Research notes · September 10, 2026
Seven notes, six publishers: Oracle's blowout RPO backlog, a quantum supply-chain map crowning GlobalFoundries, and a case for Bordeaux as a dollar hedge
On 2026-09-10 six publishers put out seven pieces. TMT Breakout filed two intraday wraps covering Oracle and Adobe earnings alongside a Goldman Sachs/Citi tech conference; SemiAnalysis mapped the six operational hurdles behind data-center behind-the-meter power; PhotonCap built a 23-company quantum supply-chain scorecard naming GlobalFoundries the top-ranked name; FundaAI surveyed ten anonymous enterprises on AI spending patterns; James Bulltard posted a personal trading diary on options flow and a technical test of SPY's 50-day average; and Cassandra Unchained argued for holding bonded European fine wine as a dollar hedge. Every statement below is attributed to the publisher that made it, with the publisher's own numbers. Quant Data did not verify any of these claims and takes no view on them.
What was published
| Publisher | Ticker | Title | Points extracted |
|---|---|---|---|
| Cassandra Unchained | macro | Shorting the Dollar by Drinking Free First Growth Bordeaux | 31 |
| FundaAI | macro | Deep|LLM: Enterprise AI Adoption Survey (Vol. 3) – Diverging Spending Outlooks, with Production Use Cases Scaling Alongside More Granular Cost Controls | 23 |
| James Bulltard | — | 9/10/26 Recap | 18 |
| PhotonCap | GFS | [Investment Map] 23 Companies in the Quantum Cycle: From Materials to Logical Qubits | 20 |
| SemiAnalysis | — | What is So Hard About Behind-The-Meter Power For Datacenters? Part 1 | 35 |
| TMT Breakout | ORCL | TMTB EOD Wrap; GS and Citi Day 3 Conference Recaps; Oracle and Adobe first takes | 45 |
| TMT Breakout | — | TMTB Morning Wrap | 30 |
"Points extracted" counts the discrete claims (numbers, theses, risks, catalysts, falsification points) pulled from each piece during ingestion. It measures how much a piece asserts, not how good it is.
What each piece said
Cassandra Unchained — Shorting the Dollar by Drinking Free First Growth Bordeaux
Cassandra Unchained argues that bonded European fine wine — First Growth Bordeaux and Burgundy — functions as a dollar hedge, citing the Liv-ex fine wine index down 25-30% from its October 2022 peak, a 1900-2012 academic study showing 4.1% real annual returns net of costs, and the dollar's central-bank reserve share falling from 72% in 2001 to just under 57% in 2025. The publisher reports a 5-year rolling correlation of -0.74 between the Liv-ex 100 and the dollar index but calls its own method 'more Boy Scout than Army Ranger.'
FundaAI — Deep|LLM: Enterprise AI Adoption Survey (Vol. 3) – Diverging Spending Outlooks, with Production Use Cases Scaling Alongside More Granular Cost Controls
FundaAI's third channel-check round of ten anonymous enterprises finds account coverage has outpaced usage depth, with 10-20% of users generating 50-55% of token volume across multiple firms. Budgets keep expanding but are shifting from unconstrained spending to granular controls: a UK law firm cut monthly overspend from 1.8x to 1.2x of budget via default models, and a European logistics firm cut consumption 20-25% via caching. FundaAI reports one Asian gaming company's paid API spend fell from over $14.3M toward $11.4-12.9M monthly after introducing token quotas and local open-weight models.
James Bulltard — 9/10/26 Recap
James Bulltard's trading diary reports SPY testing its 50-day moving average for the first time since late July, with oil crossing $100/barrel and rising rates after PPI as headwinds; the momentum scanner showed no tech names, only energy and defensives. The author closed short puts in META, NVDA and AMZN, added utility longs NRG and PCG, and flagged options flow in WMB (16,150 $67 puts sold, implying a ~1.6M-share buildup below the 200-day), GOOGL ($355 calls) and FCX ($80 calls) after copper's selloff.
PhotonCap · GFS — [Investment Map] 23 Companies in the Quantum Cycle: From Materials to Logical Qubits
PhotonCap scores 23 quantum-supply-chain companies on a five-axis framework (government backing, mode neutrality, physical bottlenecks, revenue, and photonic convergence) and finds foundry GlobalFoundries ranks first even after removing its favored convergence axis. The publisher reports the four pure-play quantum stocks are down a median 2.2% over the past year while seven component/metrology names are up a median 161%, that only 9 of 23 names show statistically significant 64-day 'quantum beta,' and that Commerce Department grant-letter recipients gave back most of their gains within two days of signing definitive agreements.
SemiAnalysis — What is So Hard About Behind-The-Meter Power For Datacenters? Part 1
SemiAnalysis maps six hurdles facing behind-the-meter (BTM) data-center power — contracts/financing, permitting, gas supply, equipment, labor and physical islanding — and says no single path (gas turbines, fuel cells, or solar-plus-storage) currently clears all six. It reports 75 GW of firm binding BTM orders tracked (about 20 GW added in Q2 2026 alone), a 1 GW off-grid plant costing roughly $5B against inference revenue potentially reaching $100B/GW/year, a projected 288,000-worker US data-center construction labor shortfall by 2027, and gas-turbine order books sold out past 2030.
TMT Breakout · ORCL — TMTB EOD Wrap; GS and Citi Day 3 Conference Recaps; Oracle and Adobe first takes
TMT Breakout's EOD wrap covers Day 3 of the Goldman/Citi tech conferences plus Oracle and Adobe earnings. Oracle's RPO rose $26B in the quarter versus a Street estimate of about $2B, with IaaS growth accelerating to 121% year over year, above the buy-side's 116-118% threshold; Adobe's ARR growth decelerated to 11.2% from 12.5%. The wrap also relays Dell's $61B AI server orders and $95B backlog, Arista's purchase commitments rising to $9.6B through 2027, and a same-day 10-year Treasury yield spike of 7-14bp toward 5% that hit semiconductors broadly.
TMT Breakout — TMTB Morning Wrap
TMT Breakout's morning wrap previews the Citi/GS conference day, noting TSMC's August revenue up 53% year over year (68% of Q3 guidance midpoint), and relays a batch of sell-side moves: JPMorgan raised its META target to $820 on Muse optimism while Oppenheimer countered that only 19% of consumers would pay for an AI agent; JPMorgan initiated SK Hynix at $245 on a five-year memory upcycle; Piper initiated AMD, ARM, AVGO and MRVL all at Overweight; and Oppenheimer raised Google's target to $400, projecting TPU external sales reaching $170B by 2028.
Directional statements on the record
Each row is a dated fact: a named source expressed a direction on a named ticker. Where a publisher was relaying a broker's rating change, the originating voice is in the "via" column. None of this is a recommendation from Quant Data.
| Publisher | Via | Ticker | Stance | Statement |
|---|---|---|---|---|
| TMT Breakout | JPMorgan | META | Bullish | Raised price target to $820 from $640, based on 23x 2028E EPS of $35.44 and optimism on the Muse AI agent. |
| TMT Breakout | Oppenheimer | META | Bearish | Countered that Muse monetization faces a high bar — its survey found only 19% of consumers would pay for an AI agent and 8% would trust Meta with passwords. |
| TMT Breakout | JPMorgan | 000660.KS | Bullish | Initiated SK Hynix at Overweight, price target $245, citing a five-year-plus AI memory upcycle with 2027 EPS CAGR of about 34%. |
| TMT Breakout | Piper | AMD | Bullish | Initiated Overweight/$600 on AMD, Overweight/$320 on ARM, Overweight/$460 on AVGO and Overweight/$270 on MRVL, citing agentic-AI-driven compute demand across custom silicon and CPUs. |
| TMT Breakout | Oppenheimer | GOOGL | Bullish | Maintained Outperform, price target $400, projecting TPU external sales of about $170B cumulative by 2028. |
| TMT Breakout | JPMorgan | SHOP | Bullish | Maintained Overweight, price target $185, arguing the AI-disintermediation selloff (~13% pullback) was overdone since Shopify captures agentic referral traffic. |
| TMT Breakout | Needham | AAPL | Neutral | Maintained Hold, citing an unresolved upgrade cycle and Apple lagging on agentic AI versus Meta's Muse. |
| TMT Breakout | StoneX | XYZ | Bullish | Upgraded Block to Buy from Hold (no target or detailed rationale given in the note). |
| TMT Breakout | Oppenheimer | AEHR | Bullish | Initiated Outperform, price target $120, citing an AI burn-in inflection and an underappreciated second major processor customer ramp. |
| TMT Breakout | Oppenheimer | GLW | Bullish | Raised 3Q26/FY26 estimates and kept Outperform/$200 as a top pick, citing the $20B Springboard fiber run-rate arriving a quarter early. |
| James Bulltard | author's own account | INTC | Bullish | Re-added short puts (selling puts is bullish exposure) at the $80 strike, up from $75, calling it a double-bottom level. |
| James Bulltard | author's own account | NRG | Bullish | Added 3,000 shares at $112.45 and sold covered calls at the $145 December strike, calling NRG 'easily a $150+ name.' |
| James Bulltard | author's own account | WMB | Bullish | Sold puts (selling puts is bullish exposure) at the $67 strike expiring 10/23, following what the author reads as a larger trader's options flow implying a ~1.6M-share buildup below the 200-day average. |
| James Bulltard | author's own account | META | Exited | Closed short puts in META (along with NVDA and AMZN) after the market lost its 21-day EMA, citing a rules-based reduction of tech exposure. |
| James Bulltard | author's own account | PCG | Bullish | Added 5,000 shares at $14.30, expecting a 10-20% swing gain in the utility. |
| James Bulltard | author's own account | GOOGL | Bullish | Holds short puts at the $290 strike (selling puts is bullish exposure) expiring in December, reading $355-strike call buying (7,000 contracts, 10/2 expiry) as a bet on a sharp Google rebound. |
Numbers the publishers put on the record
- TMT Breakout · ORCL: RPO rose $26B in the quarter vs a Street estimate of about $2B; RPO reached $664B; IaaS growth accelerated to 121% y/y (120% cc), above the buy-side's 116-118% threshold.
- TMT Breakout · ADBE: ARR growth decelerated to 11.2% y/y from 12.5% prior quarter; ARR reached $27.5B; shares fell 2.5%.
- TMT Breakout: 10-year Treasury yield rose 7-14bp toward 5%, 30-year toward 5.4% (near 2007 highs); semiconductors fell 2.6% on the day.
- TMT Breakout · TSM: August revenue NT$514.8B, up 53% y/y, equal to 68% of the Q3 guidance midpoint.
- TMT Breakout · META: JPMorgan target raised to $820 from $640; Oppenheimer survey found only 19% of consumers would pay for an AI agent and 8% would trust Meta with passwords.
- SemiAnalysis: 75 GW of firm binding behind-the-meter orders tracked (about 20 GW added in Q2 2026); a 1 GW off-grid plant costs about $5B against potential inference revenue of $100B/GW/year.
- SemiAnalysis: Projected 2027 US data-center construction labor shortfall of 288,000 workers; only 10,200 boilermakers nationwide, shrinking by about 800/year.
- PhotonCap · GFS: GlobalFoundries scores 27.5/30 (top rank) on PhotonCap's GVM framework despite a statistically insignificant 64-day quantum beta of 0.12 and a 37% one-year stock gain.
- PhotonCap: Four pure-play quantum stocks down a median 2.2% over the past year vs seven component/metrology names up a median 161%; only 9 of 23 companies show statistically significant quantum beta.
- FundaAI: Across a 10-company survey, 10-20% of AI users generate 50-55% of total token volume; a UK law firm cut monthly overspend from 1.8x to 1.2x of budget via default-model routing.
- Cassandra Unchained: Liv-ex fine wine index down 25-30% from its October 2022 peak; the dollar's central-bank reserve share fell from 72% in 2001 to just under 57% in 2025Q3.
- James Bulltard · WMB: 16,150 contracts of the $67 put (10/23 expiry) sold, which the author reads as implying a buildup of about 1.6M shares below the 200-day average.
- SemiAnalysis: CCGT construction cost rose from about $750/kW in 2018 to about $2,000/kW currently, with close to half of the increase flowing to EPC firms.
- TMT Breakout · GOOGL: Oppenheimer projects TPU external sales reaching about $170B cumulative by 2028, with annual TPU revenue rising from about $19B (2026) to about $101B (2028).
Reproduced as stated by each publisher. Not verified by Quant Data.
What would prove them wrong
Several publishers name the observation that would falsify their own view. Those are the most useful sentences in any research note, so they are collected here.
- TMT Breakout · ORCL: Whether RPO net adds in coming quarters keep pace with this quarter's $26B, and whether the 850MW capacity build-out converts to revenue as scheduled.
- TMT Breakout · LITE: Whether a second high-power (120-150mW+) laser source supplier (e.g., Coherent) emerges before 2H27; management says it does not know.
- TMT Breakout · HPE: Whether Cloud & AI segment margin holds the FY27 guide of 13%, down from 17% in Q3.
- TMT Breakout · META: Muse subscriber/revenue disclosure will show whether JPMorgan's implied ~115M paying-subscriber assumption or Oppenheimer's low-willingness-to-pay survey is closer to reality.
- TMT Breakout · 000660.KS: Whether SK Hynix sustains a >50% FCF payout ratio and the ~34% 2027 EPS CAGR through 2026-2028 quarterly results.
- TMT Breakout · TSM: Whether September and Q3 actual revenue land at the high end of the $44.6-45.8B guide.
- SemiAnalysis: Whether Nexus Hubbard's 150 MMcf/d gas contract (enough for ~750MW) is expanded to match its 5,230MW pending permit, or whether that permit turns out to be a placeholder.
- SemiAnalysis: Oracle's Project Jupiter permitting ruling is due by November 23, 2026, and the Transwestern Green Chile pipeline needs FERC approval by around October 2026 to hit its revised 2027 gas-supply timeline.
- SemiAnalysis: Whether most behind-the-meter plants eventually connect to the grid (bridge power) or expand indefinitely off-grid (a standalone macrogrid) over the next 1-2 years.
- PhotonCap: Whether the remaining six Commerce Department LOIs convert to definitive agreements in 2026Q4-2027Q1; failure would undercut the government-backing axis for Group B names.
- PhotonCap: Whether Tier 1-4 component/foundry companies ever break out standalone quantum revenue; until they do, the convergence/neutrality scores measure relationships, not dollars.
- PhotonCap · COHR: If Coherent discloses a named quantum-customer supply agreement, its convergence score would rise, pushing it into the top two of PhotonCap's ranking.
- FundaAI: If the share of surveyed companies reporting an absolute decline in paid token spend (from caching/routing/local models) rises from 1-in-10 toward 3-in-10 or more, model-vendor token revenue assumptions would need reworking.
- Cassandra Unchained: If the 5-year rolling Liv-ex/dollar-index correlation reverts toward zero and the dollar's reserve share stabilizes or rises, the case for wine as a currency hedge weakens.
- James Bulltard · SPY: Author says a close below the 50-day (with the 100-day about 1% lower as support) would confirm technical breakdown; a reappearance of tech names atop the momentum scanner would falsify the 'market is breaking down' read.
Where Quant Data's own data touches the same tickers
For tickers named above that have a stock page on this site, the settled options positioning from the same period. These are Quant Data's computations from end-of-day open interest, dated to the settlement shown, and are unrelated to what any publisher said.
| Ticker | Settled | Front expiry | Max pain | Last price | Put/call OI |
|---|---|---|---|---|---|
| AAPL | 2026-09-17 | 2026-09-21 | $330.00 | $336.29 | 0.53 |
| AMD | 2026-09-17 | 2026-09-21 | $530.00 | $609.27 | 1.23 |
| AVGO | 2026-09-17 | 2026-09-21 | $347.50 | $356.91 | 0.50 |
| GOOGL | 2026-09-17 | 2026-09-21 | $345.00 | $353.47 | 0.53 |
| INTC | 2026-09-17 | 2026-09-21 | $109.00 | $118.40 | 1.04 |
| META | 2026-09-17 | 2026-09-21 | $665.00 | $714.93 | 0.77 |
| NVDA | 2026-09-17 | 2026-09-21 | $215.00 | $222.27 | 0.89 |
Get these notes by email
Same text as this page, each publishing day. Free.
Common questions
What did paid research publishers publish on 2026-09-10?
7 pieces from 6 of the publishers tracked by Quant Data: Cassandra Unchained, FundaAI, James Bulltard, PhotonCap, SemiAnalysis, TMT Breakout. Seven notes, six publishers: Oracle's blowout RPO backlog, a quantum supply-chain map crowning GlobalFoundries, and a case for Bordeaux as a dollar hedge.
Are these Quant Data’s numbers?
No. Every figure on this page is attributed to the publisher that stated it and is reproduced as stated. Quant Data did not verify them. The only Quant Data numbers on the page are the counts of pieces and statements, and the settled options positioning shown for tickers that have a max pain page on this site.
Is a "bullish" entry in the statements table a recommendation?
No. The table records that a named source expressed a direction on a named ticker on a dated day, the way a clippings file would. It is not a signal and not advice. It exists so that the statements can be checked against what prices did afterwards.
Sources: paid research publishers accessed through the GeoScope aggregator under a licence that permits redistribution. Statements and figures are attributed to their publishers and reproduced as stated; Quant Data has not verified them and expresses no view on them. Educational only: not investment advice, not a recommendation, not a forecast.