Week Ahead: Nvidia, PCE and Warsh Test the Rate-Strained Rally

The Setup

The market enters August 24-28 with a three-part burden of proof. Nvidia must show that artificial-intelligence spending can still outrun extraordinary expectations, Wednesday's core PCE report must validate the recent easing in price pressure, and Federal Reserve Chair Kevin Warsh must explain on Friday how policy will respond to inflation while long-term borrowing costs remain restrictive. Any one of those catalysts can move a sector. Together they can reset the valuation framework for the whole index.

The starting point is stable rather than comfortable. The S&P 500 stands at 7,674.37 after a 1.4% weekly decline, the Nasdaq Composite at 26,180.46 after losing 2.1% for the week, and the Russell 2000 at 3,017.87 after a 1.6% weekly pullback. Friday's recovery kept the S&P 500 12.1% higher for 2026, but it did not remove the pressure from a 4.74% 10-year Treasury yield and a 5.27% 30-year yield. The latest closing levels therefore describe a market with positive trend support and a high discount-rate hurdle.

Technical positioning makes the split especially clear. SPY closed at $765.72, 1.9% above its $751.75 50-day average and 8.2% above its $707.54 200-day average, while its 14-day RSI is a neutral 54.23. QQQ closed at $713.44, effectively on its $713.44 50-day average, with a 50.15 RSI. The VIX at 15.13 and the dollar index at 98.80 argue against broad panic, but QQQ's loss of technical cushion says the market will demand earnings confirmation from the companies carrying the growth premium.

The core tension is that economic momentum and rate relief no longer point in the same direction. August flash services PMI accelerated to 56.8 and the composite index reached 56.0, consistent with solid third-quarter growth, while manufacturing remained expansionary at 53.2. Strong activity supports earnings, yet it also makes it harder for the bond market to dismiss inflation and term-premium risk. This week can resolve that tension only if core PCE cools without a downward GDP revision and Nvidia's forward demand remains strong.

The Market Dashboard

The dashboard shows why this is a confirmation week rather than a simple dip-buying setup. Equity indexes retain double-digit year-to-date gains, credit spreads remain contained and volatility is low. At the same time, oil, the two-year yield and the long end of the Treasury curve are all materially above their year-opening levels.

Treasury's official August 21 curve puts the three-month yield at 3.88%, the two-year at 4.24%, the 10-year at 4.74% and the 30-year at 5.27%. That upward slope is not a recession alarm by itself; it is a warning that term premium and inflation uncertainty remain expensive. A sustained 10-year break above 4.80% would likely matter more for equity multiples than a small upside surprise in any single earnings report.

The Economic Calendar

The calendar is concentrated around Wednesday's 8:30 a.m. ET data cluster and Friday's 10:00 a.m. Jackson Hole address. Tuesday supplies a consumer and housing read, Thursday checks labor-market stability, and Friday pairs a weak sentiment baseline with the Fed chair's reaction function. The BEA schedule confirms that the July personal-income report and the second estimate of second-quarter GDP arrive together Wednesday.

Core PCE is the highest-impact release because the expected 0.2% monthly increase and 3.2% yearly pace would cool the inflation narrative without creating a deflation scare. A 0.1% monthly reading with stable spending would favor long-duration growth and Treasuries. A 0.3% or higher print would strengthen the case for a September hike, especially with oil still above $87. The key composition question is whether services prices are cooling beyond travel while goods inflation remains contained.

GDP matters through the mix, not the headline alone. The advance estimate put second-quarter annualized growth at 1.5%. An unchanged headline with stronger final demand and corporate profits would support cyclicals and smaller companies. A downward revision led by consumption would turn softer PCE into an earnings concern. Tuesday's housing releases and Thursday's claims then determine whether that weakness is broadening or remains concentrated in rate-sensitive activity.

Earnings in Focus

Wednesday after the close is the earnings center of gravity. Nvidia sets the standard for AI infrastructure, Salesforce and CrowdStrike test enterprise software demand, and Synopsys measures semiconductor-design investment. Marvell follows Thursday with a direct read on custom accelerators and networking. The August 26 and August 27 calendars supply the current consensus estimates below.

Nvidia's $92.04 billion revenue consensus implies roughly 97% yearly growth, so another nominal beat is not enough. The market needs evidence that deployment of its newest systems is scaling without a damaging margin trade-off, plus a third-quarter outlook that supports the next leg of consensus estimates. The company will post results around 4:20 p.m. ET and host its 5:00 p.m. ET call. Watch data-center revenue, gross margin, supply timing and the handoff from current platforms to Vera Rubin. Nvidia's estimate momentum is the strongest in the group, which increases the bar for a positive share reaction.

Salesforce is the value counterweight within software. At about 13.5 times next-year consensus earnings, it needs durable current remaining performance obligation growth and proof that Agentforce adoption is producing revenue rather than only engagement. CrowdStrike has the opposite setup: revenue growth near 25% and a clean 4-for-4 beat record support the business, but a forward multiple above 120 times leaves little tolerance for weaker net new annual recurring revenue. Synopsys must show that electronic-design automation demand can offset integration and execution concerns.

Marvell's $2.71 billion consensus calls for 34.9% revenue growth and 38.8% EPS growth. The important number is management's third-quarter revenue guide: expectations near $3.0 billion imply another double-digit sequential step. A strong guide would broaden AI leadership beyond Nvidia; a weak one would suggest that custom-accelerator and networking expectations have moved ahead of the delivery schedule.

Thursday morning adds a lower-income and discretionary demand check. Dollar General is expected to earn $2.01 on $11.19 billion of revenue, Dollar Tree $1.13 on $4.85 billion, Best Buy $1.38 on $9.58 billion and Burlington $2.19 on $3.02 billion. Traffic, shrink, tariff costs and inventory discipline matter more than headline revenue. Dollar stores outperforming electronics and apparel would confirm trade-down behavior, not broad consumer acceleration.

Fed Watch & Rate Markets

The federal-funds target remains 3.50%-3.75%, and futures still favor patience in September. Pricing derived from CME 30-day fed-funds futures assigns a 61.0% probability to no change at the September 16 meeting and 39.0% to a 25-basis-point hike. By October 28, the distribution is 47.0% at the current range, 44.1% one hike higher and 9.0% two hikes higher. By December 9, only 28.9% remains at the current range, with 45.2% at 3.75%-4.00%, 22.5% at 4.00%-4.25% and 3.5% at 4.25%-4.50%. The rate monitor was updated August 22 and is based on CME futures prices.

Those probabilities make Friday's speech a communication test. The July minutes showed that the 9-3 decision to hold included three officials who preferred a hike, while other participants warned that higher rates could become necessary if inflation stayed elevated. Warsh does not need to pre-commit to September. He does need to explain which evidence would outweigh the other side of the mandate: a hotter PCE print, a reacceleration in inflation expectations, weaker hiring or a renewed oil shock.

The curve gives him little room for ambiguity. At 4.24%, the two-year yield already prices meaningful tightening risk. At 4.74%, the 10-year embeds both policy uncertainty and term premium. The latest available ICE BofA high-yield option-adjusted spread is only 2.75 percentage points, while the 30-year mortgage rate is 6.65%, according to FRED credit data and the Freddie Mac survey. Corporate credit remains calm even as households face restrictive borrowing costs.

Sector & Asset Class Radar

Five areas sit at inflection points, and relative performance will reveal whether the market is treating this as a growth confirmation week or a valuation reset. The strongest confirmation would be simultaneous breadth across chips, software, consumer cyclicals and smaller companies.

Semiconductors and AI infrastructure

Nvidia and Marvell can either repair breadth or deepen concentration risk. Nvidia at $214.72 has cooled into earnings, while Marvell is still up nearly 179% for 2026 despite Friday's 5.6% decline. The constructive signal is both stocks rising on guidance, with broader chip participation. Nvidia rising while suppliers and networking names fall would say the market trusts only the dominant platform.

Enterprise software and cybersecurity

Salesforce, CrowdStrike and Synopsys separate profitable software durability from multiple dependence. Salesforce has rallied above its 200-day average but remains down 21% for the year, while CrowdStrike is up almost 64% and carries a far richer earnings multiple. Strong bookings across both would confirm healthy enterprise budgets. Divergence would favor cash-flow discipline over pure growth.

Consumer and retail

Dollar General, Dollar Tree, Best Buy and Burlington will extend the retail evidence from the prior week. Favor companies with traffic growth, clean inventory and limited price-based sales gains. A nominal sales beat driven mainly by tariffs or price increases is a weaker signal than unit growth. Consumer discretionary can recover if guidance stabilizes and the 10-year yield moves below 4.65%.

Energy, gold and duration

Energy remains the earnings hedge against a failed disinflation outcome, but XLE's 71.73 RSI warns against chasing the sector after a 42.3% year-to-date advance. Gold's RSI near 73.7 also signals crowded hedging demand. Long-duration Treasuries are the cleaner tactical beneficiary of soft PCE only if oil also retreats; otherwise a low inflation print may not be enough to compress term premium.

Small caps and financials

Russell 2000 leadership depends on the difference between a steep curve and a disorderly curve. Banks benefit when long yields rise for growth reasons and credit stays healthy. Smaller companies suffer when yields rise because of inflation or fiscal credibility. A Russell breakout above 3,094 with high-yield spreads below 3.0 points would confirm broadening. Failure below 2,985 would favor quality large caps.

Geopolitical & Policy Risk Monitor

The major wildcards sit outside the scheduled calendar but feed directly into oil, inflation and discount rates. Investors should treat the probability labels as portfolio-planning judgments, not event forecasts.

Oil is the fastest transmission channel. WTI above $87 and Brent above $94 can lift near-term inflation expectations, squeeze retailers and airlines, and keep the long end resistant to soft data. A Brent move through $96 would raise the probability that even a 0.2% core PCE print fails to help duration. A retreat below $90 would make the disinflation signal more credible.

Tariff pass-through is slower but broader. Retailers can protect gross margin through sourcing, productivity, lower promotions or price increases. Only the last choice directly challenges the disinflation narrative. The market should reward companies that preserve margin without sacrificing traffic and punish businesses that report nominal growth with weaker units.

Technical Levels to Watch

SPY remains above both major moving averages, while QQQ is sitting directly on its 50-day trend. That makes the week asymmetric: the broad market has room to absorb a moderate disappointment, but technology needs a quick reclaim of momentum after Nvidia and the software reports. Bollinger-band width is moderate rather than extreme, leaving space for an event-driven expansion.

For SPY, $776 is the first resistance zone and the $790.51 upper band is the momentum target. The $751.75 50-day average is first major support. A daily close below it would expose the $734.14 lower band and argue that the market is discounting both earnings and rate risk. Holding above $751.75 through Wednesday's data and earnings would keep the medium-term trend constructive.

For QQQ, $713.44 is both the close and the 50-day average, so it is the immediate pivot. A reclaim of $730 would restore momentum, and $748.72 is the upper-band objective. Failure to hold $700 after Nvidia would make $669.66 the next statistically relevant support. RSI near 50 gives neither side a momentum advantage before the catalyst.

The AlphaEdge Outlook

The primary thesis is constructive but selective. The index uptrend can survive a high-rate environment if Nvidia validates AI demand, software guidance stays durable and core PCE slows to 0.2%. That combination would support earnings without forcing the Fed toward a September hike. It would also favor a rotation from defensive mega-cap concentration into profitable software, industrial technology and selected small caps.

The thesis changes if core PCE reaches 0.3% or more while Nvidia or Marvell issues a cautious forward guide. That pairing would attack both sides of valuation: a higher discount rate and weaker expected cash-flow growth. In that scenario, low VIX and tight credit spreads would represent complacency, and SPY's 50-day average would become a downside checkpoint rather than a buying signal.

Portfolio construction should preserve optionality. Keep core exposure to profitable large caps, but avoid adding aggressively to semiconductor momentum before the reports. Pair growth exposure with cash-generating financials or energy rather than treating oil as a separate macro story. For rate-sensitive assets, wait for the 10-year to close below 4.65% before assuming that soft inflation will produce durable multiple expansion.

The contrarian opportunity is in evidence that looks only moderately good. A 0.2% core PCE print, unchanged GDP and solid rather than spectacular Nvidia guidance could be healthier for market breadth than another extreme beat. It would reduce the need to chase one company while keeping the earnings and policy paths intact. Conversely, a huge Nvidia beat paired with a hawkish Warsh speech could lift the stock and still leave the broader index weaker.

Bottom line: Stay constructive while SPY holds $751.75, but require a cool core PCE print, durable AI guidance and a clear Jackson Hole reaction function before treating the August 24-28 week as a durable return to risk-on leadership.