CNBC Fast Money
2026-07-13 · Hosted by Melissa Lee · CNBC
Executive Summary
Fast Money opened with breaking news of Apple's federal lawsuit against OpenAI over alleged trade secret theft tied to consumer hardware, naming OpenAI's chief hardware officer Tang Tan. The desk then dug into SK Hynix's record NASDAQ debut — priced at $149, closing near $168, a roughly 16% premium to its Seoul shares — with traders split on whether to chase the "IPO euphoria" (Bonawyn Eison skeptical, historical IPO drawdowns often exceed 50%) versus viewing it as validation of a durable memory supercycle (Dan Ives, bullish, calling the AI buildout comparable to the 1870s-1900s railroads). Apple's resilience amid AI-spending scrutiny, Oracle's balance-sheet concerns, Meta's best week since February 2024, and previews of bank earnings (Tuesday) and Netflix earnings (next week) rounded out the show, alongside a debate on oil's outlook given Iran tensions.
Key Stories & Changes
1. Apple Sues OpenAI Over Trade Secret Theft
Apple's federal filing (41 pages) accuses OpenAI of "stealing Apple's trade secrets and confidential information" via former employees, including Chief Hardware Officer Tang Tan
Alleges Tan used job interviews with Apple staff to extract details on unreleased products and coached departing employees on evading Apple's security procedures
Names a former Apple product developer who allegedly took a stolen laptop to OpenAI and used it for months to download confidential R&D documents, including a proprietary metal-finishing technique
Seeks an injunction plus damages; Apple reportedly appealed directly to OpenAI in February before filing suit
Traders (Guy Adami, Bonawyn Eison) said the news is unlikely to move Apple's stock materially but called it "interesting" given Tan's prior Apple role — some framed it as "sour grapes"
2. SK Hynix Debuts With a Record ADR Offering
SKHY (temp ticker): SK Hynix — Priced $149, opened $170, closed ~$168 — Largest foreign-company ADR/IPO in US history after SpaceX; ~16% premium to Seoul shares
MU: Micron — Volatile, down over prior weeks — Seen as funding source for SK Hynix; some argue its selloff already priced in this event
NVDA: Nvidia — Steady — SK Hynix is Nvidia's largest HBM supplier
Bank of America estimates 35-40% of a projected $1.5 trillion 2027 big-tech AI infrastructure spend flows directly to memory
Digitimes reports HBM prices could more than double in 2027
SK Group Chairman Che told CNBC he's doubling capacity within five years, but customers still say "that's not enough, man, we need more"
At least half a dozen leveraged single-stock products tied to SK Hynix launch the following Tuesday, rebalancing daily and potentially amplifying swings
Guy Adami: historically, IPO drawdowns from highs often exceed 50%, favoring a wait-and-see approach; Dan Ives called it a "watershed day" for Korean tech, seeing the AI cycle at "third inning, one out"
3. Apple's Quiet Outperformance
Apple trading near an all-time high, outperforming the broader "Mag 7" over the past several weeks
Traders (Guy Adami, Bonawyn Eison, Mike Khouw) attributed this to Apple functioning as a free cash flow hedge (~$126 billion trailing 12-month FCF) against hyperscalers burning cash on AI capex
Zidle (guest) noted Apple "stumbled into" AI relevance via its unified-memory Mac architecture, which is efficient for running local open-source models
Mike Khouw: nearly 1 million call contracts traded in Apple that day; sees 13-15% top-line growth ahead — "more of a growth story than some people imagine"
4. Oracle and the AI CapEx "Pound-a-Box"
Dan Ives (Wedbush) said Oracle needs to show RPO-to-revenue conversion beyond OpenAI concentration to regain investor confidence; debt levels remain a small share of its capital structure but require continued raises ($45-50 billion planned)
Desk debated buying Oracle: Guy Adami argued for a small speculative ("moonshot") position, partly betting on White House rhetoric boosting the stock as it has for Intel, Dell, and Micron
Joe Zidle called the broader AI buildout comparable in scale to the railroads of the 1870s-1900s and telecom of the late 1990s, and said CapEx has already exceeded that late-90s peak
5. Delta Earnings and the K-Shaped Economy
Delta beat estimates but shares dipped; CEO Ed Bastian said the airline expects to recover 100% of a $4 billion full-year fuel cost increase and reaffirmed full-year profit guidance
Joe Zidle: top 10% of US household income drives the majority of consumer spending, and continues to spend regardless of high fares, reinforcing a "K-shaped economy" read
Desk noted full flights and high fares signal resilient premium-consumer demand despite inflation concerns
6. Meta's Best Week Since February 2024
Meta up ~15% since Monday on AI headlines including a coding-model update and a custom-semiconductor partnership with Broadcom
Bonawyn Eison: Meta has "added optionality" by monetizing excess compute, reversing prior scrutiny over unprofitable CapEx, though this implies margin step-down from ~70% ad-business margins to ~mid-30s cloud margins
Guy Adami and Mike Khouw both bullish, citing valuation (~19x forward earnings) below peers and a technical breakout above the recent downtrend
Trends Identified
1. IPO Euphoria vs. Structural Memory Supercycle — A Genuine Debate, Not Consensus
The desk split clearly: Bonawyn Eison called SK Hynix part of "peak euphoria" with historical IPO drawdown risk, while Dan Ives and Joe Zidle framed the AI memory buildout as a multi-decade structural shift comparable to past industrial revolutions. This is a meaningful, explicit disagreement rather than a uniform bullish read, and signals investors should distinguish near-term IPO mechanics from the multi-year memory demand thesis.
2. Compute Monetization Is Becoming the New Bull Case for Mega-Caps
Both Meta (selling excess compute) and Apple (unified memory advantage) are being re-rated based on how they monetize or avoid AI infrastructure spend rather than raw AI product news — suggesting investors are shifting focus from "who's building the most" to "who profits most efficiently from what's built."
3. AI Infrastructure Spending Is Reframing as Inflationary, Not Just Cyclical
Joe Zidle's point that HBM price doubling forecasts "put the debate [on whether AI is inflationary] to bed," combined with the observation that $1.5 trillion in 2027 CapEx is commodity- and labor-intensive, signals the AI buildout is increasingly viewed as a macro inflation input, not merely a tech-sector cyclical story. ---
Sentiment Analysis
Overall Market Sentiment: Bullish but Selectively Skeptical
The desk remained constructive on AI-linked mega-caps and memory demand, but showed explicit skepticism toward IPO-specific froth and highly leveraged names like Oracle.
Risk Factors Highlighted
Historical IPO drawdown risk: Guy Adami notes newly listed IPOs often see 50%+ drawdowns from initial highs, a pattern investors should expect from SK Hynix.
Memory funding rotation, not net-new demand: Debate over whether SK Hynix inflows are cannibalizing Micron and other chip-sector capital.
Oracle balance sheet and debt reliance: Continued dependence on debt issuance ($45-50B planned) to fund AI infrastructure buildout.
Apple-OpenAI legal exposure: Could constrain OpenAI's hardware ambitions and add reputational risk ahead of its IPO.
Iran instability and oil price risk: Brian Sullivan flags a possible power vacuum/civil war risk in Iran that could push oil toward $100/barrel if the National Iranian Oil Company strikes.
AI capex "commoditization" concern: Some worry too many companies believe they'll capture outsized share of the AI TAM, implying an eventual shakeout of weaker-balance-sheet players.
Meta's margin dilution from cloud monetization: Shifting excess compute to a cloud-like business could compress overall margins from ~70% (ads) toward the mid-30s.
This episode was covered in today's [The Market Signal — 2026-07-13](https://marketsignal.beehiiv.com/p/the-market-signal-2026-07-13), a cross-source synthesis of multiple podcast reports.