Nasdaq Leads S&P 500 to Record as AI Infrastructure Demand Overwhelms Flat Inflation

The Nasdaq is one of the cleanest tells in the August 13 closing snapshot: tech leadership pushing the S&P 500 to its first close above 7,800 while the Dow industrials barely moved. The QQQ gained 1.16% to 732.07, and the SPY rose 0.70% to 777.88 — a record close of 7,798.99 on the S&P 500 itself^[1][2]. The DIA added just 0.14%^[1]. This is not a broad-rally tape. It is a concentration tape, and the concentration is in AI infrastructure.

The Inflation Backdrop: Tame Enough to Hold

July CPI came in at 3.46% year-over-year, in line with expectations, with a key core metric matching its slowest pace since March 2021^[2][3]. Consumer prices barely rose as gasoline declined for a second straight month^[4]. The data reinforced market expectations that the Federal Reserve will hold the policy rate steady at its September meeting, with the fed funds rate currently at 3.63%^[5].

The macro dashboard is consistent with a soft-landing mosaic: unemployment at 4.1%^[5], real GDP growing at 2.1% year-over-year^[5], and the 10-2 yield curve positively sloped at +0.48%^[5] — no inversion, no recession signal. The VIX sat at 15.15, roughly 9.7% below its year-ago level^[5], and high-yield credit spreads tightened to 2.72%^[5]. Both reflect a market pricing very little near-term distress.

The one indicator that refuses to cooperate is consumer sentiment, at 49.5 — still down 18.5% year-over-year, though it did bounce 10.5% month-over-month^[5]. The dissonance between a sentiment series near historical lows and equity markets at records is worth holding in mind, even if it has not yet translated into spending weakness in the GDP data.

CoreWeave: The Infrastructure Demand Signal

The single strongest catalyst this week came from CoreWeave (CRWV), the AI cloud provider that went public in March 2025. On August 11, CoreWeave reported Q2 revenue that more than doubled year-over-year, a revenue backlog that swelled to $104 billion, and raised its 2026 revenue forecast above consensus^[6]. CEO Michael Intrator described capacity as “sold out,” citing unprecedented demand for CoreWeave’s GPU cloud^[6]. The company announced new business with both Anthropic and Meta in the quarter^[6]. Shares jumped roughly 14% in extended trading after the report^[6].

This matters beyond CoreWeave’s own market cap. A $104 billion backlog is a leading indicator of hyperscaler and enterprise commitment to AI compute spending over multi-year horizons. If the backlog converts at projected margins, it argues that the AI infrastructure capex cycle is not a speculative build-out — it is pre-contracted demand. That reading supports the premium the market is willing to pay for chips, networking, and data-center real estate.

Meta’s Compute Pivot: From Cost Center to Revenue Line

Meta (META) was the strongest mega-cap on the session, closing up 2.78% at $594.97^[7]. The move extends a run that began the prior week, when Meta laid out its “Meta Compute” strategy to monetize surplus AI infrastructure by selling compute capacity to third parties^[4]. Mark Zuckerberg confirmed publicly that selling compute “makes sense” and disclosed that buyers had offered Meta a “significant premium” for its compute capacity — which he turned down, calling selling at this stage “foolish”^[4].

Wolfe Research estimated that each gigawatt of compute monetized at a ~$25 billion annual rate could lift Meta’s EPS by roughly 20%^[4]. That is a large number, and it reframes Meta’s AI spending from a pure cost burden — the narrative that punished the stock for much of the past year — into a potential high-margin revenue stream. The strategy would put Meta in direct competition with AWS, Azure, and Google Cloud, and it poses a competitive threat to neoclouds like CoreWeave, Nebius, and IREN, which fell on the Meta Compute news^[4].

The tension is real: CoreWeave’s blowout results validate the demand for AI compute, but Meta’s entry validates the supply-side concern that hyperscaler compute could eventually commoditize the neocloud margin profile. Both things can be true simultaneously — demand is enormous, and the players capturing it may multiply.

Tesla and the Autonomy Re-rating

Tesla (TSLA) closed up 3.80% at $339.96^[7], breaking above its 50-day moving average and drawing heavy call-options volume^[8]. The catalyst is not a single news event but a confluence: Cathie Wood’s ARK Invest published work arguing that nearly all of Tesla’s future value sits in robotaxis rather than vehicle sales^[8], and Dan Ives of Wedbush assigned a $600 price target by valuing Tesla as an AI-robotics company rather than an automaker^[8].

The bullish thesis depends on a 55% full-self-driving attach rate on North American deliveries translating into a recurring software revenue stream^[8]. Against that stands the reality of negative free cash flow and a Q2 EPS miss^[8]. The stock is being re-rated on the option value of autonomy, not on current earnings — a bet that is either visionary or premature depending on whether robotaxi deployment timelines compress or slip.

Workday and the Take-Private Signal

Workday (WDAY) soared approximately 18% — its biggest one-day gain in over a decade — after Reuters reported that Silver Lake is in talks to acquire the enterprise software company^[3]. The stock was the single best performer in the S&P 500 on the day^[3].

A take-private of this size signals that private equity sees value in public enterprise-software names that the public market has underpriced. If a deal materializes, it could be a catalyst for further M&A interest in the large-cap software space — a sector that has lagged the AI-themed rally despite being a primary distribution channel for AI features into enterprises.

The Narrowness Question

The dispersion is the story. The Nasdaq 100 outperformed the Dow by roughly 100 basis points^[1]. Health care was effectively flat, energy was dead weight, and small caps barely participated. The rally is being carried by a handful of AI-adjacent mega-caps — META (+2.78%), TSLA (+3.80%), AAPL (+1.00%), MSFT (+0.90%)^[7] — while AMZN (-0.80%) went the other way^[7].

A market where the S&P 500 makes record highs on the strength of five names while two-thirds of the index’s sector ETFs gain less than 60 basis points is a market pricing a specific thesis, not a broad recovery. The thesis is that AI infrastructure spending is the dominant growth vector in the economy, and that tame inflation gives the Fed room to stay accommodative enough to fund it.

What to Watch Next

• Fed speakers and the September meeting: With CPI at 3.46% and the funds rate at 3.63%^[5], real rates are slightly positive. Any hawkish surprise from Fed speakers next week could challenge the low-VIX, tight-spread backdrop that is enabling record equity prices.
• PPI and retail sales (August 14–15): Wholesale inflation came in lower than expected^[2], adding to the disinflation narrative. Retail sales will test whether the weak consumer sentiment reading (49.5)^[5] is showing up in actual spending data.
• Nvidia earnings (late August): As the sole GPU supplier powering CoreWeave’s sold-out capacity and Meta’s compute build-out, Nvidia’s results will be the most direct test of whether the AI infrastructure demand signal is translating into revenue at the expected scale.
• Meta Compute details: Watch for any further specifics on pricing, gigawatt capacity available for external sale, and timeline. The gap between Meta’s internal compute needs and what it could sell externally is the variable that determines whether this becomes a $25-billion-per-gigawatt business or a marginal side project.
• Tesla robotaxi milestones: Any concrete deployment timeline — a city, a date, a regulatory approval — would crystallize the autonomy re-rating thesis. Slippage would test whether the current 3.8% rally is built on durable conviction or options-driven momentum.
• Workday/Silver Lake: Deal terms, financing structure, and whether competing bids emerge will calibrate the private-equity appetite for large-cap software, a sector that has been quiet on the M&A front.

The base case implied by today’s tape is continuation: tame inflation, Fed on hold, AI infrastructure demand validated by contracted backlog, and mega-cap concentration driving index-level records. The risk case is that concentration cuts both ways — if the AI capex thesis disappoints on any single data point, the narrow leadership that produced the record highs becomes the mechanism for a sharp reversal. The VIX at 15 says the market is pricing very little of that risk. Whether that complacency is warranted depends almost entirely on what Nvidia reports later this month.