Beneath the Records: AI Chips Split, Energy Unwinds

The S&P 500 closed Wednesday at $769.77 on the SPY, down just 0.20%^[1] — barely a flicker after Tuesday’s record-setting 1.79% surge to 7,736.52^[2]. But the flat headline masks a day of violent crosscurrents underneath. The Dow gained 0.44%^[1] while the Nasdaq-100 tracker QQQ fell 0.90%^[1], and the gap between them — the Dow outperforming the Nasdaq by roughly 130 basis points — is the real story of the session.

What drove the split? Three forces pulled in different directions: an AI chip war that rewarded Nvidia and punished AMD, an Alphabet brain drain that lopped 4% off one of the market’s largest weights, and a geopolitical de-escalation in the Strait of Hormuz that gutted energy stocks. The VIX sat at 15.99^[3] through it all — markets are not pricing fear, they are pricing reallocation.

The AI Chip Split: SpaceX Picks a Winner

No single symbol captured the session’s divergence better than the semiconductor complex. NVDA rose 3.43% to $219.22^[4], building on Tuesday’s momentum after Elon Musk used SpaceX’s first-ever earnings call to declare the company would build its AI infrastructure exclusively on Nvidia’s Vera Rubin architecture^[5]. “We think the Vera Rubin architecture is the best architecture,” Musk told investors, unveiling plans for a joint Starmind AI1 satellite that would perform AI computing in low Earth orbit using Nvidia’s latest chips^[5]. Reports that Nvidia’s B200 systems are sold out added to the bullish read^[6].

AMD told the opposite story. The chipmaker posted Q2 revenue of $11.5 billion and beat on both the top and bottom lines^[7], with data center sales doubling year-over-year^[7]. But its Q3 revenue guidance of approximately $13 billion failed to satisfy investors who had bid the stock up 126% year-to-date^[7]. Shares plunged 7.04% to $482.05^[4], falling as much as 9% in after-hours trading Tuesday before settling into Wednesday’s regular-session loss.

The SpaceX earnings call crystallized a question that has been building for months: in a market where AMD has doubled to become the credible AI alternative, does a single customer’s exclusive commitment to Nvidia reshape the competitive landscape? The stock market’s answer on Wednesday was uncompromising. AMD’s data center business is growing — CEO Lisa Su said she expects server revenue to rise more than 80% in 2027^[7] — but “growing” is not the same as “winning” when the benchmark is Nvidia’s $119 billion in forward supply commitments^[6].

SpaceX itself had a complicated session. The company reported Q2 revenue of $7.8 billion, up 92% year-over-year, with AI revenue surging 247%^[5]. But capital expenditures jumped more than sixfold, and CFO Bret Johnsen disclosed an additional $6.7 billion in contracted cloud services revenue^[5] — a signal of enormous investment ahead. SpaceX shares have shed roughly $1 trillion in value since their June IPO^[5], and the AI spending trajectory is clearly making investors nervous even as the revenue ramp accelerates.

Alphabet’s AI Brain Drain

Alphabet was the day’s largest-cap casualty. GOOGL fell 4.03% to $362.43^[4], with the decline triggered by an announcement that Jeff Dean — one of Google’s earliest employees and the architect of its AI strategy for 15 years — is departing to found Discovery Loop, an AI startup focused on scientific discovery^[8]. Google veterans Oriol Vinyals, Quoc Le, and Sanjay Ghemawat will join him^[8]. Simultaneously, Demis Hassabis, co-founder and head of DeepMind, will step back from day-to-day management to become the lab’s chairman^[8].

The departures land at a sensitive moment. Alphabet has been grappling with investor concerns over AI infrastructure costs since reporting its first-ever negative free cash flow as a public company^[9]. The antitrust overhang, while lighter than feared — Judge Mehta rejected the DOJ’s push for a Chrome divestiture in his September 2025 remedy ruling^[10] — still requires Google to share search data with qualified competitors^[10]. Now the market is asking whether the talent that built Google’s AI moat is leaving at exactly the moment ChatGPT and Perplexity are eroding its search dominance.

Two interpretations deserve weight. The bearish case: Dean’s departure signals that Google’s best AI minds see more upside outside the company than within it, and the cumulative talent drain will narrow the gap between Google and competitors faster than the antitrust ruling ever could. The bullish case: Google remains a $400 billion revenue machine, the antitrust remedy was a “slap on the wrist”^[10], and leadership transitions at DeepMind have historically been followed by stronger execution, not weaker. The 4% decline suggests the market is weighing both stories but tilting toward the first.

Oil Unwinds, Energy Sells Off

The energy sector was Wednesday’s worst-performing major group. XLE fell 2.10%^[1] as crude prices extended Tuesday’s sharp decline. Brent slid 5.3% to close at $79.36 per barrel on Tuesday, with WTI losing 5.7% to settle at $75.77^[11], after Treasury Secretary Scott Bessent said a deal to reopen the Strait of Hormuz could come “today or tomorrow”^[11].

The unwind is significant in context. The U.S.-Iran war had built a substantial geopolitical risk premium into oil prices over the preceding months, and blowout earnings from Exxon Mobil, Chevron, and Valero showed how the conflict generated massive short-term profits for producers^[11]. President Trump’s decision to call off new strikes against Iran^[11], combined with OPEC+’s approval of a September output increase of roughly 188,000 barrels per day^[11], is systematically removing that premium. For energy investors, the question is whether the war-trade gains of the past six months were a durable earnings floor or a one-time windfall that the market is now giving back.

The Defensive Bid

While tech and energy sold off, healthcare caught a clear bid. XLV rose 1.26%^[1], the strongest sector performance of the day. Financials (XLF) gained 0.22%^[1] and consumer discretionary (XLY) added 0.28%^[1], while industrials (XLI) were essentially flat at -0.03%^[1]. Small caps, as measured by IWM, fell 0.64%^[1] — underperforming the large-cap benchmarks, which is consistent with a risk pared-back session rather than a broad risk-off day.

The rotation pattern — Dow up, Nasdaq down, healthcare leading, energy lagging — is the kind of tape that could mean either of two things. Either it is healthy profit-taking after a record run, with capital rotating from crowded AI positions into lagging sectors as earnings broaden. Or it is an early warning that the AI-led rally is losing internal momentum, and the breadth improvement investors have been waiting for is arriving via AI weakness rather than non-AI strength. The distinction matters: the first is constructive, the second is not.

The Macro Backdrop

The macro environment remains in a zone that does not force dramatic choices. The latest FRED snapshot shows unemployment at 4.2%^[3], CPI inflation at 3.46% year-over-year^[3], the fed funds rate at 3.63%^[3], and real GDP growth at 2.1%^[3]. The 10-year Treasury sits at 4.75%^[3], and the 2s10s yield curve is positively sloped at 0.45%^[3] — not inverted, not flat.

Two indicators warrant attention. Consumer sentiment remains depressed at 49.5^[3], down 18.45% year-over-year, even as it bounced 10.49% month-over-month^[3]. And the high-yield credit spread, at 2.84%^[3], ticked up 9 basis points on the month^[3] — a small move, but credit spreads are a leading indicator worth monitoring if the trend continues. The VIX at 15.99^[3] confirms that volatility expectations remain subdued; whatever is happening beneath the surface, the options market is not pricing a break.

The FRED analog search flagged mid-2006 as the most similar macro period^[3] — a time when the Fed had paused rate hikes, inflation was cooling but above target, and the economy was growing modestly. That period preceded the 2007-2009 unwind, but it also featured another 12 months of equity gains before the turn. Historical analogs are context, not forecasts.

Earnings Season Scorecard

Wednesday’s session capped a dense earnings window. The key results:

The pattern is clear: companies that beat and raised (Palantir) were rewarded; companies that beat but couldn’t exceed elevated expectations (AMD) were punished; and companies facing structural concerns beyond a single quarter (Alphabet) were sold on narrative, not just numbers.

What to Watch Next

• Hormuz timeline: Bessent said a deal could come “today or tomorrow” — that was Tuesday. If an agreement materializes, expect energy to extend losses and risk assets to catch a further bid. If talks stall, the unwind reverses quickly.
• AMD after-hours drift: AMD closed the regular session at $482.05^[4] but was trading at $479.24 in after-hours^[4], down another 0.58%. Whether the stock stabilizes or continues lower will signal whether the guidance disappointment is being digested or deepening.
• Alphabet’s next AI move: With Dean, Vinyals, Le, and Ghemawat confirmed for Discovery Loop, the market will watch who replaces them and whether Google’s AI roadmap slips. Koray Kavukcuoglu, now reporting directly to Pichai^[8], is the key name to track.
• SanDisk (SNDK) earnings: Scheduled for after the close on August 5^[8], SanDisk’s results will offer another data point on whether the AI-driven memory and storage cycle is accelerating or plateauing.
• Credit spreads: The 9-basis-point monthly uptick in HY spreads^[3] is small but directional. A continued rise alongside equity weakness would mark the kind of quiet indicator that precedes broader risk repricing.
• Consumer sentiment: At 49.5^[3], sentiment remains near historic lows despite the monthly bounce. The next University of Michigan preliminary reading will test whether the recovery has legs or is a dead-cat bounce.

The base case remains that this is a healthy rotation within an earnings-supported bull market — GDP is growing 2.1%^[3], the Fed is on hold at 3.63%^[3], and earnings are broadening. But the AMD and Alphabet sell-offs are reminders that the AI trade, which carried the market to records, is now carrying the most concentration risk. The rotation beneath the flat index level is where the real information lives.

FN2 Research provides market commentary and educational analysis, not personalized investment advice. All prices reflect market data as of the cited timestamps.