CNBC The Exchange

2026-05-28 · Hosted by Kelly Evans · CNBC

Executive Summary

The Exchange explored the new trillion-dollar memory club (Micron, SK Hynix) through two opposing lenses: Gene Munster (Deepwater Asset Management) warning of diminishing returns from massive earnings growth already priced in, and Vivek Arya (Bank of America) arguing AI demand is growing 4–5x faster than supply can accommodate, justifying the valuations. The episode also covered the mega-IPO pipeline (SpaceX targeting $80 billion), cybersecurity as the “new semi” trade despite Z-Scaler’s 30%+ collapse, Fed policy uncertainty with rate cuts still potentially on the table, and Robinhood’s AI agent trading launch. Small caps hit a record with the Russell 2000 on pace for its best quarter in 6 years, even as options traders built unusual put protection.

Key Stories & Changes

1. Trillion-Dollar Memory Club — Bulls vs. Bears

  • Broadly bullish long-term; concern is about diminishing returns from already-exceptional growth rates

  • Micron revenue growth: +57% (Nov Q), +197% (Feb Q), +260% est. May Q, potentially +360% per Counterpoint Research

  • NVIDIA parallel: Calendar 2027 earnings estimates went from 10% growth → 42% growth in 6 months; stock only up ~20% in that period — “diminishing return” dynamic

  • “A year ago I thought we were in the third inning. Now I think we’re in the second inning” — AI transformation still massively underappreciated

  • Concern: IPO excitement around SpaceX/OpenAI/Anthropic may be funded by selling existing winners like Micron/SK Hynix

  • Verdict: “We’re more in the fine camp” — stocks will be okay but better places to be

  • AI creating demand growing at 4–5x the speed the industry can provide supply — “never seen this before”

  • HBM versions progressing: V3 → V4 → V5; each version requires 3–4–5x more wafers for same capacity — structural supply constraint

  • DRAM pricing: Up 7–8x vs. last year; up 60–70% sequentially in Q2; expects another 5–10% Q3; some deceleration starting early next year

  • Memory pricing may be locked in via long-term supply agreements — higher multiple justification

  • SOX index trading at 25–26x forward — essentially same as start of year despite 70%+ move, meaning ALL the gain is earnings growth, zero multiple expansion

  • AI demand broadening: compute → CPUs → memory → optics → EDA (Synopsys, Cadence)

2. Mega-IPO Pipeline — Is There a Liquidity Crunch?

  • SpaceX reportedly targeting $80 billion for its IPO next month

  • Dan Primack (Axios): Not worried about liquidity crunch — 2021 set the all-time US IPO proceeds record; even if SpaceX/OpenAI/Anthropic each raise $80B, as a percent of total market cap (much larger than 2021), it’s not more than 2021

  • “Capital is not finite… if people want to get in, they’ll find a way”

  • BUT: Index inclusion changes being made NOW (not a year ago) specifically to accommodate these companies — Primack called this “what happened here” without naming names

  • FOMO + index-driven pressure = institutional investors feeling they “need to be in all three”

  • Primack also flagged quantum computing startup (Quantinuum?) targeting $1 billion IPO next week — received $100M from US government

3. Cybersecurity — Z-Scaler Collapse vs. “New Semi” Thesis

  • Z-Scaler down 30%+: Revenue guide weaker than expected; blaming soaring memory costs and departure of key sales executives

  • Sema Modi: CFO Kevin Rubin cited sales leadership disruptions; Wall Street not buying the rationale

  • UBS cut target from $260 → $225; WEDD cited need for better execution

  • Victoria Green (G-Squared Private Wealth): Called Z-Scaler “one of the weaker picks” — hold, not sell on a down 30% day

  • Prefers CrowdStrike and Palo Alto — both use Anthropic-powered AI for enhanced offerings; trusted by enterprises (Palo Alto: 85% of Fortune 500/100)

  • Bug ETF (broad cyber): Down 5% today but up 37% in past 2 months — the “new semi” comparison

  • Next week: Palo Alto and CrowdStrike earnings — expected to “beat and raise”; CrowdStrike’s net new ARR is the key metric to watch

4. Trillion-Dollar Club — Who’s Next?

  • GEV (GE Vernova): Powering data centers; nuclear, transmission, renewables

  • Lam Research: Making machines that build semiconductors

  • Sandisk: Memory storage for AI

  • IBM: Underrated quantum computing story

  • All at ~$250 billion market cap currently; would require major gains but trajectory is plausible given AI tailwinds

  • Note: These are longer-term (years, not 6 months) candidates

5. Russell 2000 Record — Options Market Turning Cautious

  • Russell 2000 at all-time high; on pace for 9th positive week in 10; 20% gain in that period; best quarter in 6 years

  • Options market: IWM has 71% put premium (vs. 38% for S&P, 57% for NASDAQ) — strikingly elevated

  • Volume: nearly 3x more puts than calls traded; 100,000+ more puts bought than sold

  • One trader put $8 million into July 17 put spreads betting IWM could drop 7% by mid-July

  • Oliver Rennick (SIBO): Put skew in small caps may reflect interest rate sensitivity; bonds on 60-basis-point bounce so not outwardly bearish

  • Message: If market needs to cool, small caps are most at risk

6. Fed Policy — Rate Cut Still Possible?

  • Rate hike odds rising: ~60% probability for March next year; higher by next summer

  • Carol Schleif (BMO Wealth Management): Does NOT expect rate hike this year; rate cuts still possible

  • Key argument: Inflation is “demand-pull” from AI buildout — companies paying up for chips and talent — not a traditional wage-price spiral that forces Fed response

  • Fed has new leadership (Warsh) and must balance labor market, growth limits, and inflation — more factors than usual

  • Consumer paradox: Terrible sentiment readings but consumer stocks leading market today (spending behavior doesn’t match sentiment)

  • Schleif’s positioning: Overweight tech, communication services, industrials, financials; market needs to broaden for momentum to continue

7. Robinhood AI Agent Trading

  • Same story covered across all CNBC shows: Robinhood first retail broker to let clients connect external AI agents (ChatGPT, Claude) to execute trades

  • Kelly Evans’s take: Liability question central; “at least I can sue a mutual fund if they make a mistake”

  • Potential for “flash crash” risk if all AI agents make the same stock picks simultaneously

  • Herding effect: AI strategies that converge could create amplified market reactions that hedge funds capitalize on

1. AI Earnings Acceleration Creating “Diminishing Return” Paradox

The NVIDIA parallel Gene Munster drew is compelling: earnings estimates for calendar 2027 rose from 10% to 42% growth over 6 months, yet the stock only gained 20% in that period. Applied to Micron, the law of large numbers becomes the enemy — exceptional growth rates get harder to exceed off higher bases. This dynamic suggests that even if the AI thesis is correct, the biggest gains in memory and chip stocks may have already occurred, and future returns will require identifying the next layer of beneficiaries (EDA, interconnects, power infrastructure).

2. Cyber Bifurcation Accelerates — Platform vs. Point Solution Divergence

Z-Scaler’s 30%+ collapse on execution issues, set against the broad cyber ETF’s 37% two-month run, crystallizes a sector dynamic: the market is aggressively pricing the winners (Palo Alto, CrowdStrike) and punishing the laggards (Z-Scaler, point-solution vendors). The platform consolidation thesis — fewer, broader vendors managing more of the enterprise security stack — is being validated in real time by spending patterns and share dynamics. Palo Alto’s 85% Fortune 500 penetration is the moat investors are paying for.

3. Options Market Divergence as Early Warning Signal

The extreme put skew in IWM (71% put premium vs. 38% for SPX, 57% for QQQ) is a notable signal that sophisticated options traders are positioning for small-cap underperformance even as the index hits all-time highs. This divergence between price action and hedging behavior has historically been a leading indicator of rotational stress. The concern is interest rate sensitivity — small caps are disproportionately affected by higher-for-longer rates, and with PCE data due Friday, the risk window is near.

4. IPO Superycle as Market Stress Test

The combination of SpaceX ($80B), OpenAI, and Anthropic preparing for public offerings represents a capital absorption challenge not seen since 2021’s record year. But unlike 2021, this supply is thematically concentrated in AI — meaning it competes directly for the same pools of capital that are already allocated to NVIDIA, Micron, and the hyperscalers. Dan Primack’s dismissal of liquidity concerns may be correct in aggregate, but the concentration risk is real for AI-specific portfolio managers. —-

Sentiment Analysis

Overall Market Sentiment: Cautiously Optimistic with Pockets of Concern

Market hitting new records but sophisticated money (options traders, macro strategists) building protection; AI thesis debated but still broadly supported.

Risk Factors Highlighted

Memory earnings deceleration: DRAM pricing sequential growth expected to slow starting early next year as new capacity comes online; sequential growth is what drives stock performance

Z-Scaler contagion to broader software/cyber: Even if execution-specific, sudden guidance collapse raises investor caution about all software narratives

Russell 2000 put activity: 71% put premium, $8M put spread betting 7% decline — sophisticated money hedging small-cap record highs ahead of inflation data

AI mega-IPO capital concentration: SpaceX ($80B) + OpenAI + Anthropic all drawing from same AI thematic capital pool; concentrated supply risk

Rate hike risk returning: 60% probability by March next year; if Fed is forced to hike, small caps and rate-sensitive names most vulnerable

Palo Alto/CrowdStrike earnings risk: Positioned as “beat and raise” certainties — high bar creates downside if execution stumbles

Quantum computing as early-stage speculative surge: Quantinuum targeting $1B IPO after $100M government grant — early warning of hype cycle reaching into less-proven technologies

Consumer-sentiment divergence: Terrible sentiment readings vs. strong actual spending; if sentiment ultimately wins (i.e., consumers stop spending), cyclical thesis breaks

This episode was covered in today’s The Market Signal — 2026-05-28, a cross-source synthesis of multiple podcast reports.

Keep Reading