Market Brief(X) — May 11–May 14, 2026

2026-05-15 • Twitter

Executive Summary

This week saw an explosive AI capex supercycle trade converge with a historic Trump-Xi summit, producing a euphoric rally that peaked mid-week before a sharp macro-induced reversal. The dominant narrative threaded AI model monetization (Anthropic at $45B ARR, OpenAI at $35B) directly into physical semiconductor bottlenecks — memory, storage, optics, and advanced packaging — with the Cerebras IPO serving as the emotional crescendo. By Thursday night, however, surging Japanese and US long-end yields, combined with sticky PPI/CPI prints and unresolved Iran/oil risk, forced a violent reset. The central tension now: AI fundamental momentum is undeniably real and accelerating, but the macro cost of capital is rising simultaneously — and the market is being forced to price both at once for the first time since the March lows.

Key Themes & Trends

AI Model Monetization Goes Exponential — and the Market Re-Prices the Entire Stack

The week’s foundational signal came from revenue scale: Anthropic hit $45B ARR and OpenAI reached $35B, with combined annualized revenue adding ~$4B/month (@zephyr_z9). This validated the thesis that AI spending has converted from speculative capex to revenue-generating infrastructure. The downstream implication, articulated forcefully by @FundaAI, is that US TMT funds (notably Coatue and Altimeter) are rushing to rebalance from software/internet into semis/infrastructure — with Coatue’s semi+infra exposure jumping from 35% to 58% of book. This institutional rotation, still in early innings, is the structural bid beneath the entire AI hardware complex.

Memory Supercycle Intensifies — Supply Cliff Meets Demand Explosion

Memory was the epicenter of speculative flow. @jukan05 reported Samsung and Kioxia cutting legacy 2D NAND production, triggering a supply cliff — with MLC NAND spot prices up 300%+ since late 2024. Simultaneously, a looming Samsung semiconductor union strike threatened to remove 3-4% of global DRAM supply (@jukan05), acting as jet fuel on already-burning prices. @KotlinerBTC documented MU’s IV exploding to two-year highs with Put/Call skews at -13% to -17% — what he called a “full boil” sentiment extreme. The structural case: Samsung’s Chey Tae-won warned that if supply isn’t increased, “customers will find ways to use less memory” (@jukan05). Contrarian note: @ShanghaoJin repeatedly challenged whether DDR commodity revenues deserve growth-stock P/E multiples, arguing the HBM thesis is sound but the DDR valuation overlay is speculative.

Trump-Xi Summit: Tactical Wins, Unresolved Structural Questions

The President’s Beijing visit with 30+ CEOs (Jensen Huang, Elon Musk, Tim Cook, Larry Fink) dominated mid-week price action. @NullableX framed it explicitly as a Miran playbook event: China opening markets to absorb dollar liquidity in exchange for geopolitical concessions. Immediate deliverables included H200/RTX 5090 sales approval into China (@jukan05), a 200-plane Boeing order, and a White House readout claiming Xi agreed to help mediate Iran and keep Hormuz open (@qinbafrank). However, @qinbafrank’s detailed post-summit analysis revealed a critical gap: the Chinese official readout did not mention Iran or Hormuz, and maintained a hard Taiwan red line. @SV_Nomad was bluntest: “Why would China give high consideration to bail Trump out?” The summit was a tactical success — markets rallied on the optics — but the structural deliverables remain ambiguous.

Heterogeneous AI Inference: The SRAM Land Grab

@fi56622380 (VIP) published a defining framework on heterogeneous AI inference: the post-GTC consensus is that pure SRAM-route AI inference (Cerebras standalone) is economically unviable, but SRAM chips embedded as decode-FFN accelerators within mainstream AI ASIC pipelines (Nvidia LPX, Google TPU/Marvell, AWS Trainium/Cerebras, ByteDance/Qualcomm) unlock massive TAM expansion. @labubu_trader (VIP) built long CBRS position around this thesis, with detailed valuation scaffolding: if 7,000 Rubin racks carry LPX at $3M/rack, LPU alone is a $21B revenue business. The key risk, flagged by @LinQingV, is integration depth — superfical prefill/decode splitting doesn’t work economically; deep integration of Cerebras into Trainium’s FFN layer is the make-or-break. CBRS opened at $350 (IPO $185), a 90% first-day pop (@Balder13946731).

Intel’s Foundry & Packaging Renaissance

$INTC transitioned from meme to institutional thesis. @LinQingV produced a landmark technical deep-dive on EMIB-T, arguing it may reshape advanced packaging over the next two years by solving the power delivery bottleneck for HBM4, with a cost advantage over TSMC CoWoS-L at large reticle sizes. @jukan05 reported Tesla’s AI6 chip moving from TSMC to Intel Foundry under administration pressure; Apple M7 confirmed on 18A-P; MediaTek designing AWS Trainium with EMIB (@jukan05). @KotlinerBTC documented INTC’s gamma squeeze dynamics and the extraordinary signal of the nearest-dated max pain being “abandoned” — option sellers capitulating. However, @kayliatyyy noted INTC’s Thursday underperformance versus NVDA/MU, speculating that LBT missing the China trip mattered.

From Semis to Space: The Broadening Bull Market

The rally widened dramatically. $RKLB broke $100 and accelerated as institutional volume surged (@tradergokux), with @qinbafrank noting his position was near 2x. SpaceX IPO prospectus expected as soon as next week, targeting a $1.25T+ combined valuation (@Balder13946731). $NBIS delivered a blowout quarter (ARR +674% YoY, contracted capacity >3.5 GW) and was flagged by multiple traders as the neocloud winner (@tradergokux, @LeoYuen13). The breadth thesis, advanced by @tradergokux, was that “this is what a strong bull market looks like” — money rotating across sectors rather than concentrating.

The Optics Re-Rating is Underway

$LITE surged 17% on Nasdaq-100 inclusion plus Innolight’s 2.4T demand signal (@FundaAI). @FundaAI articulated the structural repricing: LITE is transitioning from cyclical optical component maker to AI datacenter bottleneck. @qinbafrank contextualized the TAM math: optical market ~$260B in 2026 vs memory at $3.7-4T — the size differential explains market cap gaps, but the growth rate and technology lock-in for optics players with EML, InP, and CPO capabilities is compelling. @zephyr_z9 noted Nokia and Cisco both releasing 2.4T products, and @AntonLaVay detected institutional LEAPS accumulation in NOK.

Market Sentiment

Short-term (next 1-2 weeks)

Euphoria rapidly gave way to caution by Thursday’s close. The early-week tape was “insanity” (@tradergokux) and “once or twice per lifetime environment” (@ZaStocks). By Friday morning, the tone had pivoted sharply: Japanese CGPI print triggered a US long-end selloff, 30Y broke 5.1%, and @mat78704 — who had been remarkably accurate with intraday timing — projected a low on May 18-19 and a high on May 20 before a larger top in June. @Balder13946731 warned: “if UVXY rises, avoid the sharp edge.” @qinbafrank explicitly recommended partial profit-taking, citing the unsustainable co-movement of oil, yields, and equities all moving higher simultaneously. Short-term posture: profit-take into strength, await the May 18-19 dip for re-entry.

Long-term (weeks to months)

The structural bull case remains intact but is now facing its first genuine macro test since March. @kayliatyyy articulated the core dilemma: “Can structural AI-driven deflation outpace geopolitical inflation? Short term no, long term yes.” @TJ_Research warned that the Fed ship is turning — Warsh’s first meeting will likely shift from easing bias to neutral — and that “big tech CAPEX + strong consumption = inflation risk.” @ZaStocks maintained the 1996-97 analog (not 1999) — implying more room to run but with pullbacks. @NullableX tied everything to the Miran framework: if China agrees to absorb dollar liquidity by opening its markets, the dollar floodgates open; if negotiations fail, rate cuts are blocked. The long-term conviction is bullish but conditional on the Iran/oil resolution path.

Sentiment Balance & Shift

The within-window shift was dramatic: Monday-Wednesday was characterized by disbelief-suspension and FOMO (@TJ_Research: “the music hasn’t stopped, we’re just in the stage where everyone is dancing”). Thursday afternoon through Friday morning introduced a macro anxiety overlay that had been absent since the March lows. The split is now between traders who see this as a healthy pullback within a raging bull (@labubu_trader: “nothing burger, still bullish”) and macro commentators who see the oil-rate-rally co-movement as unsustainable (@qinbafrank: “oil, bond yields, and stocks all moving in the same direction is problematic”). This tension is the most informative feature of the current market — and has not been resolved.

Key Figures & Assets

Short-term / Technical Trades

  • $RKLB — @tradergokux (High, trader) + @AntonLaVay (Medium, investor) convergence. Institutional accumulation confirmed via volume signature and diminishing ATR post-breakout. Key level: holding $100 psychological support. Risk: SpaceX IPO may rotate space-premium capital away.
  • $MRVL — @tradergokux (High) + @KotlinerBTC (High) tracking. Hit fresh ATHs; the 5-year base breakout implies a secular move, but curious Put/Call divergence — IV spiking yet traders buying Puts for protection, signaling less pure conviction than memory names.
  • $FCEL — @tradergokux flag breakout, +58% in 4 sessions. Thin liquidity, high ADR name.
  • $TSLA — @tradergokux identified a weekly MACD golden cross and flag breakout above the multi-month downtrend. @KotlinerBTC documented massive 475C positioning for June expiry. @Balder13946731 flagged gamma squeeze potential above 450. China FSD approval and SpaceX IPO serve as catalysts.

Long-term / Fundamental Positions

  • $Q (Qualys? — actually a materials play) — @tradergokux (High) is building this to his 3rd largest long-term holding. US semiconductor materials supplier, benefiting from $12B Project Vault, with a new Delaware super-factory positioning as the sovereign materials center for US fabs. Reported AI-related revenue +25%, EBITDA +52%. Thesis: national security asset, similar to INTC in strategic importance but under-covered.
  • $CRCL — @TJ_Research (VIP) + @qinbafrank (High) both hold long-term positions. The Clarity Act advanced out of Senate Banking Committee 15-9 (@TJ_Research). TJ’s thesis: CRCL is not just a stablecoin issuer but a payment infrastructure platform — ARC chain, Agent Stack, Managed Payments — in very early innings. Added small position on China trip narrative.
  • $NOK — @AntonLaVay (Medium, investor) detected LEAPS accumulation via rising long-dated IV, positing 13F institutions are building positions. @KotlinerBTC (High) confirmed a structural shift in NOK’s trading character since mid-February, with gamma squeeze dynamics above the massive 12-strike call wall. Thesis: optical networking beneficiary, Infinera integration creating a US-centric optical franchise.

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

  • @labubu_trader (VIP) — Converted MU and SNDK 2x ETF positions to shares (same size), de-leveraging ahead of expected 10-20% correction in memory. Re-entered MU at 5d EMA test. Sold half VICR and WYFI at CBRS IPO; sold all DGXX and BTBT; established CBRS long at $300 and $320 (@labubu_trader). Still long CBRS with small position, evaluating further adds based on price action.
  • @NullableX (VIP) — 70% allocated; shifted 30%+10% cash into China-related equities (TSLA, NVDA, BABA, other China names) for the Trump-Xi summit trade. Explicitly event-left-side positioning. By Wednesday, marked the trade as profitable and flagged intention to “sell the news” into Friday’s outcome.
  • @TJ_Research (VIP) — Trimmed 5% of INTC position for the first time since March lows, using a trailing stop methodology (10% pullback from highs). Remains broadly long but beginning systematic profit-taking on extended names.
  • @AntonLaVay (Medium, investor) — Added MRAM calls during the Tuesday CPI dip, citing relative strength as the selection signal. Added ORCL Jan 2027 500C as “lottery tickets” (@AntonLaVay). Maintains PS as a portfolio hedge (Ackman-controlled, counter-market).
  • @mat78704 (High, trader) — Called SPY top at 12:15pm Thursday for profit-taking, projecting next low May 18-19 and next high May 20. Remains bullish on SPX/QQQ structurally, advising against shorting.

Off-Theme Highlights

  • $MRAM — Both @AntonLaVay and @ShanghaoJin are flagging Everspin as a long-dated military/aerospace persistent memory play. TSM’s technology symposium confirmed the shift from eFlash to MRAM/RRAM for sub-28nm nodes (@zephyr_z9), providing a massive secular tailwind. @AntonLaVay added on the CPI dip based on relative strength. The stock has a small float and is vulnerable to liquidity squeezes in both directions.
  • $AEHR — @AntonLaVay outlined the wafer-level burn-in thesis: AI ASICs at higher TDP + SiPho in CPO (non-replaceable) = mandatory WLBI testing. The thesis is that AEHR’s FOX-XP system is becoming a required tool for both the ASIC ramp and the CPO transition. The setup is technically sound but the timeline is uncertain — design-in now, volume in 2027-28.

Notable Perspectives & Insights

Franktradinglog on Portfolio Construction: “The Organic Whole” @Franktradinglog (High, trader) produced what may be the most analytically rigorous post of the week — a mathematical defense of portfolio diversification that goes far beyond platitudes. His key argument: owning ten semiconductor stocks is not diversification, it’s a 10x leveraged bet on the same factor with correlation → 1 in risk-off. True diversification lives on the correlation matrix, not the ticker count. The punchline is geometric: “σ_p² = w₁²σ₁² + w₂²σ₂² + 2w₁w₂ρσ₁σ₂” — and the volatility drag equation “g ≈ μ - σ²/2” means reducing portfolio vol from 30% to 20% adds 2.5 percentage points to long-term compound returns, all else equal. This is worth reading in full for any investor running concentrated momentum books.

AntonLaVay on “Follow the LEAPS” @AntonLaVay shared a practical institutional-tracking methodology: monitor long-dated OI increases and LEAPS IV expansion in stocks within hot sectors. His logic: 13F filers build positions through LEAPS, not short-dated options; short-term OI is noisy with retail flow; but sustained long-dated IV expansion and OI growth leaves footprints that can’t be hidden. He claims this flagged NOK’s institutional accumulation before it became consensus. Simple, replicable, and analytically sound.

fin’s SRAM Heterogeneous Inference Framework @fi56622380 (VIP) delivered the week’s most important technology strategy analysis. The core insight: SRAM-route AI chips (Cerebras, Groq LPU) are economically unviable as standalone full-stack inference — but become structurally valuable when integrated as decode-FFN accelerators within heterogeneous inference pipelines. Nvidia’s Rubin + LPX architecture is the template: prefill and decode-attention on HBM GPUs, decode-FFN on SRAM LPUs. This framework values Groq at $200B+ (not $20B) and provides the valuation architecture for CBRS. The critical execution variable: how deeply Cerebras integrates with Trainium, not whether it partners.

kayliatyyy on the Macro-Cyber Risk Nexus @kayliatyyy flagged a critical, under-discussed signal: “Recently, both Japanese and US bond markets have started to react. That feels very suspicious.” The JGB selloff → US long-end contagion channel has historically been a reliable early warning for broader risk repricing. Combined with CPI/PPI beats, oil above $100, and a new Fed chair about to take office, her warning to “pay more attention to macro than before” deserves weighting despite the overwhelming AI narrative dominance.

NullableX on the Miran Playbook in Real Time @NullableX connected Trump’s “open up China” rhetoric directly to the Miran framework: dollar devaluation requires foreign asset absorption to avoid a “fireworks” scenario. China’s willingness to open its capital account and accept dollar inflows would greenlight a massive liquidity expansion. His read: Trump is following the Miran script methodically, and the summit was a negotiation over what price the US pays for that accommodation. Taiwan, Iran, and tech transfers are the bargaining chips.

The Read

The AI Capex Boom is Real — and is Now Being Priced at Peak Momentum

The fundamental signal is unambiguous: Anthropic and OpenAI are generating revenue at a scale that validates the infrastructure buildout thesis. This is no longer a “hope” trade — it’s a monetization cycle. The institutional rotation from software to semis (Coatue, Altimeter) is evidence that sophisticated allocators are repricing the entire stack. Every bottleneck — HBM, advanced packaging, optics, power, networking — is becoming a separately investable theme. This is the structural backbone of the rally, and it has not been invalidated.

But the price action has front-run the fundamentals to a degree that creates near-term vulnerability. Memory IV ranks at 92-100 (@KotlinerBTC). CBRS opened at 90% above IPO. Semis have gone parabolic. The market is now pricing perfection — and perfection leaves no room for bad news.

The Macro-Narrative Collision is the Real Story

The week’s most important development was not the summit or the IPO — it was the return of macro as a binding constraint. CPI at 3.8%, PPI at 6.0%, 30Y breaching 5.1%, Japan triggering a global bond selloff, oil at $102 — all while equities hit all-time highs. This is an unstable configuration. @qinbafrank correctly identified the “grey rhino” — sustained high oil becomes a persistent inflation impulse that blocks the rate-cut path, and the market hasn’t fully priced the probability that Warsh’s Fed may have to lean hawkish.

The unresolved variable is Iran. The summit produced White House claims of Chinese mediation assistance, but the Chinese readout conspicuously omitted Iran. If Hormuz remains blocked through summer, oil will grind higher, CPI will stay elevated, and the “AI valuations + tightening financial conditions” combination will stress the system. The market is betting on a diplomatic resolution; the bet has not yet paid off.

The Near-Term Setup: Buy the Dip, But Which Dip?

@mat78704 and @labubu_trader both expect the pullback to be shallow and buyable — targeting May 18-19 lows, May 20 highs. The weight of evidence supports this: earnings are too strong, the institutional bid is too deep, and the administration’s incentive to maintain asset prices is too powerful for a sustained breakdown. However, the macro tail risk (oil spike, bond market dysfunction, Warsh hawkish surprise) is non-trivial and growing. Position sizing and correlation management — @Franktradinglog’s framework — are the appropriate risk tools here.

The shift within this window — from euphoric “once in a lifetime” to cautious “take profits into strength” — is itself the signal. The easy part of this rally is likely behind us. The next leg requires navigating genuine macro cross-currents, and the market’s ability to shrug off bad news is being tested for the first time since March.

What to Watch

  1. NVDA Earnings — Wednesday May 20 — The single most important event of the next two weeks. @KotlinerBTC shows IV already breaking above prior earnings-week highs, with Put/Call skew at the most bullish level in a year. The setup: expectations are elevated, the China H200 approval is a tailwind, and Jensen’s presence on the China trip signals strategic positioning. A beat-and-raise would validate the entire semi ecosystem; a guidance disappointment would trigger a sharp de-rating of extended positions. @ShanghaoJin argues NVDA is being overlooked as “the top student” while everyone chases comeback stories — watch for a potential relative strength surprise.

  2. FedSpeak and Warsh Transition — Ongoing — Miran resigned from the Fed board effective at Warsh’s swearing-in (@kayliatyyy). Warsh’s first public comments will be scrutinized for any shift from the Powell-era easing bias. Any hint of rate-hike discussions would be the most significant macro shock the AI rally has faced. @kayliatyyy: “we should not completely rule out the possibility that the Fed could start talking about raising rates again.”

  3. China’s Iran Statement — Imminent — The White House claimed China agreed to mediate; China has not confirmed. The divergence in official readouts identified by @qinbafrank is the key monitoring point. A formal Chinese commitment to mediation would be a bullish risk-on catalyst; silence or denial would feed the “grey rhino” narrative and pressure oil-sensitive assets.

  4. SpaceX IPO Prospectus — As Soon As Next Week — CNBC reports the prospectus could drop next week with a June 8 roadshow (@Balder13946731). $70-75B IPO at $1.25T+ combined valuation. The liquidity drain risk is real — @Balder13946731: “if it really drains, it could drain dry.” Watch the impact on TSLA (Elon selling to fund participation?), RKLB/ASTS (rotation within space theme?), and broader market liquidity.

  5. US 10Y 4.6% Threshold — @qinbafrank identifies 4.6% as the pain threshold where equities begin to feel real pressure. Currently at 4.53% and rising. Watch whether Bessent deploys Treasury buyback tools to defend this level upon return from China. A clean break above 4.6% would be a risk-off signal across duration-sensitive assets.

  6. Google I/O — Tuesday May 20 — @jukan05 reports a new Gemini model roughly on par with GPT-5.5 will be announced. This is a direct competitive response and could reshape the AI model hierarchy narrative that has been dominated by Anthropic/OpenAI this cycle. Positive surprise would benefit GOOGL and the broader AI infrastructure trade; a disappointment would reinforce the “Anthropic/OpenAI duopoly” narrative.

  7. Corsica267’s Settlement/Liquidity Framework — In @Corsica267’s Thursday post, he flagged that Friday and next Tuesday’s settlements will cause capital and rate fluctuations from Thursday through Wednesday’s night session. His tactical read: “Gold will have volatility, Nasdaq relatively impacted, SPX likely stable. Combine with 3-month T-bill rates, DXY, and Put/Call ratio — if the decline is confirmed as funding-driven, buy the bounce and go long during Wednesday’s day session.” This is a specific, tradeable liquidity roadmap worth tracking against realized price action.