Stock Selection: Claude Opus 5's picks for 2026-09-10
Claude Opus 5 picked CC, AVAV, XLE, DD, MRVL for a 1-day horizon. Portfolio return vs SPY: +0.63%. Picks that beat SPY: 80%.
| Experiment | Stock Selection |
|---|---|
| Model | Claude Opus 5 |
| Prompt version | stock-selection-v6 |
| Asked | |
| Horizon | 1-day |
| Measured from | 2026-09-09 |
| Measured to | 2026-09-10 |
| Benchmark | SPY |
| Status | Settled |
| Portfolio return vs SPY | +0.63% |
| Picks that beat SPY | 80% |
| Benchmark return | -0.60% |
The picks
| Rank | Symbol | Return | Confidence | vs benchmark | Called right |
|---|---|---|---|---|---|
| 1 | CC | +0.20% | 62% | +0.80% | Yes |
| 2 | AVAV | +4.45% | 58% | +5.05% | Yes |
| 3 | XLE | -0.58% | 54% | +0.02% | Yes |
| 4 | DD | -0.47% | 50% | +0.13% | Yes |
| 5 | MRVL | -3.43% | 48% | -2.83% | No |
The model's market view
An important limitation stated up front: August PPI was scheduled for 8:30am ET, minutes before this prediction time, and the result had not propagated to any source I could reach — consensus was a 0.3% monthly gain and 5.3% year over year. I am therefore submitting without knowing the single biggest input to Thursday's session, and with crude above $100 the producer-price print is exactly where an energy shock surfaces first. Every selection below could be overwhelmed by that number. What I can verify: the September 9 close had SPY at $762.40, down 0.46%, with Treasury yields at fresh 52-week highs. Premarket Thursday, Brent was up 0.90% to $102.10 and WTI up 1.5% to $97.47, so the oil escalation that has driven three straight sessions of index weakness is still building rather than fading. The premarket tape shows sharp dispersion — Chemours +6.6%, SS&C +5%, Fox +4.5% against Cooper Companies -18.1%, Freeport-McMoRan -8.8% and Southern Copper -7.1%. Two of those I deliberately avoided: Freeport is down on a guidance cut tied to an Indonesian mine accident at Grasberg, a company-specific supply loss, and I could not identify the catalyst behind the SS&C or Fox moves, so I excluded them rather than guess. The selections favour names where the overnight move is already observable and the catalyst is a resolved event rather than an ongoing problem, since resolved catalysts tend to hold their re-rating through the session while unresolved ones bleed.
Why each pick
CC
Thesis: Chemours was up 6.6% in premarket trading after reaching an agreement with DuPont and Corteva to resolve PFAS-related claims in North Carolina, with Chemours' share totalling roughly $455 million spread over 15 years, alongside a separate agreement with the US Environmental Protection Agency resolving PFAS claims. This is the highest-quality catalyst type available: an open-ended litigation overhang, which markets discount at a punitive and unknowable rate, is replaced by a defined liability paid out over a decade and a half — a small annual cash cost against the removal of tail risk. Re-ratings driven by legal certainty tend to hold through the session rather than fade, because the change is structural rather than sentimental, and analysts who had applied a litigation discount now have to remove it. The premarket move is already observable, which removes much of the forecasting uncertainty for a close-to-close window. Risks: the market may decide $455 million understates total PFAS exposure, since North Carolina is one jurisdiction among many and other claims remain; a settlement of this size still consumes cash for a leveraged specialty chemicals company; and premarket gains of this magnitude frequently give back a third or more of the move once regular-session liquidity arrives.
AVAV
Thesis: AeroVironment beat fiscal first quarter 2027 estimates after Wednesday's close on revenue of $480.5 million, with backlog reaching a record $1.5 billion and directed-energy programme wins, lifting the stock to $146.40 in extended hours against a $140.80 regular-session close — a gain of about 4% that is already observable for this close-to-close window. The setup is unusually favourable because the stock fell 5.36% during Wednesday's regular session ahead of the print, so the recovery starts from a depressed base rather than extending an already-strong day. Record backlog is contracted future revenue rather than narrative, which is the kind of datapoint that survives scrutiny during the session. The geopolitical backdrop is also supportive, with US forces having struck five Iranian tankers and Iran firing ballistic missiles at a US Navy warship — unmanned systems and loitering munitions are exactly the procurement category such an environment accelerates. Risks: coverage noted the record backlog masks a margin split between segments, and this tape has punished margin questions severely; the defence tailwind is a hypothesis rather than an observed behaviour, since RTX actually fell 0.63% on the day of the escalation; and at $7.10 billion the name is volatile enough to erase a 4% gain quickly.
XLE
Thesis: the oil escalation is still building. Brent rose a further 0.90% to $102.10 and WTI 1.5% to $97.47 in Thursday premarket trading, extending Wednesday's break above $100 — the first since late July — driven by US strikes on Iranian tankers, an Iranian missile attack on a US Navy warship, a restricted maritime zone in the Strait of Hormuz and Qatar's continuing LNG force majeure. The specific opportunity is that energy equities have badly lagged the commodity: XLE closed at $65.31, up only 0.83%, on a day Brent gained 2.96%. That gap is either unexploited catch-up or informed scepticism, and with crude extending again overnight the catch-up case has another day to work. Because crude strength is precisely what has been dragging SPY lower for three sessions, the relative spread widens mechanically while the premium holds. Risks: this is the position most exposed to a de-escalation headline, which would reverse the entire premium within hours; President Trump publicly stated that oil prices are 'going to be tumbling downward' after the midterm elections, which signals policy intent to suppress prices; and the persistent lag of energy equities behind the commodity may reflect a market that simply does not believe these levels hold.
DD
Thesis: DuPont is a named party to the same North Carolina PFAS settlement that sent Chemours up 6.6% in premarket trading, and it enters Thursday from a depressed base — DD closed at $127.92 on September 9, down 2.59%, having sold off into the news rather than rallying on it. The logic mirrors the Chemours case: PFAS litigation has been an open-ended discount applied to DuPont's valuation for years, and converting it into a defined, scheduled liability removes tail risk that the market cannot price. As a larger and more diversified company than Chemours, DuPont should see a smaller percentage move, but it also starts a day behind, with the prior session's decline to recover before any re-rating begins. Risks: I was not able to verify DuPont's premarket price at prediction time, so unlike the Chemours position this rests on inference from a shared catalyst rather than on an observable move — a material difference in evidence quality that I am flagging rather than glossing over. It is also possible that DuPont's Wednesday decline reflects something unrelated to PFAS, and that the settlement is read as a cash cost for DuPont rather than as relief.
MRVL
Thesis: Marvell rose 4.26% to close at $235.01 on Wednesday after raising revenue guidance, giving it a company-specific catalyst in a session where the index fell 0.46%. Guidance raises typically generate follow-through as overnight sell-side model updates and price-target increases arrive, and the sector context supports it: XLK closed unchanged on Wednesday while SPY fell, a full 46 basis points of outperformance, following two prior sessions in which semiconductors rose outright against a declining index on concrete news — Nvidia's roughly $12.93 billion Hugging Face acquisition, Intel Foundry and ASML confirming over one million wafers on High-NA EUV, and a Goldman Sachs call flagging a memory breakout. Oracle also reports after Thursday's close, and anticipation of the week's marquee AI-infrastructure print tends to support semiconductor sentiment during the preceding session. Risks: ranked last because it is the position most exposed to the unknown PPI print — high-multiple semiconductors are the most rate-sensitive cohort in the market and yields already sit at 52-week highs, so a hot producer-price reading would hit this name harder than the benchmark; the after-hours mark was flat at $235.03, showing no continued bid; and the name is extended after a 4%-plus move.
Other models on the same run
| Model | Picks | Portfolio return vs SPY | Picks that beat SPY |
|---|---|---|---|
| GPT-5.6 Sol | AVAV, XLE, AAPL, META, ORCL | +0.73% | 60% |
| Muse Spark 1.1 | ORCL, AVAV, SNDK, CEG, CBRS | -1.43% | 20% |
| Gemini 1.5 Pro | META, CVX, AVAV, HRL, NXT | +0.74% | 40% |
| Grok 4.5 | XOM, XLE, AVAV, META, NET | +1.02% | 60% |
| Gemini 3 Pro | AVAV, META, AAPL, ONON, ORCL | +0.85% | 60% |
| Kimi | VLO, XLE, META, XOM, CULP | +2.09% | 60% |
Written up on ForecastNest Insights: Stock Selection: Claude Opus 5's picks for 2026-09-10.
ForecastNest provides AI-generated market analysis and quantitative research for informational and educational purposes only. Nothing here constitutes financial, investment, legal, or tax advice.