Research cutoff: Saturday, August 29, 2026, roughly 08:30 BST.
U.S. price cutoff: Friday, August 28 regular-session close.
Horizon: roughly the next 1–7 U.S. trading sessions.
Weekend note: U.S. markets are closed today, so this report is based on Friday's close and the setup for next week rather than pretending there is a Saturday premarket.
The quick read
Friday gave us a useful reality check.
Nvidia's earnings on Wednesday night were strong enough to restart the AI trade on Thursday, but Federal Reserve Chair Kevin Warsh reminded investors on Friday that excellent AI demand does not cancel out the interest-rate problem.
The result was a split market rather than a broad tech collapse:
- Nvidia fell 4.6% after Thursday's 8.7% earnings rally.
- AMD fell 2.3% and Broadcom slipped modestly.
- Microsoft, Alphabet, Apple, Meta and Amazon all finished higher.
- Amazon was the standout of the megacaps, up almost 4%.
- The Nasdaq lost 0.52%, but still gained 0.85% for the week.
That distinction matters. The market did not reject the AI demand story. It repriced the most crowded semiconductor trade while rates moved higher and capital rotated toward other large technology platforms.
My working view for next week is therefore:
Do not treat Friday's chip weakness as proof that AI is rolling over. But do not automatically buy every semiconductor dip either. Relative strength is currently cleaner in Microsoft, Google and Amazon.
The two events that can change that view quickly are Broadcom earnings on Wednesday, September 2 and the August U.S. jobs report on Friday, September 4.
Market pulse
| Market | Friday close | Friday move | Weekly move | What it says |
|---|---|---|---|---|
| S&P 500 | 7,711.76 | -0.25% | +0.49% | Near records, but rates remain a valuation headwind |
| Nasdaq Composite | 26,402.42 | -0.52% | +0.85% | AI/tech trend intact, Friday breadth was weak |
| Dow Jones | 53,559.99 | -0.02% | +0.53% | Essentially flat Friday |
Friday's breadth was worse than those modest index declines suggest. Reuters counted 1,494 Nasdaq advancers versus 3,272 decliners. That is not panic, but it is a reminder that a handful of megacaps are doing a lot of work.
Warsh was the macro catalyst. He said the Fed still has "more work to do" unless inflation is clearly moving toward the 2% target. Late Friday, fed-funds futures implied roughly a 57% probability of a September rate hike, according to Reuters.
That keeps long-duration growth stocks sensitive to every major economic print from here.
The tech board
| Stock | Friday close | Friday move | 1–7 day read |
|---|---|---|---|
| NVDA | $217.55 | -4.57% | Fundamentals excellent; post-earnings volatility high |
| MSFT | $513.53 | +1.68% | Strongest clean setup |
| AAPL | $319.70 | +1.63% | Relative strength + Sept. 9 product catalyst |
| GOOGL | $346.59 | +1.74% | Catch-up candidate; regulatory risk eased |
| META | $578.02 | +1.21% | Constructive, but less urgent catalyst-wise |
| AMZN | $266.43 | +3.97% | Friday leader; attractive if breakout holds |
| TSLA | $348.75 | -1.71% | Relative weakness; needs technical repair |
| AMD | $465.58 | -2.33% | AI exposure strong, price needs stabilization |
| AVGO | $368.79 | -0.74% | Major event risk: earnings Wednesday |
The most interesting part of the table is not Nvidia's decline. It is the divergence between semiconductors and the big cloud/platform companies.
Nvidia: the earnings story is still extremely strong
Nvidia's Wednesday report was not a marginal beat.
The company reported quarterly revenue of about $96.2 billion, with data-center revenue around $89 billion, and guided the following quarter to approximately $108 billion. Management also gave unusually long-range guidance indicating roughly 70% revenue growth for the next fiscal year, far above the growth rate Wall Street had been modeling.
That is why NVDA jumped 8.7% Thursday.
Friday's 4.6% decline therefore looks more like a combination of profit-taking, a rates shock and post-earnings digestion than evidence that the underlying AI cycle suddenly broke.
The caution is on the supply side. Nvidia has warned that memory and component shortages can constrain how quickly it converts demand into shipments, and its outlook still excludes China data-center revenue because export rules remain uncertain.
My NVDA decision zones
These are not analyst targets. They are simply short-term price areas derived from this week's trading.
- $215–218: first area I want to see hold after Friday's selloff.
- $228–230: the first meaningful reclaim area; Thursday closed near $228 and traded above $230 intraday.
- A clean break below the lower zone would make me stop treating every red day as a simple buy-the-dip opportunity.
For a 1–7 day trade, I prefer proof of stabilization over chasing the fundamental story.
Microsoft: probably the cleanest setup on the board
Microsoft finished Friday at $513.53, up 1.68%, while the Nasdaq fell.
That relative strength matters more to me than another semiconductor sympathy move.
The broader software narrative also improved materially this week. Salesforce surged after raising its annual forecasts and expanding its AI partnership with Anthropic, while CrowdStrike rallied after raising its own outlook. Reuters noted that those results helped calm the fear that generative AI simply destroys incumbent software economics.
For Microsoft, that is a helpful backdrop: it sits on both sides of the AI trade, selling cloud compute while also monetizing AI through software and enterprise distribution.
MSFT levels
- $505 area: first pullback zone. Friday's low was about $504.87 and the prior close was roughly $505.
- $517–518: Friday's high and the nearest obvious breakout test.
If the market gets a rates-driven pullback next week and MSFT holds around $505 while the Nasdaq is weak, that is the kind of relative strength I would rather buy than a random semiconductor bounce.
Amazon: Friday's strongest megacap
Amazon rose 3.97% to $266.43, making it the clear Friday leader among the large names in this report.
I would not invent a single-company headline to explain all of that move. The more useful context is that Nvidia's outlook reinforced the scale of hyperscaler AI infrastructure demand. Nvidia also said AWS and Nvidia plan to deploy another 2 million GPUs across global infrastructure during 2027–2028.
That supports the longer-term AWS AI-compute thesis even if it does not explain every percentage point of Friday's rally.
AMZN levels
- $257–260: the breakout-support area I would like to see hold on a pullback.
- $267–268: Friday's high zone and immediate breakout test.
The risk now is simply chasing. A nearly 4% one-day move is much less attractive to buy at the top than after a controlled retest.
Google: still one of my preferred catch-up ideas
Alphabet closed at $346.59, up 1.74% on Friday while the Nasdaq fell.
There was also a genuine company-specific positive development. Google changed its European spam/search enforcement policy to address EU Digital Markets Act concerns. The adjustment removes certain manual demotion actions for EEA users and publishers, reducing the immediate risk of a DMA penalty around this specific case. The European Commission said it will continue monitoring implementation.
That does not remove Alphabet's broader regulatory risk, but it does remove one near-term problem without damaging the core AI thesis.
GOOGL levels
- $340–342: first support/observation area, around Friday's low.
- $349–350: Friday's high and the next breakout test.
This remains attractive to me because Google combines improving relative strength with a stock that has not had the same explosive post-earnings move as Nvidia.
Apple: a different catalyst from the AI-chip trade
Apple rose 1.63% to $319.70 on Friday.
Its next obvious catalyst is already on the calendar: Apple has officially scheduled a special event for September 9, and Reuters expects the company to unveil its next iPhone lineup, with a potential first foldable iPhone among the products investors are watching.
Apple also refreshed its Mac mini and Mac Studio lineup this week with newer chips and a stronger AI-performance pitch.
This makes AAPL a useful diversifier inside a tech watchlist. Its near-term setup is not dependent on Broadcom matching Nvidia's AI forecast.
I would not chase simply because the event is approaching, but relative strength into an announced catalyst is better than owning a weak chart and hoping the catalyst repairs it.
Meta: constructive but not my first choice next week
Meta finished at $578.02, up 1.21%.
The important thing here is relative behavior: Meta joined Microsoft, Google, Apple and Amazon on the positive side Friday even as the broader Nasdaq slipped.
For a short-term trade, I would use Friday's range rather than manufacture a new fundamental catalyst:
- $571–572: first area where I want buyers to appear.
- $589–590: Friday's high zone and near-term breakout test.
Between those levels, I view META as constructive but less compelling than MSFT or GOOGL because the next immediate catalyst is less obvious.
Broadcom: next week's most important single-stock test
Broadcom closed at $368.79, down 0.74% Friday. The bigger issue is not the move. It is Wednesday's earnings report.
Broadcom is one of the cleanest tests of whether Nvidia's demand visibility extends to custom AI accelerators and networking.
There is a useful warning from Marvell. Marvell fell more than 10% Friday despite strong results because investors disliked the timing of revenue from its Google AI-chip agreement. Reuters reported that meaningful contributions from that deal are not expected until fiscal 2029.
The lesson is straightforward:
The market still believes in custom AI silicon, but it is no longer rewarding a huge future opportunity if the revenue arrives later than investors hoped.
That makes Broadcom's guidance and timing commentary at least as important as the headline quarter.
AVGO levels before earnings
- $365–369: Friday's support/close area.
- $376–377: Friday's high zone.
But those levels have limited value through earnings because the stock can gap straight through them. I would treat AVGO as an event watch, not a normal technical trade, until the report is out.
AMD: this dip needs proof
AMD fell 2.33% to $465.58 Friday.
That makes it tempting as a catch-up trade because Nvidia just validated AI demand. But the price action itself has not confirmed the thesis yet.
- Around $465: first stabilization area because Friday closed almost on the session low.
- $478–480: first useful reclaim area.
If AMD cannot stabilize while software/cloud megacaps remain strong, I would rather wait than assume semiconductor weakness must immediately mean opportunity.
Tesla: wait for repair rather than predict the bottom
Tesla fell 1.71% to $348.75 Friday.
Its setup is different from Nvidia, Microsoft or Google. Tesla is particularly sensitive to rates and its valuation leaves very little room for a macro shock to be ignored.
Friday opened around $357 and finished below $350. I would therefore want to see roughly $355–357 reclaimed before calling the short-term chart repaired.
Until then, this is lower on my list. There is no reason to force a Tesla trade when other tech names are showing cleaner relative strength.
Relative-strength screen: where money actually went Friday
Friday's ranking is revealing:
Relative leaders
- AMZN +3.97%
- GOOGL +1.74%
- MSFT +1.68%
- AAPL +1.63%
- META +1.21%
Relative laggards
- AVGO -0.74%
- TSLA -1.71%
- AMD -2.33%
- NVDA -4.57%
I do not automatically read the bottom group as "cheap" and the top group as "too late."
The better question is whether the laggard has a reason to catch up and whether the price starts confirming it.
That leaves me with three different trade types:
- MSFT / GOOGL: clean relative-strength setups.
- AMZN: breakout/momentum setup that needs a controlled retest.
- NVDA / AMD: potential dip trades that need stabilization first.
- AVGO: event-driven setup where Wednesday's guidance matters more than Friday's chart.
The macro risk is not finished
Warsh's speech moved the rate debate back to the center of the market.
The August payroll report is due Friday, September 4. A Reuters poll currently expects roughly 58,000 jobs added and unemployment of 4.1%, following a surprising 23,000 decline in employment in July.
The market problem is slightly awkward:
- A very weak jobs report can raise growth concerns.
- A very strong jobs/wages report can strengthen the case for a Fed hike.
- A middle-of-the-road report is probably the friendliest outcome for expensive technology stocks.
Manufacturing and services data also arrive during the week.
There is another caution hiding under the index: global equity funds snapped a 13-week inflow streak in the week through August 26, according to Reuters. U.S. equity funds saw about $22.3 billion of net outflows, even while technology funds continued attracting money.
That is a market where investors still want AI exposure, but are becoming more selective about the rest of the risk they are carrying.
Saturday AI ecosystem note: OpenAI vs Cursor
One fresh development arrived after Friday's close.
Reuters reported Saturday that OpenAI plans to stop supplying models to Cursor after the coding-tool company became owned by SpaceX. OpenAI has proposed a November 12 cutoff, while Cursor says discussions are still ongoing. Anthropic, meanwhile, said it plans to increase Claude compute support for Cursor.
This is not a direct trading catalyst for the nine public stocks above, but it is worth watching for the broader AI platform battle. Model access is becoming a strategic distribution weapon, not merely an API-price competition.
For Microsoft, Amazon and Google, that matters because cloud/model distribution increasingly determines where developer workloads land.
What I'd watch next week
This is a watch ranking, not a recommendation to buy all nine names at Monday's open.
1. Broadcom — biggest event
Wednesday earnings can either confirm Nvidia's AI visibility across custom silicon/networking or expose the same timing problem that hit Marvell. I would rather react to the report than gamble in front of it.
2. Microsoft — best clean setup
Strong Friday relative performance, improving software sentiment and a clear $505 pullback zone. This is my favorite normal 1–7 day setup if rates cause a broad dip.
3. Google — best catch-up setup
Friday relative strength plus a small regulatory de-risking event. I like $340–342 as the first area to judge whether buyers are still there.
4. Amazon — strongest momentum
The move is excellent; the entry is less excellent after +4%. I want either a hold above the breakout or a retest toward $257–260.
5. Nvidia — strongest fundamentals, messiest entry
The earnings story remains exceptional. I simply do not want to confuse "great company" with "must buy this exact candle." The $215–218 hold or a $228–230 reclaim would tell me more.
6. Apple — catalyst approaching
Relative strength plus the September 9 event makes AAPL worth keeping on the screen even though it is not the purest AI trade.
7. Meta — constructive middle ground
Friday was good. I would become more interested on a controlled hold around $571–572 or a convincing push through the Friday high zone.
8. AMD — needs stabilization
Great industry exposure, weak Friday tape. Let the chart prove the dip is being bought.
9. Tesla — weakest setup for me
I want the $355–357 area back before treating the recent weakness as repaired.
Bottom line
The easiest mistake after Thursday would have been: Nvidia crushed earnings, therefore buy every AI stock.
Friday showed why that is too simple.
AI demand remains extremely strong. Nvidia's forecast makes that hard to dispute. But the Fed has reintroduced a real discount-rate problem, and investors are distinguishing between companies with immediate monetization, companies whose AI revenue is farther out, and companies whose valuations leave no margin for error.
For me, the best risk/reward next week is not chasing the stock with the biggest earnings headline.
It is watching for:
- MSFT around $505 if the market pulls back;
- GOOGL around $340–342 if relative strength holds;
- AMZN $257–260 on a clean retest;
- NVDA $215–218 only if buyers actually defend it;
- and AVGO after Wednesday's earnings, when we know whether Nvidia's visibility is translating cleanly into Broadcom's part of the AI stack.
Then Friday's jobs report can reset the whole rate equation again.
Sources
Checked August 29, 2026:
Markets and macro
- Reuters — Wall Street ends lower after Fed Chair Warsh reaffirms inflation fight
- Reuters — Wall St Week Ahead: Jobs report, Broadcom results pose next hurdles for stock market rally
- Reuters — Global equity funds snap 13-week inflow streak
AI, chips and software
- Reuters — Nvidia forecast lifts Nasdaq and S&P 500 after earnings
- Reuters — Nvidia forecasts quarterly revenue above estimates
- Reuters — Marvell selloff deepens as investors seek clarity on Google AI deal payoff
- Reuters — Salesforce raises annual forecasts and expands Anthropic AI partnership
- Reuters — CrowdStrike raises annual revenue forecast on strong cybersecurity demand
Company catalysts
- Reuters — Google changes spam policy in EU to avert antitrust fine
- Apple — September 9 special event
- Reuters — Apple sets September 9 date for next iPhone launch event
- Reuters — Apple launches faster Mac mini and Mac Studio
- Reuters — OpenAI to cut off AI models for SpaceX-owned Cursor
Market research only, not individualized financial advice. Short-term levels are observation zones derived from recent trading, not guarantees; earnings and macro data can gap prices through them.
