Stock Selection: DeepSeek's picks for 2026-10-23
DeepSeek picked RNG, RHI, CCI, AMT, SKHY for a 1-day horizon. Portfolio return vs SPY: +0.59%. Picks that beat SPY: 60%.
| Experiment | Stock Selection |
|---|---|
| Model | DeepSeek |
| Prompt version | stock-selection-v8 |
| Asked | |
| Horizon | 1-day |
| Measured from | 2026-10-23 |
| Measured to | 2026-10-26 |
| Benchmark | SPY |
| Status | Settled |
| Portfolio return vs SPY | +0.59% |
| Picks that beat SPY | 60% |
| Benchmark return | +0.67% |
The picks
| Rank | Symbol | Return | Confidence | vs benchmark | Called right |
|---|---|---|---|---|---|
| 1 | RNG | -0.31% | 72% | -0.98% | No |
| 2 | RHI | +1.50% | 70% | +0.83% | Yes |
| 3 | CCI | +2.21% | 68% | +1.55% | Yes |
| 4 | AMT | -0.18% | 66% | -0.84% | No |
| 5 | SKHY | +3.03% | 62% | +2.37% | Yes |
The model's market view
The October 23 session is shaping up as a stock-picker's tape rather than an index trade: SPY is roughly flat to slightly lower while single-name catalysts are producing double-digit moves. Two clusters dominate the premarket and early intraday action: earnings beats with raised guidance (RNG, RHI) and the SpaceX cell-site story lifting tower REITs (CCI, AMT). A third, quieter bid is in AI memory (SKHY). The main risk to all five is that a weak broader tape compresses gains or that crowded post-catalyst moves see profit-taking into the close. I weight catalyst specificity and early-session volume confirmation over raw percentage gain, and I avoid names where the catalyst is a pure short-squeeze or where the premarket move has already fully faded.
Why each pick
RNG
RingCentral reported Q3 revenue of $676M (+5.9% y/y) and non-GAAP EPS of $1.33 versus the $1.27 consensus, then raised FY26 revenue guidance to $2.65–$2.657B and adjusted EPS to $5.07–$5.12. The guidance raise above market estimates is the key driver because it forces an immediate lift in forward EPS expectations, and the stock had closed down 3.64% on October 22 before the release, leaving room for a gap-and-go. The latest intraday quote I can verify is $81.30 at 10:46 AM ET on October 23, up roughly 2.4% from the prior close, which means the stock has given back a large part of the premarket 16% jump. That pullback is the main risk: GF Value pegs intrinsic value near $39.60, about half the current price, and insider selling has been persistent. The catalyst is substantive but already partially reflected; the remaining edge is that sell-side estimates still need to catch up to the raised guide. Invalidation for the long thesis is a break below the October 22 close of $79.42, which would signal the gap is failing. I used an intraday regular-session quote, not a premarket print.
RHI
Robert Half beat Q3 EPS with $0.49 versus the $0.47 consensus, and the quarter contained the first year-over-year revenue growth in its Talent Solutions segment since 2022, up 2%. Operating income jumped to $74.8M from $13.6M a year earlier, showing significant operating leverage, and management guided Q4 revenue about 3% above the prior-year period. The catalyst is substantive because it changes the near-term earnings trajectory, not just sentiment. BMO lifted its target to $50 and Baird to $53, both Outperform, which supports estimate revisions in the coming days. The latest intraday quote I can verify is $43.28 at 10:35 AM ET on October 23, up about 14.9% from the $37.68 prior close, so the market has already repriced a large part of the beat. The main risk is that the stock is now well above its 50-day average near $38.75 and could see profit-taking after a gap of this size; the premarket indication was around $41.50, and the intraday extension to $43+ shows momentum but also crowded positioning. Invalidation is a close back below the October 22 close of $37.68, which would fill the earnings gap. I used an intraday regular-session quote.
CCI
Crown Castle is the most direct liquid beneficiary of the Bloomberg report that SpaceX is in talks to secure access to roughly 80,000 U.S. cell sites, with Crown Castle named as one of the tower operators in discussions. The catalyst is event-driven and speculative in the sense that no definitive agreement has been signed, but it is substantive because a SpaceX national wireless buildout would materially increase tenant demand across Crown Castle's tower portfolio over time. The stock closed at $78.83 on October 22 and traded as high as $90.71 intraday on October 23, a gain of roughly 15%, making it one of the strongest large-cap movers in the session. The latest intraday quote I can verify is $90.50, which means the move is holding near the high rather than fading. The main risk is headline reversal: if the SpaceX talks stall or another tower operator is chosen first, the event premium could compress quickly. A secondary risk is that REITs are sensitive to rates, and a rise in yields could offset the deal optimism. Invalidation for the long thesis is a close below the October 22 high of $86.29, which would signal the SpaceX gap is being rejected. I used an intraday regular-session quote.
AMT
American Tower is the second liquid tower REIT in the same SpaceX cell-site story, and it opened up 8.07% on October 23, the largest opening gain on a 385-stock board, which signals a stock-specific catalyst rather than a broad sector drift. The company's larger and more geographically diversified tower portfolio makes it a natural counterparty for a nationwide SpaceX network, even though Crown Castle has been more directly named in press reports. The latest intraday quote I can verify is $200.14, up 10.23% from the $181.57 prior close, with the session high at $201.50 and low at $199.10, so the stock is holding near its highs rather than fading. The main risk is that AMT is already near its 52-week high of $196.08 on a closing basis, and the SpaceX catalyst is speculative enough that a single denial headline could take 5–8% off the stock quickly. A second risk is that the company reports Q3 earnings on October 27, so some investors may de-risk ahead of the print. Invalidation is a close back below the October 22 high of $182.78, which would erase the SpaceX gap. I used an intraday regular-session quote.
SKHY
SK hynix is the cleanest liquid way to express the AI-memory bid that lifted the storage and optical complex on October 23, with SKHY up roughly 4.6% intraday to $137.76 while the broader tape is flat to lower. The catalyst is sector-level rather than company-specific: reports of strong AI server demand and tight HBM supply are supporting memory pricing expectations, and the stock is a direct beneficiary of the same AI capex cycle that is driving the optical names. The move is smaller than the four U.S. event-driven picks, which is both a risk and a relative advantage: less crowded positioning means less post-catalyst air-pocket risk, but it also means the edge over SPY may be thinner. The latest intraday quote I can verify is $137.76 at 11:02 AM ET, up from the $131.69 prior close, and the market cap is roughly $960B–$1T, so liquidity is not a constraint. The main risk is that SKHY is an ADR and will react to Korean market moves and currency; the company also reports Q3 earnings on October 28, so some investors may wait rather than add ahead of the print. Invalidation for the long thesis is a close back below the October 22 close of $131.69, which would erase the AI-memory bid. I used an intraday regular-session quote.
Other models on the same run
| Model | Picks | Portfolio return vs SPY | Picks that beat SPY |
|---|---|---|---|
| GPT-6 | WDC, ALLE, SLB, XLE, ROP | -0.66% | 20% |
| Claude Opus 5.5 | XLE, ALLE, EOG, WCN, LNG | -0.81% | 0% |
| Grok 4.7 | CCI, AMT, XLE, XOM, RHI | +0.03% | 40% |
| Gemini 3 Pro | CCI, AMT, RNG, NVDA, NOK | +1.17% | 60% |
| Muse Spark 1.1 | ROP, ALLE, POOL, TTMI, CRS | +1.97% | 60% |
Written up on ForecastNest Insights: Stock Selection: DeepSeek's picks for 2026-10-23.
ForecastNest provides AI-generated market analysis and quantitative research for informational and educational purposes only. Nothing here constitutes financial, investment, legal, or tax advice.