Extreme Movers: Claude Opus 5's picks for 2026-09-17
Claude Opus 5 picked INTC, AMD, TSM, GS, C for a 1-day horizon. Average directional return: +0.18%. Directions called correctly: 40%.
| Experiment | Extreme Movers |
|---|---|
| Model | Claude Opus 5 |
| Prompt version | extreme-movers-v8 |
| Asked | |
| Horizon | 1-day |
| Measured from | 2026-09-17 |
| Measured to | 2026-09-18 |
| Benchmark | SPY |
| Status | Settled |
| Average directional return | +0.18% |
| Directions called correctly | 40% |
| Benchmark return | -0.18% |
The picks
| Rank | Symbol | Direction | Price change | Return in the called direction | Confidence | vs benchmark | Called right |
|---|---|---|---|---|---|---|---|
| 1 | INTC | up | -0.34% | -0.34% | 58% | -0.16% | No |
| 2 | AMD | up | +2.56% | +2.56% | 54% | +2.74% | Yes |
| 3 | TSM | up | +1.56% | +1.56% | 52% | +1.74% | Yes |
| 4 | GS | down | +1.89% | -1.89% | 51% | -1.71% | No |
| 5 | C | down | +0.98% | -0.98% | 49% | -0.80% | No |
The model's market view
Source session is Wednesday 16 September 2026, FOMC day, and it produced an unusually clean split rather than a broad move. The Fed raised rates 25 basis points to 3.75%-4.00% in a unanimous 12-0 vote, the first hike since 2023, and the dot plot was firmly hawkish: 16 of 18 officials see at least one more hike this year, four pencilled in two more, the median stays near 4.1% through end-2027 with no cuts in 2027, and the return to 2% core inflation is pushed to 2029 (vantagemarkets.com, fool.com). The index damage was concentrated, not general: the Dow fell 631.21 points or 1.21% to 51,461.90 with Goldman Sachs leading it lower, the S&P 500 dropped 0.45% to 7,551.81, and the Nasdaq Composite finished essentially flat at 25,978.42, down 0.01%. Banks took the brunt, with the SPDR S&P Bank ETF down 2.6% for its worst session since 27 February, and Goldman Sachs, Wells Fargo, Bank of America and Citigroup each falling more than 3% while JPMorgan lost 1.5% (cnbc.com/2026/09/15/stock-market-today-live-updates.html). Semiconductors went the other way on a genuinely company-specific catalyst: Reuters reported SK Hynix is in talks to manufacture memory chips in the US for the first time, possibly leasing part of Intel's $100 billion Ohio campus, and SOXX gained 2%, tripling QQQ's advance (247wallst.com, cnbc.com/2026/09/16/intel-sk-hynix-us-memory-chips.html). My structural read is that both moves are substantive rather than speculative and both should persist for more than one session. The Intel catalyst is a reported commercial negotiation reinforced by multiple analyst target raises, not sentiment; the bank selloff reflects a repricing of the policy path itself, which does not reverse because a day has passed. Two honesty notes. First, I verified prices and percentages for the three semiconductor picks but could only verify that the banks fell more than 3% without an exact figure or a refreshable quote, so I report what the source actually says and omit what I could not confirm. Second, my freshest usable quotes for ranks 1 to 3 are Wednesday closes, roughly eighteen hours stale; the one intraday datapoint I have for today is a percentage move, not a dollar price, which I describe rather than convert and present as a quote.
Why each pick
INTC
Ranked first because it is the only pick with a substantive catalyst and confirmed follow-through already visible today. Intel closed at $101.05 on 16 September, up 4.03%, on 117.5 million shares against a three-month average of 109.5 million, after Reuters reported talks with SK Hynix to manufacture memory chips in the US for the first time, potentially by leasing part of Intel's planned Ohio production site or through a joint venture with cloud firms. This is a reported commercial negotiation, not a sentiment story, and it was reinforced the same day by Melius Research reiterating Buy with a $165 target and Tigress Financial lifting its target to $145 from $118, plus an Altera IPO filing that crystallises asset value. A stock market roundup timestamped 07:47:53 EDT on 17 September reports Intel trending up a further 3.22%, so the move is extending rather than fading. Risks: the talks are explicitly exploratory with no decisions made, South Korea's trade ministry told Reuters any deal involving national core technology is subject to review under its Industrial Technology Protection Act, and the Ohio plants are not due until 2030, so nothing here affects near-term earnings. The stock is up 171% year to date, which makes it vulnerable to profit-taking. Price note: $101.05 is Wednesday's regular-session close, about eighteen hours stale; today's move is reported to me as a percentage rather than a dollar quote, so I have not converted it. Market cap is an estimate derived from price and approximate share count.
AMD
Ranked second as the broader semiconductor continuation. AMD closed at $512.50 on 16 September, up 1.65%, participating in a session where SOXX gained 2% and tripled QQQ's advance while the S&P 500 fell 0.45%. The relative-strength signal is the point: on a day when a hawkish Fed knocked 1.21% off the Dow, the semiconductor complex rose, which says the market is treating this cohort as the place to be in a higher-for-longer regime rather than as the first thing to sell. AMD also carries its own recovery narrative from Monday's AI-pacing selloff, which was driven by an opinion essay on the pace of frontier model development rather than by any change to orders or guidance, and which President Trump publicly dismissed on competitiveness grounds. Risks: AMD has roughly doubled in 2026, so the valuation offers no cushion if the long end pushes further above 5%, and semiconductors are the longest-duration equity exposure in the market. Morgan Stanley has flagged weakening DRAM demand with building inventory and NAND oversupply, and a 1.65% gain is a modest signal to extrapolate from. There is no scheduled AMD catalyst inside the window. Price note: this is Wednesday's regular-session close, roughly eighteen hours stale; I could not obtain a fresher intraday quote. Market cap is an estimate.
TSM
Ranked third as the lowest-variance way to express the same continuation. Taiwan Semiconductor closed at $417.72 on 16 September, up 1.23%, alongside the broader foundry and chip advance. The specific logic for including it rather than a fourth high-beta name is that the SK Hynix story is fundamentally about where advanced manufacturing capacity gets built, and a wave of investor attention on domestic foundry economics tends to lift the incumbent leader alongside the challenger, even when the headline names a competitor. TSM is also the largest and most liquid pure foundry exposure available in US listings, which damps the idiosyncratic risk that dominates ranks 1 and 2. Risks: a 1.23% gain is the smallest move among my three continuation picks, so there is the least momentum to extend, and the reported Intel-SK Hynix arrangement is arguably competitive with TSM's US expansion rather than supportive of it. Geopolitical headline risk around Taiwan is unforecastable over a single session and is the tail that matters most here. As a roughly $2 trillion name it will also track the index closely, so the likeliest failure mode is simply not moving enough to score. Price note: Wednesday's regular-session close, roughly eighteen hours stale. Market cap is an estimate derived from the quoted price and approximate share count.
GS
Ranked fourth as the cleanest short side of Wednesday's split. CNBC reports Goldman Sachs led the Dow's 631-point decline and fell more than 3%, with the SPDR S&P Bank ETF shedding 2.6% for its worst session since 27 February. The continuation case is that this was a repricing of the policy regime, not a one-day reaction: the dot plot now projects no cuts in 2027, a median rate near 4.1% through next year and inflation above target until 2029, which raises funding costs and credit provisioning assumptions across a multi-quarter horizon. The dollar index also popped 0.6% to 100.21, its highest since 31 July, which tightens conditions further. Regime repricings of this kind typically bleed over several sessions as estimates are revised, unlike headline shocks that mean-revert. Risks: real and specific. A first hike with a steepening curve is the textbook net-interest-margin trade, and if the long end holds above 5% while the front end stabilises, banks are the natural beneficiaries and this leg inverts. Goldman is also more capital-markets-driven than spread-driven, so higher volatility can help its trading revenue. Honesty note: the source states only that the stock fell more than 3%, so -3 is a floor rather than the exact figure, and I could not verify a current quote or a market capitalisation, so both are omitted rather than invented.
C
Ranked last and carrying the thinnest verified data, which I want to state before the thesis. CNBC names Citigroup among the banks that fell more than 3% on 16 September, alongside Goldman Sachs, Wells Fargo and Bank of America, while JPMorgan lost only 1.5%. That dispersion within the group is the reason for picking Citigroup rather than JPMorgan: the market discriminated, selling the names with weaker perceived capital positions and more consumer credit exposure hardest, and a 3%-plus decline against JPMorgan's 1.5% is a relative-quality verdict that tends to persist while the rate path is being repriced. With the Fed projecting inflation above target until 2029 and no cuts in 2027, the borrower stress channel runs for quarters, not hours. Mast Investments CIO Yung-Shin Kung has argued the hike disproportionately strains the bottom of a K-shaped economy by raising debt-servicing costs without relief, which is a direct consumer-credit threat. Risks: heavily discounted bank stocks are prone to sharp mean-reversion bounces after a capitulation day, and the entire sector could rally if the curve steepens rather than flattens. Honesty note: as with rank 4, -3 is the floor the source states rather than an exact figure, and I could not verify a current price or market cap, so both are omitted.
Other models on the same run
| Model | Picks | Average directional return | Directions called correctly |
|---|---|---|---|
| GPT-5.6 Sol | DLXY, MEDS, XBIO, AEHL, CUPR | +3.65% | 60% |
| Muse Spark 1.1 | WLTH, SWKS, ANGX, UNFI, SUNE | -1.76% | 40% |
| DeepSeek V3 | GNRC, FLNC, DLXY, MEDS, NBIS | +6.96% | 60% |
| Gemini 1.5 Pro | SMTC, SABR, CIFR, JBHT, ALHC | -1.62% | 20% |
Written up on ForecastNest Insights: Extreme Movers: Claude Opus 5's picks for 2026-09-17.
ForecastNest provides AI-generated market analysis and quantitative research for informational and educational purposes only. Nothing here constitutes financial, investment, legal, or tax advice.