Market Brief(X) — May 19–May 22, 2026

2026-05-23 • Twitter

Executive Summary

The AI infrastructure capex super-cycle remains the dominant narrative, with the rally this week broadening decisively from core semiconductors into CPUs, power, solar, and even legacy OEMs like Dell. The signal is no longer just about GPU scarcity—it has migrated up the stack to CPU orchestration for Agentic AI and down to the physical bottlenecks in power delivery (800V HVDC), packaging substrates, and optical interconnects. However, the week closed with a sharp geopolitical reminder: a potential US-Iran ceasefire framework and a severe Chinese regulatory crackdown on cross-border brokerages injected tail-risk uncertainty into an otherwise euphoric tape. The central tension is whether the broadening infrastructure trade can absorb the macro and geopolitical shocks that are rapidly multiplying, or if the dispersion itself signals a late-cycle chase for marginal beta.

Key Themes & Trends

The Agentic AI Bottleneck Has Shifted from GPU to CPU

The most consequential thematic pivot this week is the market’s recognition that Agentic AI workloads invert the data center compute hierarchy. Multiple commentators flagged that the CPU-to-GPU ratio is moving from 1:4–1:8 toward 1:1 or even higher, as orchestration, tool-calling, and serial logic become the binding constraint. @qinbafrank provided the foundational thesis mid-week, noting that Agentic workflows can push CPU dynamic energy consumption to 44% of total, a 3-4x increase over traditional AI. This was validated by @jukan05, who reported that AMD is accelerating its Helios rack-scale platform deployment with Taiwanese ODMs ramping production in Q3–Q4. @fi56622380 added that Qualcomm’s recent ASIC order book from ByteDance further confirms the CPU demand surge. The convergence here is exceptionally strong: macro commentators (@qinbafrank, @FundaAI), industry analysts (@jukan05, @fin), and traders (@tradergokux on $AMD, $INTC, $ARM) are all aligned on the structural nature of this shift. The key watchpoint: if CPU TAM estimates continue being revised upward—$200B by 2030 per Nvidia, per @fi56622380—the DRAM demand implied by higher CPU core counts and per-core memory (17GB per Vera CPU core vs. 6.7GB for Grace) has not yet been fully priced into memory stocks.

800V HVDC: The Next Infrastructure Layer Is Power Delivery

A deep, technically rich theme emerged mid-week around NVIDIA’s push for 800V High-Voltage Direct Current architecture in next-generation AI racks. @qinbafrank published a comprehensive analysis detailing how the shift from 54V to 800V is not an incremental upgrade but a platform-level rearchitecture of data center power delivery, driven by the physics of 1MW+ racks. The ecosystem beneficiaries span SiC/GaN power semiconductors (Navitas, Infineon, Innoscience), power system integrators (Delta, Vertiv, Schneider), and component suppliers. @qinbafrank explicitly flagged NVTS as a high-beta GaN play on this theme, noting its inclusion in NVIDIA’s 800V reference design. The theme’s credibility is reinforced by the quantifiable BOM analysis from Morgan Stanley on Vera Rubin racks: power value share is up 32% and cooling up 12%, per @qinbafrank. This is a multi-year capex theme with clear supplier mapping—early-stage but investible.

Optical Interconnects: The “Physical Bottleneck” That Caps AI Scaling

Herman Jin (@ShanghaoJin) delivered the week’s most forceful articulation of a contrarian supply-chain constraint: optical communications as the ultimate physical bottleneck for AI scaling. In a series of posts, he argued that “what is truly constraining token compute is optical interconnects” and that the entire AI dream is “being held up by small companies and small bottlenecks one by one” (@ShanghaoJin). His hierarchy: single-card performance depends on GPU architecture + HBM; rack-level on NVLink and thermal; 10,000-card clusters on optical interconnects, power conversion, and switch fabric. This view was amplified by @jukan05 reporting that Korean memory makers are actively discussing decoupling HBM from GPU packages using optical interconnects to overcome the “memory wall” and GPU shoreline constraints. @LinQingV added that if GPU-HBM optical interconnects materialize, the demand volume for CPO components would shift from network-equipment scale to chip scale, creating a TAM expansion for precision optical alignment equipment (ficonTEC/罗博特科). @zephyr_z9 tempered enthusiasm by noting this is a “2029 or beyond” story, with optical interconnects reaching switches and XPUs first. The profile mix here—industry analysts (Jukan, Zephyr, LinQingV) and a VIP macro commentator (ShanghaoJin)—gives this theme analytical weight even though the timeline is long.

Memory Super-Cycle: Undersupply Thesis Intact, But Japanese Pricing Philosophy Disappoints

The memory supply-demand outlook tightened further. @zephyr_z9 cited UBS research forecasting DRAM undersupply until Q2’28 and NAND until Q4’27, with CoWoS capacity reaching 210k WPM by end-2027—a cycle longer and tighter than historical norms. However, a revealing divergence opened between Korean and Japanese supplier behavior. @jukan05 documented that Nittobo, despite commanding critical T-glass supply for IC substrates, has explicitly stated it will not raise prices, prioritizing market share and customer relationships. @LinQingV sharply observed that Japanese materials companies with monopoly positions but unwillingness to exercise pricing power are “not shareholder-friendly” and unlikely to generate alpha. This is a nuanced but critical insight: the memory shortage thesis is real, but not all shortage beneficiaries will monetize equally. Korean memory makers (Samsung, SK Hynix) and Micron are the “pricing power” plays; Japanese materials companies are the “volume without price” disappointment. Late Friday, @Balder13946731 flagged an emerging risk: CXMT DDR5 appearing in Corsair consumer products, a potential signal that Chinese memory is beginning to penetrate global supply chains outside HBM.

US-Iran: “Deal Framework” Narrative Collides with Weekend Escalation Risk

The week closed with a geopolitical whipsaw. Early confidence in a negotiated settlement—per @qinbafrank citing Pakistani mediation progress and a potential 60-day ceasefire extension with gradual Strait of Hormuz reopening—lifted risk assets Friday. But by Saturday, reports emerged that Trump canceled weekend plans to remain at the White House, and US military/intelligence officials were placed on standby (@Balder13946731). @AntonLaVay offered the most analytically sharp take: Trump’s priorities are oil prices and bond yields, and continuing strikes on Iran would worsen both—so escalation is against US interests, but Iranian retaliatory credibility (they said they’d block the Strait and did) means de-escalation cannot be assumed. This is the defining macro uncertainty entering the holiday-shortened week.

China Regulatory Crackdown: A Policy Shock with Portfolio Implications

On Friday, China’s SEC announced severe penalties on Futu, Tiger Brokers, and Long Bridge for illegal cross-border securities operations, including confiscation of all illegal proceeds and a two-year mandatory liquidation of existing mainland client accounts (@qinbafrank). TIGR dropped 40%+ pre-market. The implications extend beyond the broker stocks themselves. @NullableX provided the most consequential analytical framing: this is not merely a regulatory action but a signal of how Beijing intends to manage capital flows in an era of fiscal pressure. The state needs domestic capital directed toward domestic industrial priorities—semiconductors, AI, advanced manufacturing—and views outbound portfolio investment as “disobedient capital.” @Franktradinglog and @RichTerry123 reinforced the practical implication: Chinese retail flows into US equities, which have been a marginal bid for years, will be structurally reduced over the next two years. The policy risk premium for Chinese ADRs and any stock dependent on mainland flows has been repriced higher.

Solar & US Domestic Energy Supply Chain: A Sub-Radar Convergence

A concentrated but high-conviction theme around US solar manufacturing emerged from multiple high-weight voices. @NullOreo_ initiated on $FSLR mid-week, citing its position as a US-domiciled, financially sound leader at a time when China is blocking solar manufacturing equipment exports to Tesla. He then expanded the thesis to $GLW (Corning), which is expanding Michigan solar capacity and sits at the intersection of AI optical interconnects and US solar supply chains (@NullOreo_). @AntonLaVay concurred on $FSLR as a “stable, reliable” beta play. @labubu_trader added that Chinese PCB strength does not negate TTMI’s domestic advantage, reinforcing the “domestic supply chain security” narrative. This theme draws on a trader (NullOreo), a macro-investor (AntonLaVay), and a short-term trader (Labubu) arriving at the same conclusion through different lenses—a robust convergence signal.

Market Sentiment

Short-term (next 1-2 weeks)

Extremely bullish with tactical caution around the long weekend. The tape on Friday showed breadth expansion: Dell, HP Enterprise, and legacy computer OEMs surged double-digits (@Balder13946731), software rallied on $INTU’s earnings (@TJ_Research), and QCOM broke out with +13% in two sessions (@tradergokux). @Franktradinglog described it as “a very comfortable week,” rotating from software hedges into semis and solar. @TraderGoku summarized the operating principle: “Know when to press the gas.” The AAII survey showed only 31.7% bullish, which @labubu_trader flagged as a contrarian bullish signal. However, @mat78704 and @Corsica267 both warned of a potential liquidity-driven dip around May 27–29 before a rally into early June. @Corsica267 specifically flagged a Liquidity Grab signal for the NQ.

Long-term (weeks to months)

Structurally bullish with an emerging divergence. The dominant view, articulated by @ZaStocks and @TradeBrigadeCo, is that AI represents a transformational technology cycle that justifies sustained long exposure, with pullbacks as buying opportunities. @AntonLaVay contributed a framework-shifting analysis: AI infrastructure profit-pool companies are not “long-duration growth” vulnerable to higher rates, but “short-duration growth + industrial bottlenecks + pricing power + current-period earnings revision.” This reframing supports the bull case even in a higher-for-longer rate environment. However, @TJ_Research provided a crucial data-driven caution: his revised analysis of NDX forward P/E shows 24.3x, with 27.5x as the historical upper bound, implying ~13% upside to the zone where he would “begin reducing beta”—a narrowing runway. @ShanghaoJin captured the cognitive dissonance: “Macro is dusk, micro is dawn—I’m squeezed in the middle like a schizophrenic.”

Sentiment Balance & Shift

The week began with cautious optimism after NVDA earnings, shifted to aggressive risk-on by Thursday as CPU and power infrastructure themes caught bid, and closed Friday with a sharp edge of geopolitical anxiety. The most notable shift: the “sell the news” response to NVDA earnings (stock declined -1.77%) was absorbed and reversed within two sessions as the narrative expanded beyond NVDA to the broader infrastructure ecosystem. This is a healthy dispersion signal—the rally is no longer solely dependent on NVDA’s market cap. However, @NullableX noted a potential warning: DRAM stocks were flat-to-down on a day SOXX surged, suggesting the memory trade may be losing momentum even as the broader semi complex accelerates. The trader cohort remains uniformly bullish; the macro cohort is constructive but increasingly watchful of valuations and geopolitical tail risk.

Key Figures & Assets

Short-term / Technical Trades

  • $AMD: Multiple traders flag a High Tight Flag breakout to all-time highs. @tradergokux notes the pattern has completed with <20% max drawdown and tight Doji consolidation on the 5-day MA. Catalysts: Middle East ceasefire would accelerate upside. Risk: a false breakout if the HTF pattern fails at the 122.8 level, per @tradergokux.
  • $ARM: @tradergokux describes it as the technically strongest semiconductor, breaking a 2-year consolidation with weekly volume +50% on the third day post-breakout. He added back his full position on the 229 breakout. @qinbafrank provides the fundamental overlay: ARM is the most undervalued of the CPU “Big Three” (INTC, AMD, ARM) given its architectural advantage in Agentic AI workloads. Convergence of a top trader and a VIP commentator on the same name is a strong signal.
  • $QCOM: @tradergokux flagged a +13% move in two sessions. @FundaAI and @jukan05 confirm fundamentals: QCOM is providing ASIC services to ByteDance and Amazon, with 1M LPU-like AI ASIC units shipping to a Chinese CSP by end-2026 at ~$4,000 ASP.
  • $INTC: @tradergokux sees a 3-week flag consolidation as healthy HTF behavior. Calls it a “lead horse” that started the rally first and is consolidating first. Breakout level: 122.8.
  • $FSLR: @NullOreo_ initiated on China blocking solar equipment exports to Tesla, emphasizing FSLR’s US-domiciled, financially sound positioning as a clean domestic energy supply chain play. Catalyst: US solar manufacturing buildout narrative.
  • $GLW: @NullOreo_ identified as a dual-theme play: AI optical interconnects + US solar supply chain (Corning expanding Michigan solar capacity). CEO explicitly cited “domestic security” as the driver.
  • $LUNR: @KotlinerBTC analysis shows the stock is in a steep uptrend after NASA contract win, with implied volatility still “lying flat” despite the price surge—suggesting the explosive move has not yet occurred. He prefers LUNR over ASTS for the SpaceX IPO catalyst play, citing lower competitive overlap with SpaceX. Call buying concentrated around the 40/50 strikes for the SpaceX IPO window.
  • $NOK: @AntonLaVay, @jdhasoptions, and @labubu_trader all flagged NOK after Nvidia highlighted AI-RAN in its earnings call. JD holds LEAPS with +60% profit, targeting $20 by next year on sell-side estimates that exclude AI-RAN optionality entirely. Labubu added immediately after Jensen’s mention. AntonLaVay framed AI-RAN as a “free long-dated option” within the NOK thesis.

Long-term / Fundamental Positions

  • $BB (Blackberry): @qinbafrank continues to build the thesis around QNX as the embedded safety OS for Physical AI, with FedRAMP recertification and a 5% share buyback authorization adding near-term catalysts. The core logic: QNX + NVIDIA IGX Thor creates a unified safety-critical edge AI platform for robotics, medical, and industrial applications.
  • $NOK: Beyond the short-term AI-RAN catalyst, @AntonLaVay has built a detailed long-term thesis around the “Americanization” of NOK’s management team, with Infinera CEO David Heard effectively running a reverse-integration of Nokia’s optical networking business. The Q1’26 disclosure framework now separates Optical/IP growth assets from legacy mobile infrastructure, enabling a potential sum-of-the-parts re-rating.
  • $IBM: @qinbafrank positioned IBM as a long-term quantum computing play after the US government’s $10B investment in its “Anderon” quantum chip fab, framing it as “Intel-style government strategic investment in a fallen giant” with a 2026 quantum advantage target.
  • Power Infrastructure (Vertiv, Eaton, Schneider, Delta): @qinbafrank mapped the 800V HVDC supply chain comprehensively. These are multi-year compounders on AI data center power delivery, with Vertiv seen as the most direct early beneficiary.

Trading Activity & Holdings (VIP & High-Weight Traders)

  • @labubu_trader (VIP): Bought $NOK immediately after Nvidia’s Jensen called out AI-RAN during the earnings call @labubu_trader. Stopped out of $CBRS at $280 from a $300/$320 entry for a loss @labubu_trader. Bought $NVDA at the $215 max pain level, exactly hitting the low of the day, and opened a $215 put butterfly for +180% gain @labubu_trader.
  • @tradergokux (High): Added back full position in $ARM on the 229 breakout after trimming in prior weeks @tradergokux. Confirmed swing entry in $LITE on the pullback @tradergokux.
  • @AntonLaVay (Medium-weight, but high signal on NOK/Japan): Bought back 6981 (Murata) after realizing Nvidia’s physical AI/robotics narrative creates incremental high-end MLCC demand @AntonLaVay. Holds $NOK LEAPS and explicitly stated he is using LEAPS to ride through volatility rather than trading around events @AntonLaVay.
  • @jdhasoptions (High): Heavy LEAPS position in $NOK with +60% unrealized gain, conviction comparable to “buying INTC at 40 last year” @jdhasoptions.
  • @mat78704 (High): Positioned for a May 29 low, plans to buy $IBT calls or $NAIL as a homebuilder rebound play @mat78704.

Off-Theme Highlights

  • $DELL: @jukan05 flagged Dell surging on CPU-driven Agentic AI server demand, noting that CPU servers for Agentic AI share many components with general-purpose servers but carry higher ASPs—a structural tailwind for Dell’s existing strengths. Validated by Trump explicitly naming Dell (@Balder13946731).
  • $ENPH: @tradergokux identified a weekly Stage 2 breakout with a Cup & Handle pattern on heavy volume.
  • $HPE: @tradergokux flagged a triangle breakout “surfing EMA8” on daily, framing it as a mid-term swing AI play.
  • $AEVA: @tradergokux posted a retrospective case study on a VCP breakout that delivered +400% in two months—educational, not an active call, but signals his attention to lidar/autonomous sensing names.
  • $IBM: @qinbafrank built a position on the quantum computing thesis, noting IBM was the first buy on US government quantum investment news.

Notable Perspectives & Insights

@AntonLaVay: “High Rates Won’t Kill AI Infrastructure Stocks”

The week’s most framework-shifting analysis. AntonLaVay argued that AI infrastructure profit-pool companies are not traditional long-duration growth stocks vulnerable to DCF compression from higher rates. Instead, they exhibit “short-duration growth + industrial bottleneck + pricing power + current-period earnings revision” characteristics. Rising rates hurt weak competitors more than strong incumbents by raising barriers to new capacity, while customers pre-pay and sign long-term contracts, effectively shortening the business’s duration. The policy implication: “You cannot mechanically say high rates = AI infrastructure stocks should fall” (@AntonLaVay).

@NullableX: The China Crackdown Is Not About Securities Law—It’s About Capital Control for Industrial Policy

NullableX provided the deepest structural read on the Futu/Tiger crackdown. His argument: (1) China needs to plug the fiscal hole left by the property downturn, and capital gains tax on offshore assets held through gray channels represents lost revenue. (2) More importantly, Beijing needs domestic capital directed toward domestic industrial upgrading—semiconductors, AI, advanced manufacturing—and views outbound portfolio investment as capital that “doesn’t listen.” (3) The logical endpoint is a global taxation system modeled on the US, requiring cooperation with the US to track Chinese citizens’ global financial flows (@NullableX). This elevates the regulatory action from a sector-specific event to a structural signal about China’s capital account management trajectory.

@jukan05: Japanese Materials Companies Are Structurally Bad at Monetizing Scarcity

A sharp, investible insight: Nittobo, with critical T-glass supply for AI packaging substrates, explicitly refuses to raise prices, citing customer relationship preservation and market share defense (@jukan05). This contrasts with Korean memory makers who aggressively price for scarcity and reinvest profits into technology moats. The implication: not all “shortage beneficiaries” are equal. Companies with pricing power and willingness to use it (SK Hynix, Micron) will outperform those with market dominance but a cultural aversion to price increases (Nittobo, Ajinomoto). @LinQingV concurred: “Monopoly positions without pricing power monetization are hard to generate alpha from.”

@TJ_Research: NDX Valuation Is Not As Extended As Previously Believed

A data-driven corrective to his own earlier caution. After discovering issues with Bloomberg’s composite forward P/E calculation for the NASDAQ Composite (3,000+ constituents), TJ_Research shifted to the NDX (NASDAQ 100) as a cleaner metric. The NDX forward P/E is 24.3x, with 27.5x as the historical upper bound—implying ~13% upside to the zone where he would begin reducing beta. His prior call that valuations had returned to “expensive” territory was revised to “not very expensive,” and his beta-reduction timeline has been pushed out (@TJ_Research). This revision from a VIP commentator who is naturally cautious on valuations is a significant sentiment signal.

@qinbafrank: The Nvidia Edge Computing Re-segmentation Is a Valuation Narrative, Not Just Accounting

An underappreciated signal from Nvidia’s earnings: the company restructured its reporting into two platforms—Data Center and Edge Computing—with Edge including AI-RAN, automotive, robotics, and AI PCs. @qinbafrank argues this is a deliberate attempt to reframe Nvidia’s TAM from “hyperscaler capex dependent” to “AI operating system for the physical world,” introducing a third growth curve (Physical AI) beyond cloud and enterprise (@qinbafrank). This is the kind of narrative engineering that can support multiple expansion even as Data Center growth rates inevitably decelerate.

The Read

The Broadening Is Real—And That’s Both Bullish and a Warning

The most significant development this week is not any single stock move but the structural broadening of the AI infrastructure trade. The rally has migrated from NVDA/GPU-centric to CPU (AMD, INTC, ARM), power delivery (NVTS, Infineon, Vertiv), optical interconnects, packaging substrates, solar/energy, and even legacy OEMs (Dell, HPE). This is what a healthy secular bull market looks like: capital flows to the next marginal bottleneck as the previous one gets priced. The convergence across commentator types—industry analysts identifying the bottlenecks, macro commentators framing the TAM expansion, traders executing on the technical setups—is as robust as it gets. The breadth argument from @TradeBrigadeCo (“breadth expansion via earnings catchup in the Other 493”) is supported by the data.

However, breadth expansion in a late-cycle environment can also signal that the easy alpha from the core theme has been harvested and capital is now chasing marginal beta in second- and third-derivative plays. The CPU theme, while fundamentally sound, is already being framed as “the next big thing” by multiple voices simultaneously—a consensus that forms quickly in a social media-driven market can also unwind quickly. The Japanese materials dynamic—where a genuine bottleneck exists but the bottleneck-holder refuses to monetize—is a reminder that not every “shortage” translates to shareholder returns.

The Geopolitical Overhang Is Being Underpriced

The market’s reaction to the US-Iran situation has been binary and headline-driven: rally on “deal framework” headlines, sell off on “escalation” headlines. But @ShanghaoJin and @AntonLaVay both point to a deeper structural tension that the market has not fully priced. The US has committed its credibility to confronting Iran; backing down without a decisive outcome weakens deterrence globally, including in the Taiwan Strait. Iran has demonstrated willingness to actually execute on Strait of Hormuz disruption threats, unlike the US which has not followed through on “bomb them back to the stone age” rhetoric. This is not a symmetric negotiation—it’s a game of credibility chicken where both sides have reasons to escalate and constraints on backing down. The Pakistan-brokered talks are a positive signal, but the weekend’s activity (Trump canceling plans, military on standby) suggests the risk of a break in negotiations is non-trivial. Oil above $100 and 10Y above 5% remain the “tail risk that nobody is positioned for” scenario.

The China Policy Shock Has Portfolio Construction Implications

The Futu/Tiger/Long Bridge crackdown is not merely a sector event. Combined with @NullableX’s structural analysis, it signals that Beijing is systematically closing off channels for domestic capital to flow offshore. The practical consequence: the marginal Chinese retail bid that has supported US tech and growth stocks for years will diminish. More importantly, the policy risk premium for any asset class dependent on Chinese household capital flows—US-listed Chinese ADRs, Hong Kong stocks, crypto on-ramps, and even the “Chinese retail flow into US tech” dynamic—has been structurally repriced. For the portfolio manager, this reinforces the case for being overweight US-domiciled, domestically-oriented companies (the $FSLR, $GLW thesis) and underweight anything dependent on Chinese capital flows or regulatory goodwill.

What to Watch

  • US-Iran Negotiation Outcome (Ongoing, with potential Memorial Day weekend catalyst): Pakistan’s Army Chief visit to Iran produced a statement describing “encouraging progress.” An MOU for a 60-day ceasefire extension with gradual Strait of Hormuz reopening would be a significant risk-on catalyst, particularly for oil-sensitive sectors and the memory/HBM trade that depends on stable energy costs. A breakdown in talks would spike oil and VIX. Watch the Strait of Hormuz shipping data and Iranian/ US official statements over the holiday weekend. Per @qinbafrank, the next formal negotiating round is expected after May 31.

  • MRVL Earnings (May 27) and AVGO Earnings (June 3): @Balder13946731 flagged both as “relatively bullish” setups. MRVL is critical for validating the custom ASIC and optical interconnect themes that have driven the week’s narrative. A strong report would confirm that the AI infrastructure trade extends well beyond NVDA. A disappointment would test the breadth thesis. @KotlinerBTC notes MRVL IV Rank is at 86, tied for highest across his coverage universe—the options market is pricing a significant move.

  • SpaceX IPO (June 12) and Pre-IPO Positioning: The SPCX IPO date is now formally set. @Balder13946731 and @Balder13946731 have provided detailed instructions for Schwab and Robinhood participation. @KotlinerBTC has identified a notable divergence: ASTS options are showing large-scale call selling and put buying into strength, with 10,000+ calls closed in the post-IPO June 18 expiration. This suggests the “sell the SpaceX IPO event” trade is already being positioned. LUNR, by contrast, is seeing aggressive call buying into the same window, implying a preference for names with lower competitive overlap with SpaceX. This is a tradeable divergence.

  • NVDA Post-Earnings Option Dynamics: @KotlinerBTC notes that NVDA’s IV has collapsed post-earnings and the stock neatly hit its $215 max pain level on Friday. The Put/Call volatility spread has rebounded from extreme negative territory (unprecedented in two years) back toward zero—a healthy normalization. If IV stabilizes at the new lower level and the Put/Call spread stays near zero, the setup for the next leg higher is constructive. If IV continues to drift lower toward the 30% baseline, it signals a loss of speculative interest—bearish for the broader semi complex.

  • AgentMat’s May 29 Cycle Low: @mat78704 is projecting a near-term low around May 29, followed by a rally to ATH into June 8–12, which he characterizes as a “dance of demons and monsters” period where lagging sectors (specifically homebuilders) play catch-up. His track record on short-term cycle timing has been noted by the community. If the low materializes, his buy zones are IBT calls and NAIL (homebuilder leveraged ETF).

  • TSMC Bonus Cut Rumors: @jukan05 flagged unconfirmed reports from Taiwan’s Blind that TSMC may cut employee bonuses, with employees threatening unionization. If confirmed, this would signal margin pressure or a deliberate shift toward shareholder returns at the expense of labor—both would have implications for the foundry cost structure and the broader semi capex cycle. This is unconfirmed but worth monitoring given TSMC’s centrality to every AI infrastructure thesis.