 2030, and that math gets harder at $111 oil.

There are only two outcomes on April 6: diplomatic progress that takes pressure off energy prices, or escalation that keeps them elevated. The second scenario does not just hurt sentiment, it directly raises the operating cost of every hyperscaler running NVIDIA's chips at scale, creating a quiet incentive to optimise workloads rather than expand them. Watch April 6 as an inflection, not a footnote.

The specific trigger to monitor: any Commerce Department action on chips currently classified below the H20 threshold, the product NVIDIA engineered specifically for restricted markets after earlier export controls. If that line moves, NVIDIA loses its last meaningful China revenue stream. The guidance gap between what the company reports and what it could report in an open market becomes permanent, not cyclical. This is not priced in. It is not even widely discussed. That is exactly why it matters.

If all three of these signals move negative in the same quarter, hyperscalers hedging language, energy costs staying elevated, and a new China restriction, the deceleration scenario stops being a risk and becomes a base case. That combination would warrant a fundamental reassessment of the bull thesis, not just a technical price target adjustment.

Attention is the hardest thing to buy. And everyone else is bidding too.

When people are scrolling, skipping, swiping, and split-screening their way through the day, finding uninterrupted moments where your audience is truly paying attention is the priority.

That’s where Performance TV stands out.

Check out the data from 600+ marketers on the most effective channels to capture audience attention in 2026.

Download Report

The most dangerous investing mistake of the last three years has been confusing a great business with a great stock. NVIDIA may be the most profitable semiconductor business built in a generation and yet investors who bought at $212 in October are down nearly 20%, not because the company has failed them, but because their entry price has. This is the part of investing that no earnings beat, no analyst upgrade, and no bull case ever states plainly: the price you pay is not a minor detail. It is the entire game.

When a stock is priced for perfection, even near-perfection disappoints. Markets do not reward what companies do; they reward what companies do relative to what was already expected. The market was extremely top-heavy heading into 2026, with roughly half of the S&P 500's market cap tied to just 20 stocks and when those 20 stocks start repricing, the index feels it everywhere. The patience to wait for a great company to trade at a sensible price is not a passive skill. It is an aggressive edge. In a week when the Nasdaq entered correction and oil hit $111 a barrel, that edge has never felt more worth cultivating.

NVIDIA's stock now trades at a discount to the S&P 500 on its forward price-to-earnings ratio, which is either a screaming opportunity or a valuation trap, depending entirely on whether 30%-plus annual earnings growth holds through fiscal 2028. Buying NVIDIA here without understanding the hyperscaler spending outlook is not investing. It is hoping.