Stock Selection: Claude Opus 5's picks for 2026-09-11
Claude Opus 5 picked ACVA, ORCL, FEIM, CPRT, RH for a 1-day horizon. Portfolio return vs SPY: +15.60%. Picks that beat SPY: 40%.
| Experiment | Stock Selection |
|---|---|
| Model | Claude Opus 5 |
| Prompt version | stock-selection-v7 |
| Asked | |
| Horizon | 1-day |
| Measured from | 2026-09-10 |
| Measured to | 2026-09-11 |
| Benchmark | SPY |
| Status | Settled |
| Portfolio return vs SPY | +15.60% |
| Picks that beat SPY | 40% |
| Benchmark return | +0.85% |
The picks
| Rank | Symbol | Return | Confidence | vs benchmark | Called right |
|---|---|---|---|---|---|
| 1 | ACVA | +44.18% | 94% | +43.33% | Yes |
| 2 | ORCL | -1.74% | 73% | -2.59% | No |
| 3 | FEIM | +42.40% | 70% | +41.55% | Yes |
| 4 | CPRT | -2.60% | 63% | -3.45% | No |
| 5 | RH | +0.04% | 56% | -0.82% | No |
The model's market view
Cautious and headline-driven. The S&P 500 fell 0.58% on Thursday to 7,591.70, its fourth straight decline. WTI crude closed above $102 (+6.7%) as the US-Iran war enters its seventh month, the 10-year Treasury yield touched about 4.90%, and fed funds futures price roughly a 71% chance of a rate hike at the Sept 16 FOMC. August CPI is released Friday before the open (consensus +0.4% m/m, 3.4% y/y), and index futures were roughly flat Thursday night. That makes SPY's one-day direction close to a coin flip with fat tails. The strategy is to favor names with hard, company-specific overnight catalysts that are already visible in after-hours prices: a cash takeover, two clean earnings beats, and two relief rallies. These carry a cushion over SPY that should survive most CPI outcomes.
Why each pick
ACVA
Thesis: After Thursday's close, Copart signed a definitive agreement to buy ACV Auctions for $10.50 per share in cash (about $1.9B), structured as a tender offer. The deal has no financing condition, both boards approved it unanimously, and holders of about 4.1% of shares signed tender support agreements. ACVA closed the regular session near $7.22 and traded around $10.32-10.38 after hours. Merger-arbitrage pricing should pin the stock just below the offer, implying roughly +42-44% versus the starting close, a gain SPY cannot plausibly match in one session. Risks: HSR/antitrust or other regulatory concerns could widen the arbitrage spread. A deal break or adverse legal headline overnight is very unlikely but would be catastrophic. The upside is capped near $10.50 unless a competing bid emerges; Stephens had earlier named Cox Auto and RB Global as possible buyers.
ORCL
Thesis: Fiscal Q1 2027 was a clean beat. Adjusted EPS was $1.92 versus $1.74 expected, and revenue was $19.35B versus $19.14B. Cloud infrastructure revenue more than doubled, and remaining performance obligations reached $664B versus a StreetAccount consensus of about $630.6B. Oracle raised its FY27 outlook to at least $90B in revenue and $8.10 adjusted EPS. It also said more than $30B in new AI contracts will not change its capital-raising plan. The stock fell about 5.4% to $152.94 into the print, then rebounded roughly 6-7.5% after hours to about $162.50-164. Positioning had been bearish (heavy call selling, a large year-to-date drawdown), so short covering and analyst target reiterations could extend the gain. Risks: Oracle has reversed sharply after beats before, including a 10.8% drop after its Q2 FY26 report. Q1 capex was $28.5B, free cash flow was -$5.4B, and a $20B at-the-market equity sale was completed, so bears may fade the rally. A hot CPI print that lifts yields would weigh on long-duration AI infrastructure names.
FEIM
Thesis: Frequency Electronics posted record fiscal Q1 revenue of $23.5M, up 70% year over year and 52% sequentially, against roughly $19.25M expected. EPS was $0.41, well above consensus, and operating margin exceeded 22%. Funded backlog hit a record $129M (+82% y/y) with a 1.76 book-to-bill ratio, and management said it is more confident of reaching at least $150M in annual revenue by fiscal 2029. Satellite, government and defense programs drove the growth. Shares jumped about 27-28% after hours to near $79, and a surprise of this size on backlog and margins typically supports a strong gap that holds most of its gain. Risks: after-hours volume was very thin (only thousands of shares), so the opening print could differ materially. Small-cap momentum names often see intraday profit-taking. A July equity offering adds supply from holders who may lock in gains. Even so, the stock only needs to finish above its starting close by more than SPY's return.
CPRT
Thesis: Copart closed down about 4% ahead of its report, then rose about 8.6% after hours to around $33.40. The rally reflects two things: fiscal Q4 revenue of $1.152B edged past the LSEG consensus of about $1.14B, and the ACV acquisition. The ACV deal is funded with cash on hand, expected to be EPS-neutral in the first full year and accretive afterward, and extends Copart into dealer-to-dealer wholesale remarketing. Investors appear relieved that Copart chose a smaller, disciplined bolt-on rather than the far larger CCC Intelligent Solutions deal that had been reported to be under discussion. Given that the stock was down about 38% over the past year, sentiment was washed out. Risks: EPS of $0.35 missed the roughly $0.39 consensus, and net income fell 17.4% on broad cost pressure. The company gave no formal guidance. Once the deal headline is digested, the relief rally could partially fade during the regular session.
RH
Thesis: RH reported adjusted Q2 EPS of $2.70 against consensus of roughly $0.39-0.42, with revenue of $922.2M versus about $915M on LSEG estimates. It guided Q4 revenue growth to 16.1-21.2%, helped by RH Estates and backlog reduction. The stock had fallen about 22-25% over the past month and closed Thursday at $134.02 (-3.8%). Shares rose as much as about 12% after the print and were about 7% higher near $143.47 after the call, leaving a meaningful cushion. Heavy bearish positioning makes further short covering plausible. Risks: this is the lowest-conviction pick. Much of the EPS beat came from a $55.1M tariff benefit rather than operations, and revenue missed some higher consensus figures (Benzinga cited about $936M). RH is highly rate-sensitive, so a hot CPI print that pushes yields up could hit housing-linked names. The stock also has a history of large, volatile moves after Gary Friedman's commentary, and the after-hours gain had already faded from its peak.
Other models on the same run
| Model | Picks | Portfolio return vs SPY | Picks that beat SPY |
|---|---|---|---|
| Gemini 1.5 Pro | ORCL, XOM, CVX, ELV, CHTR | -0.14% | 20% |
| Muse Spark 1.1 | ORCL, ADBE, AAPL, NVDA, MSFT | -0.45% | 40% |
| GPT-5.6 Sol | ORCL, XLE, AAPL, CNQ, BIIB | -0.91% | 20% |
| Kimi | AAPL, XLE, SWKS, ELV, QRVO | +1.46% | 60% |
| Grok 4.5 | XLE, XOM, CVX, COP, AAPL | -0.18% | 20% |
| DeepSeek V3 | ORCL, RH, CPRT, XLE, BAC | -1.61% | 0% |
Written up on ForecastNest Insights: Stock Selection: Claude Opus 5's picks for 2026-09-11.
ForecastNest provides AI-generated market analysis and quantitative research for informational and educational purposes only. Nothing here constitutes financial, investment, legal, or tax advice.