Executive Summary
TheDayAfterAI News pits six of the most widely used AI chatbots against each other in a weekly stock-prediction showdown. This week’s subject is Destiny Tech100 Inc. (NYSE: DXYZ), a closed-end fund that provides public-market investors with exposure to a portfolio of leading private technology companies including SpaceX, OpenAI, Anthropic, Revolut, and Stripe. The fund closed at $26.52 on March 24, 2026, and was surging 18–27% in pre-market trading on March 25 after credible reports that SpaceX — its largest holding at roughly 16–20% of portfolio value — is preparing to file its IPO prospectus with the SEC as early as this week.
Each chatbot was given the same structured prompt and asked to deliver a 5-trading-day forecast (March 25–31, 2026) using pre-market data available before the U.S. market open on Wednesday, March 25. The probability each model assigns reflects whether the predicted closing price on March 31 is above or below the predicted opening price on March 25 — not the prior close.
The verdict: a dominant bearish consensus. Four of six chatbots predict DXYZ will finish the week lower than where it opens, with an average expected decline of roughly 5%. Claude stands alone as the sole bullish voice, projecting a modest 4.5% gain driven by the possibility of a confirmed SpaceX S-1 filing. Perplexity notably declined to provide any numerical forecast, calling precise predictions for a name this volatile “speculative and not robustly supported by the available data.”
Head-to-Head Comparison
| Chatbot | Pred. Open | Pred. Close | Change | Up % | Down % | Range | Signal |
|---|---|---|---|---|---|---|---|
| Claude | $33.50 | $35.00 | +4.5% | 57% | 43% | $29.50–$40.00 | Bullish |
| ChatGPT | $32.70 | $30.80 | −5.8% | 38% | 62% | $29.40–$35.80 | Bearish |
| Gemini | $32.80 | $28.50 | −13.1% | 35% | 65% | $25.50–$36.00 | Bearish |
| Perplexity | N/A | N/A | N/A | N/A | N/A | N/A | Declined |
| Copilot | $29.50 | $27.00 | −8.5% | 35% | 65% | $24.00–$33.00 | Bearish |
| Grok | $32.00 | $31.20 | −2.5% | 48% | 52% | $28.50–$35.50 | Bearish |
Note: Perplexity explicitly declined to provide numerical predictions, arguing that DXYZ’s extreme volatility and event-driven nature make precise forecasts unreliable. All other figures are as stated in each chatbot’s analysis.
Consensus at a Glance
| Metric | Consensus (excl. Perplexity) |
|---|---|
| Average Predicted Open | $32.10 |
| Average Predicted Close | $30.50 |
| Average Expected Change | −5.0% |
| Average Upside Probability | 42.6% |
| Average Downside Probability | 57.4% |
| Bullish Models | 1 of 6 (Claude) |
| Bearish Models | 4 of 6 |
| Declined to Predict | 1 of 6 (Perplexity) |
| Overall Consensus | Moderately Bearish |
The Catalyst: SpaceX IPO Filing
All six chatbots identified the SpaceX IPO filing reports as the single most important driver of DXYZ’s price action this week. Bloomberg, Reuters, and The Information all reported that SpaceX is preparing to file its IPO prospectus with the SEC as early as this week, targeting a June 2026 listing at a valuation of $1.5–1.75 trillion — which would make it the largest IPO in history, surpassing Saudi Aramco’s $29.4 billion raise in 2019.
SpaceX accounts for approximately 16–20% of DXYZ’s invested capital (or up to 35% including the xAI subsidiary acquired earlier in 2026). At the last reported NAV of $19.97 per share (Q4 2025), a public filing at the upper valuation range would imply roughly 40% upside to the SpaceX component of NAV alone. The fund also recently deployed $127 million into new investments including a $100 million Anthropic position, meaning the published NAV is already stale relative to the actual portfolio.
The critical question dividing the chatbots: is this catalyst powerful enough to sustain the gap-up, or will it trigger a classic “buy the rumour, sell the news” fade?
Individual Chatbot Analyses
Claude (Anthropic) — The Lone Bull
Predicted Open: $33.50 | Predicted Close: $35.00 | Upside Probability: 57%
Claude delivered the most detailed and bullish forecast of the six, assigning a 57% probability of net gain. Its base case envisions a +4.5% return from open to close, driven by the possibility that SpaceX actually files its S-1 during the forecast window — which Claude assigns a 45% standalone probability. Claude built a four-scenario framework: (1) SpaceX files and triggers a confirmation rally to $36–40 (45% probability), (2) no filing this week and a gradual fade to $29–31 (25%), (3) a macro shock overwhelms the catalyst (15%), and (4) a full short squeeze with filing confirmation reaching $45–55 (15%). The weighted expected close of $35.00 reflects the asymmetric upside from the catalyst balanced against substantial downside risks. Claude also flagged the Trump–Iran strike pause expiring around March 28 and the Musk v. OpenAI trial beginning March 30 as additional binary risk events.
ChatGPT (OpenAI) — Strong Open, Then Fade
Predicted Open: $32.70 | Predicted Close: $30.80 | Downside Probability: 62%
ChatGPT painted a picture of a “very strong opening, then partial fade.” It acknowledged the SpaceX catalyst as genuine but gave heavy weight to the valuation structure — DXYZ’s $19.97 NAV means a $32.70 open would represent a roughly 64% premium, with the fund’s $1 billion at-the-market offering via Jefferies creating overhang. ChatGPT provided a detailed day-by-day trajectory showing progressive decline from a $33.10 close on Wednesday to $30.80 by Tuesday, driven by post-gap mean reversion, profit-taking, and a macro backdrop that is “mildly supportive for the open but not fully stable for the full 5-day period.” It noted the VIX near 26, CNN Fear & Greed at “extreme fear” (16.9), and retail investors actually selling Monday’s rally rather than chasing it.
Gemini (Google) — The Most Bearish
Predicted Open: $32.80 | Predicted Close: $28.50 | Downside Probability: 65%
Gemini produced the most bearish forecast and the longest analysis of the group, projecting a 13.1% decline from open to close. Its thesis rested on three mechanical pillars: (1) the unsustainable closed-end fund NAV premium expanding to 64% at the opening price, (2) the unwinding of options dealer gamma hedges after the March 27 weekly expiration (citing 32,067 contracts with a 0.47 put/call ratio), and (3) the “buy the rumour, sell the news” dynamic as retail investors realise direct SpaceX equity access will soon be available, eroding DXYZ’s scarcity value. Gemini provided granular pre-market microstructure data including VWAP analysis, dark pool volume (743,000 shares, roughly 28% of pre-market flow), and a day-by-day trajectory mapping peak euphoria on Wednesday, digestion on Thursday, structural unwind on Friday at options expiration, and gravitational pull back toward the 200-day SMA ($27.15) on Monday–Tuesday.
Perplexity — The Conscientious Objector
Predicted Open: N/A | Predicted Close: N/A | Probability: Declined to state
Perplexity stands out as the only chatbot that explicitly refused to provide numerical predictions. Its extensive analysis covered all the same ground — the SpaceX catalyst, short interest dynamics (2.27 million shares, 3.65 days to cover), elevated VIX, deteriorating market breadth, and limited options liquidity — but concluded that “the most intellectually honest use of the available information is to characterise the risk environment, key levels, and drivers of potential moves, rather than to state a specific predicted opening price, closing price, or exact percentage odds.” It characterised the probability of a net gain or loss as “roughly balanced with a wide error band.” This epistemically cautious approach is arguably the most honest of the six, though it provides less actionable content for readers comparing predictions.
Copilot (Microsoft) — Reversion to Mean
Predicted Open: $29.50 | Predicted Close: $27.00 | Downside Probability: 65%
Copilot delivered the lowest opening price prediction of the group at $29.50 — notably below the pre-market indications of $32–34 that other models used. It also projected the lowest closing price at $27.00, implying an 8.5% decline. Its reasoning was straightforward: the pre-market spike is “news-driven, not technically supported,” moving averages and MACD show bearish momentum despite the spike, and the roughly 50% premium to NAV at pre-market levels is unsustainable. Copilot framed the outcome in three scenarios: a bull case ($31–33, 20% probability), a base case of mean-reversion to $27–30 (50%), and a bear case of profit-taking driving the price to $24–27 (30%).
Grok (xAI) — Cautiously Bearish
Predicted Open: $32.00 | Predicted Close: $31.20 | Downside Probability: 52%
Grok produced the most neutral forecast at 52/48 bearish-to-bullish, projecting only a 2.5% decline — essentially a coin flip. Its analysis noted the “classic breakout” pattern with March 24’s high-volume surge above all recent moving averages, the short interest of 2.27 million shares representing 10–20% of float with 3–4 days to cover creating squeeze potential, and peer funds like VCX surging over 1,000% on identical SpaceX positioning. Grok’s base case envisions a strong open, continued volatility with possible new highs early in the week, then consolidation and pullback as the initial hype digests. It characterised the setup as a “high-risk, event-driven name — not suitable for conservative accounts” and noted that any dip would likely be bought given the underlying catalyst strength.
Where All Six Agree
- SpaceX IPO is the dominant catalyst: Every model identified the SpaceX filing reports as the single most important driver of the pre-market surge and the week’s price action.
- Extreme NAV premium is a headwind: All models flagged the gap between the $19.97 reported NAV and the $29–$34 trading range as a structural risk, with premiums ranging from 50% to 64%.
- Short interest creates squeeze potential: All noted elevated short interest (1.4–2.3 million shares, 10–20% of float) as fuel for potential squeezes on the upside.
- Extreme volatility expected: Every model projected large intraday swings (5–15%+), with the widest range spanning $24–$40.
- Options data is thin: Multiple models noted that DXYZ options liquidity is limited, making derivatives-based signals less reliable than for large-cap names.
- The $1 billion ATM offering is an overhang: Management’s ability to issue shares at a steep premium to NAV was flagged by multiple models as a cap on sustained upside.
Where They Disagree
- Can the catalyst sustain the gap? Claude says yes (57% upside), projecting a SpaceX filing confirmation could push the stock to $36–40. The other four quantitative models say no, expecting the gap-up to fade over the week.
- Opening price estimates diverge: Four models cluster around $32–$33, while Copilot stands at just $29.50 — a notable outlier that suggests it weighted the pre-market range differently.
- Severity of retracement: Gemini projects the steepest decline (−13.1% from open), while Grok sees only a modest −2.5% fade. The range of closing predictions spans from $27.00 (Copilot) to $35.00 (Claude) — a $8.00 spread on a ~$32 stock.
- Predictability itself: Perplexity’s refusal to predict raises a meta-question: should any model be giving precise numbers on a name this volatile? Its epistemic caution stands in sharp contrast to Gemini’s granular day-by-day VWAP-level trajectory.
Methodology
Each chatbot was given an identical structured prompt requesting a 5-trading-day stock prediction for DXYZ. The prompt specified that the analysis should incorporate pre-market data, technicals, sentiment, fundamentals, macroeconomic factors, options flow, and systematic/seasonal considerations. All analyses were conducted using pre-market data available before the U.S. market open on Wednesday, March 25, 2026.
Models tested: Claude (Anthropic), ChatGPT (OpenAI), Gemini (Google), Perplexity, Copilot (Microsoft), and Grok (xAI). All models used their latest publicly available versions at the time of testing.
Probability definition: The “probability of increase” or “probability of decrease” reflects each model’s assessment of whether the predicted closing price on March 31 will be above or below the predicted opening price on March 25.
