Salesforce, Snowflake, Marvell Earnings Reaction Guide: 454 Reports on PCE Week

The Wednesday Night Wall: 19 Names After the Bell

After-close reports Wednesday 27 May 2026 — key names and their pre-report context

The heaviest concentration of market-moving names for this week all landed Wednesday after the close. Nineteen companies reported, but five drove the bulk of the attention: Salesforce (CRM), Snowflake (SNOW), Marvell Technology (MRVL), HP Inc (HPQ), and Synopsys (SNPS).

Salesforce came in as the most watched enterprise software name. The stock has been navigating the AI platform arms race from two directions simultaneously: defending its core CRM revenue base from AI-native disruption while trying to position its own Agentforce product as the enterprise answer to that same disruption. There is no comfortable resting place for a $250 billion software company in that kind of market structure transition.

Snowflake reported the same night. The earnings intelligence summary noted the conference call opened with “AI is fundamentally…” — a phrase that tells us everything about how the company wants to frame results but almost nothing about whether the underlying consumption-based revenue model is accelerating or plateauing. Those are very different earnings stories wearing the same AI costume.

Marvell came in after its previous quarter produced a beat-and-raise that the market graded A-minus by market consensus criteria. Custom AI silicon demand from hyperscaler customers was the driver. The earnings intelligence preview noted the prior guidance was received as in-line rather than strong. That framing matters because MRVL’s after-hours reaction showed -4.59% even before the full reaction settled, which suggests the bar was elevated and the print did not clear it convincingly.

HP Inc (HPQ) was a special case. The earnings intelligence feed flagged a potential short squeeze dynamic: “It is almost like everyone shorted HPQ at the very bottom.” That commentary, combined with the note that reports expected an HPQ squeeze, signals the options market had priced a larger-than-normal move to the upside as the most likely path.

Sources: earnings calendar data, earnings intelligence conference call summaries, options data. AH = after-hours. Salesforce/Snowflake full reaction pending Thursday open.

Salesforce versus Snowflake: The Same Budget, Two Very Different Bets

Enterprise AI spending narrative — why both names cannot be right simultaneously

This is the central tension of Wednesday night’s earnings tape. Salesforce and Snowflake compete for the same dollar: enterprise AI platform budget. When a large company decides to build an AI workflow on top of its data, it is choosing between Salesforce’s CRM and Agentforce ecosystem or Snowflake’s data cloud and Cortex AI layer. The board makes one call. They do not buy both at full price.

Salesforce had the bigger structural question to answer. Agentforce is the company’s answer to the AI disruption threat to its core CRM business. The thesis: rather than being replaced by AI-native CRM tools, Salesforce becomes the orchestration layer for enterprise AI agents. That is a compelling pitch. The question is whether it converts to revenue fast enough to offset the slowing growth in the legacy CRM base. The market entered Wednesday expecting clarity on exactly that point.

Snowflake’s narrative is simpler but carries its own risk. The consumption-based model means revenue tracks actual AI workload usage. If enterprises are buying AI licences but not yet running production workloads at scale, Snowflake’s revenue growth stalls even while the AI story sounds brilliant on every conference call. The earnings intelligence summary opens with “AI is fundamentally…” which is precisely the kind of opener that sounds strong but tells you nothing about whether consumption actually accelerated in the quarter.

The sector flow analysis noted that technology ranked tenth of eleven sectors on Wednesday, gaining only 0.54% despite massive institutional dark pool activity in the space. That is the read saying: institutions were positioning in tech names in the dark while the public tape went sideways. The Salesforce and Snowflake earnings were the event those dark pool positions were built around. The reaction at Thursday’s open tells us whether that positioning was front-running a beat or hedging against a miss.

One of these two companies has the stronger hand. They cannot both win the same budget cycle simultaneously. Historically, when both report the same night, the one with the cleaner growth acceleration gets the bid and the other gets sold — not because the loser reported badly, but because the comparison immediately becomes unfavourable.

Analysis only. Not a signal or recommendation to trade either name.

How Options Were Pricing the Reports

Implied move expectations versus the SPY index straddle — the volatility divergence

The options read established that SPY’s expected move heading into Thursday was just 0.13%: the straddle priced an upper bound of $751.42 and lower bound of $749.50. That is a remarkably narrow range for a session containing both PCE and a tech earnings reaction wall. It says the market was pricing calm at the index level.

But individual earnings names tell a completely different story. AMD sat with an expected move of 4.73% and an implied volatility of 65.1% at the money. That single-stock expectation is 36 times wider than the SPY straddle. TSLA priced a 0.21% expected move individually but carried put IV skew at 199.1%, meaning the market was not pricing TSLA through its straddle alone; it was carrying substantial downside tail risk in the structure. The IV skew on AAPL read 191.6% puts-expensive, TSLA 199.1%, QQQ 198.85%.

The divergence between a compressed index straddle and elevated single-stock IV is not a mystery. It is a structural feature of earnings week. Market makers sell index volatility and buy single-stock volatility because earnings dispersion is uncorrelated at the index level. The index pin (SPY max pain $749) holds as long as the earnings moves cancel each other. If multiple names miss simultaneously, the cancellation breaks down and the index absorbs the shock.

That is the specific risk on PCE morning Thursday. Not just that PCE misses, but that PCE misses while Salesforce or Snowflake also guided cautiously, and the two events add rather than offset. The $750 call wall identified in the options analysis becomes the ceiling that falls rather than holds.

Options data from pre-close Wednesday 27 May 2026. Expected moves reflect near-term straddle pricing. IWM is the only instrument with a bearish put-to-call ratio above 1.0.

Thursday Before Open: Retail, Banks, and the PCE Collision

Thursday 28 May 2026 pre-open reports — earnings intelligence confirmed list

Thursday morning delivers a second, equally complex earnings wave before PCE lands. Best Buy, Burlington Stores, Dollar Tree, Kohl’s, Xpeng, Li Auto, Hormel Foods, Royal Bank of Canada, Toronto-Dominion Bank, and Canadian Imperial Bank all report before the open.

The consumer retail cluster: Best Buy, Burlington, Dollar Tree, Kohl’s. These four names collectively read the consumer spending environment. And they report the same morning as the Core PCE print. That is not a coincidence from a market structure standpoint: it means the retail earnings are being interpreted through the same inflation lens as the PCE data. A soft Core PCE number that prints alongside a Best Buy or Dollar Tree guidance raise is a very different market environment from a hot PCE that coincides with Kohl’s cutting full-year expectations. Both events are in the same two-hour window on Thursday morning.

Best Buy is the most watched of the four. The sector flow analysis confirmed that defensives led Wednesday and consumer staples ran +2.38%. Best Buy sits in consumer discretionary but trades defensively in a high-rate environment as a big-ticket electronics retailer. If Best Buy guides cautiously on consumer confidence, that note will hit at the same moment traders are digesting PCE. Double negative signals in the same window tend to overshoot on the downside.

The three Canadian banks: Royal Bank (RY), TD Bank (TD), and CIBC (CM). Canadian bank earnings carry specific relevance here. Canadian housing exposure is the core risk watch for this set. If one or more of the three flags mortgage delinquency acceleration or provision builds, that is a credit signal that reaches beyond Canada’s borders — it is a housing-credit read that European and US financial sectors tend to copy-trade within hours.

MongoDB and Autodesk also report Thursday. MongoDB (MDB) is the NoSQL counterpart to the Snowflake conversation: both are cloud data plays competing for AI infrastructure budget. If Snowflake guidance disappointed Wednesday night, MDB enters Thursday under that shadow regardless of its own fundamentals. Autodesk (ADSK) is an entirely different story — architecture and manufacturing software, where AI integration is slower but margins are structurally high. The read there is simpler: subscription growth and any guidance update on the buyback programme.

Report timing: before open Thursday 28 May 2026 (BBY, BURL, DLTR, KSS, RY, TD, CM, MDB, LI, XPEV, HRL). ADSK and DELL after close Thursday.

Marvell’s -4.59%: What the Semiconductor Read Is Telling Broader Tech

MRVL initial after-hours reaction — the signal for NVDA, AMD, and the semiconductor complex

Marvell’s initial after-hours drop of -4.59% deserves careful interpretation. The instinct is to read it as a miss. That might be wrong.

Marvell’s prior quarter was graded A-minus. That means the results were strong but the grade acknowledged that expectations were already elevated. An A-minus on elevated expectations is a structurally vulnerable setup: any result that does not handily exceed consensus will disappoint the positioning. The stock had run materially into the print on the custom AI silicon story. Positioning was long. The options market had call OI concentrated above the entry price. When a stock is held that way, a beat that does not exceed the whisper number is enough to trigger the unwind.

The broader read: if Marvell, which is genuinely benefiting from custom ASIC orders from hyperscalers, cannot hold its after-hours gain, it says the AI semiconductor narrative is already fully priced at current levels. NVDA sits at $212.60 with max pain at $215 and 21,959 contracts of call OI at $215. The Marvell reaction is not a direct NVDA read, but it is a temperature check on how much further the sector can run on the same AI chip demand story before the options structure rolls over.

Here is where this connects back to what the dark pool analysis identified: $5.31 billion of NVDA was accumulated in the dark on Wednesday. If Marvell’s reaction feeds through to NVDA at Thursday’s open, that $5.31 billion position is immediately underwater. The institutional buyers knew this was the risk. They bought anyway. Either they have information the Marvell print does not change, or they will cover aggressively if NVDA breaks the $210 support level that the puts at $210 (219,471 volume Wednesday) were defending.

All prices from Wednesday close. MRVL reaction -4.59% from earnings intelligence data. AAPL, NVDA prices from Wednesday options market data.

Consumer Earnings as a PCE Pre-Signal

Best Buy, Dollar Tree, Kohl’s: the consumer read before the data lands

PCE at 08:30 EDT Thursday is the official measure of what consumers spent on and how those prices moved in April. But Thursday morning’s consumer earnings reports are the actual companies telling you what they saw in their own sales data. The two sources should rhyme. When they do not, one of them is wrong.

The defensive sector leadership on Wednesday — consumer staples +2.38%, real estate +3.71%, utilities +2.92% — is consistent with a market that expects the consumer to be under pressure. Staples outperform when people trade down. Real estate outperforms when the expectation is that rates are going to fall, which only happens if inflation cooperates. If Best Buy reports soft big-ticket demand on Thursday morning and Core PCE comes in hotter than expected, both signals point the same direction. That is where the market has its worst sessions.

Dick’s Sporting Goods already telegraphed the other side. The Wednesday conference call summary from earnings intelligence: “Sports are so hot right now.” That is not a cautious consumer statement. It is a company telling you that a specific category of discretionary spending is running strong. The question is whether that strength is idiosyncratic to sport or representative of the broader consumer.

Abercrombie’s conference call was more measured: “We are still a healthy business.” That framing is defensive. “We are still healthy” is what companies say when they want to reassure investors without committing to acceleration. The Australian inflation data on Thursday already showed CPI YoY at 4.2%, above the prior 4.6% but running hot by developed-market standards. The French consumer confidence number printed 82 against expectations of 83. Global consumer confidence is softening at the margin. Thursday’s US retail earnings will confirm or contradict that read.

The Tension the Data Cannot Resolve

Here is the honest read, and it cuts against the comfortable narrative.

The options analysis showed that call buying dominated six of ten tracked instruments with a 0.606 aggregate put-to-call. The institutional dark pool analysis showed $27.7 billion of concealed accumulation, concentrated in tech megacaps. Both of those reads say: large money is positioned for continuation. That is the bullish argument, and it is data-supported.

But here is what the earnings tape is saying simultaneously: Marvell, the semiconductor name most directly exposed to the AI infrastructure spending that is supposed to drive the next leg, sold off -4.59% after hours despite posting results that its own graders called A-minus. Abercrombie’s guidance language was defensive. The sector rotation on Wednesday showed defensives leading technology by more than 3 percentage points on the same day that tech dark pool flow was running at decade-high levels. That is not a consistent picture.

The read says institutions are building positions in tech. The price action says tech is underperforming. The earnings reactions say even strong tech results are not clearing the elevated bar. All three of those things cannot be simultaneously true and comfortable for the next two weeks. One of them resolves in the wrong direction. Our honest read is that the most vulnerable scenario is not a single bad PCE print but a cascade: weak PCE data triggers a sell-the-news reaction in rate-sensitives, which hits the same real estate and utilities names that just led the sector table, while the Marvell after-hours print creates sympathy selling in NVDA, and the combination of those two effects arrives at the same $750 SPY ceiling that the call wall analysis identified as the level that breaks when the gamma support flips.

We are not calling for that scenario. We are watching for the conditions that would trigger it.

The 454-Report Week: Full Schedule Context

Week of 25 May 2026 — top names by day and sector concentration

454 reports across the week is an above-average count. For context, a typical Q1 earnings week for S&P 500 names at this point in the cycle runs 60-90 reports. The full 454 includes mid-cap, small-cap, and international names, but the concentration at the top is what matters for market impact.

BO = before open. AH = after hours. Source: earnings calendar data. 454 total reports confirmed for week of 25 May 2026.

The Economic Calendar Collision: PCE and Mortgage Data in the Same Window

Thursday 28 May 2026 — macro events surrounding the earnings wall

Before the US market opens on Thursday, the economic data has already been printing for six hours. Australia’s April CPI came in at 4.2% year-on-year, marginally below the prior 4.6% but above the 4.4% consensus. The RBA trimmed mean core measure held at 3.4%. China industrial profits (YTD) printed 18.2% year-on-year, well above the 12.0% prior — a positive demand signal from the world’s largest manufacturing economy.

The Japan 40-year JGB auction cleared at 3.840%, above the prior 3.600%. That is not a small move. Japan’s long end is where global duration stress has been building. A rising 40-year JGB yield signals that Japan’s bond market is starting to price the BoJ’s exit from yield curve control more aggressively. That reaches US Treasuries through the same carry trade unwind mechanism that periodically jolts risk assets.

At noon US time, the MBA mortgage application data landed: -8.5% week-on-week, with 30-year mortgage rates at 6.65% against 6.56% prior. That is a significant weekly drop in applications, and it comes with rates moving higher. The housing read from those two numbers is simple: affordability is contracting again. Real estate’s +3.71% session gain on Wednesday now looks more fragile as a sustained bid if mortgage applications are collapsing at the rate level.

PCE lands at 08:30 EDT Thursday. Before it arrives, the market has a Japanese duration signal, a soft Australian inflation read that is still above target, Chinese industrial profits beating, French consumer confidence missing by one point, UK gilt yields clearing at 4.550%, and an 8.5% collapse in US mortgage applications. That is not a clean macro backdrop for a bullish earnings season reaction.

All data from Thursday 28 May 2026 economic calendar. PCE is the single most important print but lands into a mixed global macro backdrop.

Three Scenarios for Thursday/Friday: How Earnings and PCE Interact

From the earnings perspective — probability-weighted outcomes

Probabilities represent our analytical weighting of the evidence. Not a recommendation. Analysis only. Scenarios are non-exhaustive.

Earnings Season Position Sizing: What We Are Watching

How we are thinking about allocation into a binary event cluster

Sizing guidance reflects how we are thinking about risk allocation. Not a directive. Individual risk management applies. Analysis only.

Three-Timeframe Earnings Verdict

Continue Reading: The Full 28 May 2026 Analysis

Each section in today’s sequence builds on the last. Start with the dark pool campaigns if you want the institutional positioning context. The options structure and magnetic strikes provide the derivative framework that constrains these earnings moves. The sector rotation confirms where real money was rotating before the prints landed.

• The $27 billion dark pool campaigns and what they signalled before the prints — Institutional Flow
• SPY’s magnetic $750 strike and the options wall that contains Thursday’s range — Option Watch
• How defensives led technology by 3 points on a day the S&P hit record highs — Sector Flow
• The VIX compressed to 16.29 entering a binary session: what that means for stops — Volatility
• Where macro regime sits as PCE arrives on a record-close market — Macro