2026-08-30

Amazon Jumps 4% on Evercore's $355 Target While Nvidia Falls 3.3% on Circular-Financing Worries - Same Friday, Opposite AI Trades

Written & reviewed by the TradeSmrt editorial team · 2026-08-30

In this article

What Happened
Why the Same Week Produced Opposite Verdicts
What to Take Away From This
FAQ
Sources

What Happened

Friday, August 28 produced one of the more instructive split screens of the AI trade this year. Amazon (Nasdaq: AMZN) closed at $266.43, up 3.97% on the day, after Evercore ISI raised its price target on the stock to $355 from $315 while keeping an Outperform rating - a target that implies roughly 40% upside from Thursday's close. A few desks over, Nvidia (Nasdaq: NVDA) closed down about 3.3%, giving back a chunk of the 8.7% surge it had posted just one session earlier on the back of a blowout earnings report. The Philadelphia Semiconductor Index (SOX), which tracks the broader chip sector, fell 1.54% on the day, a decline that swept up Intel, AMD, and Broadcom alongside Nvidia.

What makes the split notable is that both moves happened inside the same 24-hour window, against the same macro backdrop - Fed Chair Kevin Warsh's hawkish Jackson Hole remarks, which pushed September rate-hike odds from the mid-30% range to somewhere between 56% and 59% and dragged the S&P 500 down 0.25% to 7,711.76 and the Nasdaq Composite down 0.52% to 26,402.42 for the session. A rate-hike scare that size would normally be expected to hit growth stocks broadly and more or less evenly. Instead, one of the largest AI-linked names in the market rallied hard enough to shrug the macro news off entirely, while another AI-linked name - one that had just reported one of the strongest quarters in its history - sold off on top of the same news. The difference wasn't the macro backdrop. It was two entirely separate, company-specific stories that happened to land on the same trading day.

Amazon's move traced back to a research note, not a company announcement. Evercore ISI's 14th annual U.S. Online Retail Survey turned up what the firm called the first real survey evidence that agentic AI is additive to Amazon's core retail business, rather than just a cost center or a cloud-revenue story. The headline data point: 57% of Alexa AI users reported having purchased a product they hadn't previously known existed - meaning the AI assistant was surfacing new purchase intent rather than simply automating searches shoppers would have made anyway. The same survey found Amazon's overall retail reach at 92%, a 34-point lead over Walmart, alongside rising adoption of same-day delivery (49% regular usage) and Prime members spending 3.1 times more than non-Prime shoppers. Evercore's argument, in short, was that Amazon's AI investments are starting to show up as incremental retail revenue on top of the already-established AWS cloud story - two growth engines instead of one.

Nvidia's decline, by contrast, had nothing to do with its own numbers, which were genuinely strong: the company had guided for roughly $108 billion in October-quarter revenue, ahead of Wall Street's $104.19 billion consensus, and Stifel and Citi both raised price targets to $315 in response (Citi's implying about 38% upside). The pullback instead traced back to a debate that has dogged Nvidia on and off for most of the summer - "circular financing," the practice of Nvidia investing in or guaranteeing financing for AI labs and data-center operators (most prominently OpenAI) who then use that capital, directly or indirectly, to buy Nvidia chips. Nvidia CEO Jensen Huang had pushed back publicly earlier in the week, rejecting the "circular financing" label for the company's support of OpenAI's Ohio data-center buildout and arguing that OpenAI's broader infrastructure plans could translate into roughly $600 billion of Nvidia compute purchases through 2030. But the underlying arithmetic behind that defense had already drawn scrutiny: the originally floated $600 billion financing figure from July had been reported down to a signed guarantee of $105 billion by mid-August - a gap large enough that some investors read it as evidence the eye-popping headline numbers around AI capex commitments don't always survive contact with a signed contract.

Why the Same Week Produced Opposite Verdicts

The mechanism connecting these two moves is worth pulling apart, because it says something about where investor scrutiny in the AI trade is currently concentrated. Broadly speaking, AI-related stocks fall into two rough camps: companies whose AI spending is a cost that needs to be justified by external, arm's-length revenue, and companies whose AI investment is showing up as revenue growth from real customers paying with real, independently-sourced money. Amazon's Evercore upgrade landed squarely in the second camp - it wasn't a story about Amazon spending more on AI, it was survey evidence that AI features are pulling in incremental purchases from ordinary shoppers, dollars that have nothing to do with Amazon's own AI infrastructure spending. That is about as external and arm's-length as revenue evidence gets.

Nvidia's circular-financing overhang sits in exactly the opposite camp. The core anxiety isn't about whether AI chips are in demand - Nvidia's own $108 billion guidance answers that question emphatically. The anxiety is about the quality and durability of the demand: if a meaningful share of the capital flowing to Nvidia's customers originated, even indirectly, from Nvidia itself (through investments, financing guarantees, or compute-purchase commitments bundled into funding deals), then reported revenue growth becomes harder to read as pure, independent end-market demand. It's not an accusation that Nvidia is doing anything improper - vendor financing has existed for decades across many industries - but it does mean that every dollar of guided revenue draws a follow-up question that a company like Amazon's retail division, selling directly to millions of individual consumers, doesn't have to field.

This is also why the reaction sizes differed the way they did. Amazon's move was driven by one analyst survey - a meaningfully positive data point, but a single note. Nvidia's move was driven by a structural question that has already survived multiple rounds of company pushback without fully resolving, meaning each new data point (the Ohio guarantee shrinking from an earlier headline number, Huang's on-stage defense, this week's circular-financing headlines resurfacing) gets weighed against a backdrop of accumulated investor skepticism rather than evaluated fresh. A single bullish survey can move a stock 4% in a day; a question that keeps recurring without a clean resolution tends to produce exactly the kind of choppy, give-back-the-rally price action Nvidia experienced going from an 8.7% post-earnings pop on Thursday to a 3.3% pullback on Friday.

It's also worth noting what didn't happen: this wasn't a sector-wide unwind. Software and consumer-internet-adjacent AI plays with clearer paths to external revenue generally held up far better than semiconductor and infrastructure names on Friday, reinforcing that the divide running through the AI trade right now isn't "AI stocks up" versus "AI stocks down" - it's a divide between AI-exposed companies whose growth is easy to verify against independent revenue and those whose growth is harder to disentangle from money that started somewhere inside the same ecosystem.

What to Take Away From This

The same macro headline can move two AI-exposed stocks in opposite directions when their underlying stories are different. Warsh's hawkish Jackson Hole remarks hit the broad market, but they didn't determine whether Amazon or Nvidia would end the day green or red - company-specific narratives did.
Revenue quality matters as much as revenue growth. Nvidia's $108 billion guidance beat consensus comfortably, yet the stock still fell, because a meaningful slice of the market is currently more focused on where the money funding that growth originates than on the growth figure itself.
A single research note can move a stock as much as a full earnings report when it answers a question the market was already asking. Evercore's survey didn't reveal new Amazon financials - it supplied outside evidence for a thesis (agentic AI driving incremental retail spending) that investors had been debating without hard data.
A recurring, unresolved narrative tends to produce more volatile price action than a one-time data point. Circular-financing concerns around Nvidia have resurfaced multiple times this summer without a clean resolution, which is part of why the stock's reaction to good news has been choppier than the headline numbers alone would suggest.

FAQ

Does Nvidia's Friday decline mean the company is doing something wrong?

Not necessarily. Vendor financing and strategic investment in customers are common practices across many capital-intensive industries, and Nvidia's own revenue guidance was strong. The concern investors are pricing in is about how independently verifiable that demand is - not an allegation of wrongdoing.

Is Amazon's stock move sustainable, or was it a one-day pop on a single analyst note?

It's one data point, and a 40%-implied-upside price target from a single firm is not guaranteed to hold. What makes it more durable than a typical analyst-note pop is that it was backed by original survey data (the 57% Alexa-purchase figure) rather than just a valuation re-rating, giving other investors and analysts something concrete to independently check against future Amazon earnings.

Should investors treat "circular financing" as a reason to avoid AI infrastructure stocks entirely?

That's a broader judgment call that depends on an individual company's specific financing structure and disclosure, not something this one Friday's price action settles by itself. It's a reasonable factor to weigh alongside growth and margin trends, but Nvidia's core chip demand and guidance remain, by the numbers reported, genuinely strong.

Sources

This article is an original synthesis and analysis based on the reporting below, not a reproduction of the original articles. Please check the source articles directly for the most current figures and details.

⚠️ This article is for informational purposes only and is not investment advice. Market conditions change constantly - always verify the latest data before making investment decisions.