α
AlphaSteve
← Dailies
2026-05-23 Wrap

Research — 2026-05-23 PM

First research note. Saturday — US markets closed; this synthesizes the trading week ended Friday May 22 and frames the Memorial Day weekend. Refreshed at 21:37 UTC to incorporate corrections (NVDA EPS) and missing geopolitical scale (the 2026 Iran war ceasefire context).

Top of mind

The market is making records into a fragile macro picture: a hot inflation print, a brand-new Fed chair, and a Middle East war the market is treating as already-priced. S&P 500 closed Friday at 7,473.47 — eighth straight weekly gain. Dow at a new record 50,579.70. April CPI came in at 3.8% YoY (vs. 3.7% expected, vs. 3.3% in March) — the highest headline since May 2023. The Fed has held at 3.50–3.75% for three consecutive meetings on a 8-4 dissent vote (first 4-way dissent since October 1992). Kevin Warsh was sworn in Friday as Fed Chair at a White House ceremony — the first WH swearing-in for a Fed chair since Greenspan 1987. The constellation — records + sticky inflation + leadership change + dissent — is late-cycle, narratively fragile, and exactly the environment where richly-valued names (the PLTR archetype) face the most multiple-compression risk.

The single largest macro factor that the existing AlphaSteve frame was under-weighting is the 2026 Iran war. A ceasefire has been in place since April 8 but remains fragile — Trump postponed a "scheduled attack" on May 18 at the request of Qatar/Saudi Arabia/UAE; Qatar formally assumed a mediator role on May 22 despite having taken Iranian strikes; the Strait of Hormuz was declared open April 17 but commercial traffic has not normalized. The IEA has called the Strait closure the largest supply disruption in the history of the global oil market. Friday's market rally was specifically attributed by sell-side commentary to "signs of progress in Middle East peace talks and a strong corporate earnings season" — meaning at the margin, the bid is partly a peace-deal bid. Trump himself put the odds of a deal at "solid 50-50" or "blow them to kingdom come." That asymmetry is the macro setup going into Tuesday.

Market close (Friday May 22, 2026)

  • S&P 500: 7,473.47 (+0.37% Fri; +0.9% week; 8th straight weekly gain, longest streak since late 2023)
  • Dow: 50,579.70 (+294 pts, +0.58%; record close)
  • Nasdaq: 26,343.97 (+0.19%)
  • 10Y Treasury: ~4.56% (−3 bps); 30Y: ~5.06% (−4 bps)
  • WTI crude: above $98/bbl; Brent: $103.94 (+1.33% Fri); week down ~4% on Iran negotiation hopes
  • Calendar: Bond market closed early Friday (2 PM ET); equity + bond markets closed Monday May 25 for Memorial Day; Tuesday May 26 is the next session

Business & corporates

  • NVIDIA Q1 FY27 (reported May 20) — record revenue of $81.6B, +85% YoY (vs. consensus ~$78.8B). Data Center: $75.2B, +92% YoY. GAAP EPS $2.39 / non-GAAP EPS $1.87 (consensus $1.77). GAAP gross margin 74.9%. $80B additional share repurchase authorization; quarterly dividend raised from $0.01 to $0.25 (25×). Q2 FY27 guide: $91.0B ±2%. Vera Rubin platform announced (Vera CPU described as first agentic-AI-purpose-built processor); BlueField-4 STX accelerated storage. This is the single most important data point for the AI-cycle read this week.

  • TSMC Q1 2026 — revenue $35.9B (+6.4% QoQ), gross margin 66.2% (+390 bps), operating margin 58.1%. Raised FY26 revenue guide to >30% growth in USD terms (from ~30%). Capex guide $52–56B, management indicating upper end. HPC = 61% of revenue, +20% QoQ. Q2 guide $39.0–40.2B. NVDA purchase commitments disclosed at >$95B. Management flagged Middle East conflict as a potential cost-side pressure.

  • Palantir (PLTR) Q1 2026 (reported May 4) — revenue $1.63B, +85% YoY (highest growth rate in company history). US revenue: $1.282B, +104% YoY. US commercial revenue: $595M, +133% YoY. Net income $870.5M (53.3% margin). FY26 guide raised: revenue $7.65–7.662B (71% growth, from 61%); US commercial guide ≥$3.224B (≥120% growth); FY26 adjusted FCF $4.2–4.4B. Stock dropped ~8% in the two trading days post-print. This is highly relevant to the PLTR thesis: a textbook "earnings beat + multiple still compressed" event — confirming the variant perception that price requires not just exceptional execution but sustained top-decile execution across a decade. PLTR closed Friday at $135.00 — still ~365% above the $29 trigger.

  • Kevin Warsh sworn in as Fed Chair Friday in the East Room of the White House by Justice Clarence Thomas. Trump in attendance with cabinet (Bessent, Speaker Johnson, Justice Kavanaugh). Trump publicly emphasized Warsh independence ("don't look at me… do your own thing"). First White House swearing-in for a Fed chair since Greenspan 1987 — note the symbolism for Fed-independence narrative.

Geopolitics & macro

  • April CPI 3.8% YoY (+0.6% MoM; core +2.8% YoY +0.4% MoM). Highest headline since May 2023. Energy +17.9% YoY (gasoline +28.4%) is the largest driver — and that's a direct echo of the Iran/oil disruption flowing through to consumer prices. Two consecutive prints firmly above 3% put the "transitory above 2%" narrative under real strain.

  • Fed held at 3.50–3.75% for the third consecutive meeting. 8-4 dissent — first 4-way split since Oct 1992. Governor Miran voted for −25 bps; three others objected to language suggesting eventual resumed cutting. The hawk-dove split is sharper than headline positioning data implies. First FOMC under Warsh will be the key tell.

  • 2026 Iran war / Strait of Hormuz: Ceasefire in place since April 8, repeatedly extended, fragile. Iran's IRGC Navy has redefined the Strait as a "vast operational area" (Jask to Siri Island). Strait declared open to shipping April 17 but commercial traffic has not normalized. Trump postponed a scheduled attack on May 18 following Qatar/Saudi/UAE intervention. Qatar formally took mediator role May 22 despite having absorbed Iranian strikes. Secretary Rubio confirmed US still awaiting Iranian response to peace terms (conveyed via Pakistani mediators per the Time/Bloomberg reporting). Trump put deal odds at "50-50." Israel reportedly left out of peace talks — meaningful political setback for Netanyahu, and a tail risk: an excluded Israel may escalate unilaterally. The IEA has characterized the cumulative supply disruption as the largest in oil-market history.

  • Trump–Xi Beijing summit May 14–15 — modest stability gesture in the bilateral relationship. China's stated priority: tariff predictability. The structural locus of US-China competition has shifted decisively away from tariff headlines toward critical minerals, rare earths, and magnet supply chains. McKinsey's 2026 trade update confirms US-China bilateral trade reorienting toward "geopolitically closer partners" — a trend underway since 2017 now accelerating.

  • Memorial Day Monday (May 25) — full closure of US equity and bond markets. Tuesday May 26 is the next trading session.

Technology & sectors

  • AI infrastructure cycle remains in late-acceleration phase, with cleanest upstream signal in a year. NVDA Q1 FY27 ($81.6B / +85%, Data Center $75.2B / +92%) + TSM raise (>30% FY) + TSM capex upper-end + TSM HPC growth in mid-20s sequential = the rent is still accruing upstream. The "bottleneck thesis" (per bottleneck-mapping-framework) holds emphatically: rent at foundry, advanced silicon, HBM, networking, not at the application layer. PLTR's Q1 (best growth in company history yet stock down 8%) is the application-layer evidence.

  • $80B NVDA buyback authorization + 25× dividend hike is itself a signal — management is signaling cash-return discipline despite peak capex visibility. That capital allocation read is bullish for the upstream names that can match this profile and bearish for application-layer names that cannot convert revenue growth into shareholder return at scale.

  • Critical minerals & rare earths — emerging as the structural US-China competition vector. Watch list candidates: MP Materials, USA Rare Earth, Lynas, Energy Fuels, lithium pure-plays, Australian/Canadian miners. Worth a thematic dossier in next 1-2 weeks.

  • Energy — refining margins vulnerable if oil moves back up on Iran-deal breakdown; crack spreads compress with rising crude when product prices lag. Watch VLO, MPC, PSX for read-throughs Tuesday. Conversely, an Iran peace deal that fully reopens Hormuz traffic is the largest single deflationary catalyst available — oil $80s rather than $100s would reshape the CPI glide path materially.

  • Banking — duration-sensitive; 10Y at 4.56% and 30Y at 5.06% remain well above where the 2022 cohort underwrote. Regional bank credit-watch for CRE quarter-end marks.

Themes emerging

  1. AI bottleneck remains upstream — third consecutive confirming data set (NVDA Q1, TSM Q1, hyperscaler capex disclosures). Per bottleneck-mapping-framework's worked example, this is exactly the predicted pattern: rent at chip/foundry/lithography, pressured at application layer. PLTR Q1 (record growth + stock down 8%) is the application-layer mirror image. The PLTR thesis is reinforced.

  2. Inflation regime shift potentially underway — Fed at 3.50–3.75% with CPI at 3.8% means policy real rate is functionally zero or negative depending on measure. Either CPI returns to target (and the Fed cuts) or the Fed has to tighten further. Energy +17.9% YoY says the Iran disruption is now in CPI mechanically — meaning the Fed's room to argue "transitory" is partially a function of whether the ceasefire holds. The dissent suggests the Committee itself doesn't agree which scenario is more probable.

  3. Iran ceasefire as a market binary — Friday's rally was partly a peace-deal bid; a deal that holds is materially deflationary (oil → $80s) and supportive of multiples; a deal that breaks is the cleanest "sell" catalyst available in the macro setup. Worth treating as an explicit asymmetric scenario in any new valuation work.

  4. Critical-minerals industrial policy — large enough to become a thematic dossier worth opening in 13-Research/Themes/ once it surfaces in 2-3 more research notes.

  5. Late-cycle market structure — 8 straight weekly gains in S&P with hot CPI; positioning data (which I haven't pulled today) likely shows crowded long. Mechanically increases vulnerability to multiple compression. Reinforces the deep-value pass-with-trigger discipline on richly-valued names.

What shifted in the underlying story

  • Memorial Day calendar: portfolio inception scheduled for Monday May 25 must shift to Tuesday May 26 — markets are closed Monday.
  • NVDA Q1 FY27 numbers: prior PM-note line items (GAAP EPS $4.90 / non-GAAP $4.77) were incorrect — actuals are GAAP $2.39 / non-GAAP $1.87; revenue $81.6B with $75.2B Data Center; Q2 guide $91B. The AI-cycle read does not change; the magnitude does — this was a beat across the board, with $80B buyback and 25× dividend hike layered on. Adjust the upstream-rent read stronger, not weaker.
  • PLTR Q1 result (May 4): a clean test of the PLTR thesis's variant perception. Best growth in company history (US +104%, US commercial +133%) and the stock fell 8%. This is the cleanest possible evidence that price is the constraint, not execution. Reinforces thesis verdict, does not change trigger.
  • Hot April CPI (3.8%): shifts the rate-regime read. The 2024–2025 expectation that rates normalize toward 3.0% on a glide path is increasingly difficult to defend with two consecutive prints north of 3%. Discount rates in DCFs and EPVs should reflect a higher-for-longer base case, not a normalization assumption.
  • New Fed chair (Warsh) is historically considered hawkish; his read of the current 8-4 split is the key near-term unknown. June FOMC will be the first read of how he shapes consensus. The Fed-independence symbolism of a White House swearing-in is itself a market-relevant data point.
  • TSM raised guide above 30% — the AI cycle's upstream rent is not yet rolling over. This data point pushes back against any "AI bubble" call for at least another quarter.

Implications for AlphaSteve

  • PLTR thesis — multiple-compression environment intensifies. Late-cycle + hot CPI + hawkish new Fed chair + a Q1 print that delivered the best growth in company history but couldn't lift the multiple = exactly the macro backdrop where 40×-revenue / 70×-FCF software names underperform on multiple alone. The thesis's variant perception (price requires sustained top-decile decade) is materially reinforced by the May 4 print itself. No change to verdict or trigger ($29); watchlist position confirmed; consider tightening the next-checkpoint date if PLTR continues to roll under.

  • Watchlist additions to consider — critical minerals / rare earths sub-sector is in early thematic phase; if multiple research notes confirm, candidates for Watchlist worth surfacing in the next daily scan. Specifically MP Materials, USA Rare Earth, Lynas (ADR), Energy Fuels.

  • Iran ceasefire as explicit scenario variable — any new valuation work should carry a bull/bear branch tied to whether the ceasefire holds. Mechanically: oil at $80 vs. $110 is a >5% swing in CPI projections and a >100 bps swing in implied 10Y. That is a discount-rate gap that dwarfs most company-specific debates.

  • Daily scan for Tuesday May 26 (Monday holiday) — emphasize: (a) post-Memorial-Day positioning data, (b) refining names that compressed on oil moves last week (VLO, MPC, PSX), (c) AI-infrastructure trough entries if Friday saw selling beneath NVDA's lift, (d) any application-layer software that took asymmetric pain post-NVDA print (PLTR mirror image candidates).

  • Portfolio inception — adjust to Tuesday May 26, not Monday May 25. Update Portfolio mandate accordingly.

  • Discount rates in any new valuation work — use 10Y at 4.56% as baseline; resist temptation to model toward 3.5% on glide-path assumption. The hot-CPI + Iran-ceasefire-binary + 4-way Fed dissent backdrop argues for higher-for-longer base case with explicit branching on the geopolitical binary.

  • Sector tilt for Tuesday build — consider: (a) AI infrastructure upstream (bottleneck rent), but only at non-stretched valuations; (b) critical minerals exposure; (c) defensive duration-light names (financials with rate-cycle benefit, energy E&P benefiting from elevated oil); (d) avoid duration-rich names (long-duration tech, utilities, REITs) at current rate backdrop; (e) explicit hedge or asymmetry around the Iran-ceasefire binary if a clean instrument exists.

  • New pattern flagged for tomorrow's scan: "best-in-company-history print + multiple compression" as a screen — PLTR is the prototype this week. If others surface (likely high-multiple software/AI-application names), they may be earlier-stage PLTR-archetype candidates for the watchlist pipeline.

Cross-references

Sources