AI Chatbot Stock Forecast Showdown: Six Leading AI Models Predict APLD’s Price Trajectory

Introduction

At TheDayAfterAI News, we believe the best way to evaluate the predictive capabilities of today's leading AI systems is to put them to the test — side by side, on the same question, under the same conditions. In this recurring series, we task six of the most widely used AI chatbots with a single challenge: forecast the price movement of a specific stock over the coming five trading days.

This week's subject is Applied Digital Corporation (NASDAQ: APLD), a company at the centre of the AI infrastructure boom. APLD has pivoted from cryptocurrency mining hosting to high-performance computing (HPC) and AI data centre development, securing a contracted revenue backlog of US$16 billion over 15 years. With a closing price of US$33.56 on February 17, 2026, the stock entered our forecast window amid significant turbulence — most notably NVIDIA's disclosed divestiture of its entire strategic stake.

Each chatbot was given the same prompt requesting a comprehensive analysis covering technical indicators, fundamental catalysts, macroeconomic factors, options market dynamics, and sentiment indicators. They were asked to provide a predicted opening price for February 18, a predicted closing price for February 24, an estimated price range for the period, and a probability assessment for the stock ending higher or lower.

The six AI chatbots assessed in this edition are: Gemini, Claude, Perplexity, Grok, ChatGPT, and Copilot. What follows is a synthesis of their independent analyses, a head-to-head comparison of their forecasts, and our editorial observations on where they agreed, where they diverged, and what that tells us about the current state of AI-driven financial analysis.

Forecast Comparison at a Glance

The table below presents each chatbot's core predictions side by side, enabling readers to quickly compare the range of outcomes these AI models envision for APLD over the five-day trading window.

Reference closing price (Feb 17, 2026): US$33.56. Gemini also assigned a 10% probability to a neutral/flat outcome.

Consensus View: What the Numbers Reveal

Averaging across all six models, the consensus predicted opening price is approximately US$32.19 and the consensus predicted closing price is approximately US$33.80, implying a modest net gain of around US$1.60 (approximately 5.0%) over the five-day window. The average probability of a price increase stands at 55.3%, while the average probability of a decrease is 43.0%, with Gemini uniquely assigning a 10% neutral probability.

The consensus price range spans roughly US$29.26 on the low end to US$37.41 on the high end, reflecting an expected intra-week swing of over 25% — consistent with APLD's extremely elevated beta (reported between 7.0 and 7.19 across analyses) and implied volatility exceeding 100%.

Individual Chatbot Analyses

Gemini — The Cautious Bear

Gemini produced the most comprehensive single report among the six, delivering a 10-page strategic equity research document with detailed daily price forecasts, options microstructure analysis, and macroeconomic scenario modelling. It was also the only model to adopt a decisively bearish stance, assigning a 55% probability of net price decline.

Central to Gemini's thesis is the concept of a "liquidity event versus fundamental degradation" distinction. While acknowledging APLD's robust US$16 billion backlog and successful pivot to AI infrastructure, Gemini argued that the NVIDIA stake divestiture, combined with hawkish Federal Reserve anxiety and Supreme Court tariff uncertainty, would create a "wall of worry" too steep for the stock to climb in the near term.

Gemini's most distinctive contribution was its granular options analysis, including max pain calculations pinning the stock near US$34.00 through Friday's expiration, and a detailed assessment of how dealer hedging dynamics would influence day-by-day price action. Its daily forecasts traced a path from US$33.75 on Wednesday to US$32.55 by Tuesday — a controlled drift lower punctuated by temporary stabilisation around the options expiry pin.

Claude — The Balanced Pragmatist

Claude delivered a slightly bullish outlook with a 57% probability of price increase, predicting a close of US$34.00 by February 24. Its analysis stood out for the breadth of its data integration, synthesising technical indicators, fundamental catalysts, macro conditions, and sector peer performance into a cohesive narrative.

A notable feature of Claude's report was its emphasis on the NVIDIA earnings anticipation effect (reporting February 25) as a positive undercurrent for AI infrastructure names throughout the week. Claude also provided unusually detailed macro context, citing specific figures such as the S&P 500 at 6,843, the NASDAQ's five consecutive losing weeks, and gold surging past US$5,000/oz.

Claude's three-scenario framework (bull at 25%, base at 55%, bear at 20%) was clearly articulated, with the base case envisioning consolidation in the US$32.50–US$34.50 range. The report was particularly effective in identifying the Core PCE print on February 20 as the single most consequential variable for the week's outcome.

Perplexity — The Disciplined Statistician

Perplexity took the most methodologically rigorous approach, explicitly declining to provide "fabricated" price targets before establishing a defensible quantitative framework. Its initial response consisted of a detailed checklist for conducting proper short-term equity analysis, followed by a statistically derived forecast using a log-return model calibrated on approximately 90 trading days of historical data.

The result was the most conservative directional call: a 52.8% probability of price increase with a central closing estimate of US$33.95 — barely above the starting price. Perplexity's 1-sigma confidence band of US$28.86–US$39.93 was the widest among all six models, honestly reflecting the extreme uncertainty inherent in a high-volatility, high-beta stock over such a short horizon.

While Perplexity's intellectual honesty is commendable, its reluctance to integrate qualitative catalysts (FOMC minutes, options expiry dynamics, NVIDIA sentiment) into its probability estimates arguably left value on the table for readers seeking actionable guidance.

Grok — The Optimistic Catalyst Hunter

Grok projected one of the more bullish outcomes, with a predicted close of US$35.00 and a 60% probability of price increase. Its analysis placed heavy emphasis on fundamental catalysts: analyst buy ratings with price targets ranging from US$40 to US$58, the US$16 billion contracted backlog, and data centre sector tailwinds projecting US$3 trillion in cumulative investment by 2030.

Grok's technical analysis incorporated Elliott Wave theory, suggesting a corrective phase ending near US$28–US$29 with a potential impulsive rebound targeting US$35 and above. The report also noted unusual options activity, including 71,000 weekly US$25 calls, as evidence of bullish positioning.

One area where Grok's analysis was less precise was in its macro timing — it placed the GDP/PCE data release on February 19 rather than February 20, and listed the Fed Minutes and Jobless Claims on the same day. These minor inconsistencies, while not altering the overall directional thesis, highlight the importance of verifying event calendars in time-sensitive forecasts.

ChatGPT — The Cautious Technician

ChatGPT offered the most practically structured forecast, providing clear support and resistance levels, day-by-day path estimates, and explicit invalidation criteria. It predicted the highest probability of upside at 65%, yet paradoxically set a relatively modest closing target of US$32.80 — suggesting it expected the stock to recover only partially from its gap-down opening.

ChatGPT's analysis was distinguished by its attention to real-time pre-market data, citing specific bid/ask spreads (US$30.15/US$31.04) and ATR values (approximately 3.87). Its day-by-day estimates painted a picture of initial volatility followed by gradual stabilisation, with the stock grinding higher through short-covering but unable to fully reclaim the pre-gap levels.

The report's invalidation framework — break below US$30 for bear case, reclaim above US$33.50 for bull case — provided readers with clear signposts for adjusting their expectations in real time.

Copilot — The Recovery Optimist

Copilot projected the strongest absolute recovery, with a predicted close of US$34.50 representing an 11.3% gain from its predicted open of US$31.00. It assigned a 62% probability of the period ending higher, anchoring its bullish case on analyst upgrades, hyperscaler lease momentum, and concentrated call interest in the options market.

Copilot's day-by-day forecast traced a steady recovery arc: stabilisation at US$32.50 on Wednesday, gradual appreciation through Thursday and Friday (with a brief expiry-related dip), and continued momentum into the following week. Its estimated range of US$29.00–US$38.00 was among the wider forecasts, acknowledging both squeeze and flush scenarios given the 30–36% short interest.

The report was concise and well-organised but lighter on macro analysis compared to Gemini or Claude, focusing primarily on company-specific catalysts and positioning data.

Key Themes Across All Six Models

Points of Agreement

The NVIDIA exit as headline risk, not fundamental damage. All six models recognised that NVIDIA's divestiture of its APLD stake was a sentiment shock rather than an indication of deteriorating business fundamentals. Multiple models noted that the shares were absorbed by high-quality institutional allocators including Vanguard and Jane Street.

Extremely high short interest creates asymmetric risk. Every model flagged the approximately 34% short float (86 million+ shares) as a critical variable. While none predicted an imminent short squeeze, all acknowledged that this positioning creates the potential for explosive upside moves if a catalyst emerges.

Options expiry on February 20 as a key volatility event. All models identified the monthly options expiration as a structural force that would influence price action, with several noting the call-skewed open interest and the potential for pinning effects around key strike prices.

The US$16 billion backlog provides a fundamental floor. Despite varying directional biases, every model cited APLD's contracted revenue pipeline as evidence of durable business value, distinguishing it from more speculative AI-adjacent names.

Points of Divergence

Opening price estimates varied significantly. The predicted opening prices ranged from US$31.00 (Grok and Copilot, reflecting expected gap-down from NVIDIA news) to US$33.90 (Gemini, anticipating pre-market short covering). This US$2.90 spread — representing an 8.6% difference — highlights the challenge of forecasting gap behaviour in volatile stocks.

Directional bias split between bears and bulls. Gemini stood alone as decisively bearish (55% probability of decline), while the remaining five models ranged from marginally bullish (Perplexity at 52.8%) to moderately bullish (ChatGPT at 65%). This 30-percentage-point spread in directional probability underscores the genuine uncertainty surrounding APLD's near-term path.

Macro weighting differed substantially. Gemini and Claude placed heavy emphasis on macroeconomic catalysts (FOMC minutes, Core PCE, Supreme Court tariff ruling), while Grok and Copilot tilted more heavily toward company-specific and sector catalysts. This methodological divergence explains much of the variance in their conclusions.

Methodology and Quality Comparison

Beyond the numerical forecasts, the approaches taken by each model reveal important differences in analytical philosophy.

Editorial Observations

Several patterns emerge from this multi-model comparison that are worth highlighting for readers who follow our AI chatbot forecast series.

First, the models are converging on process but diverging on conclusions. All six chatbots now demonstrate competency in integrating technical analysis, fundamental data, and macro factors. The variation in their outputs stems less from analytical gaps and more from different weighting of those factors — precisely the same source of disagreement among human analysts.

Second, there is a notable "bullish bias" in AI forecasting. Five of six models predicted upside, with only Gemini taking a bearish stance. This may reflect training data patterns where stocks more frequently close higher than lower over short periods, or it may indicate a tendency for AI models to weight fundamental strength more heavily than technical deterioration.

Third, the gap-down pricing challenge exposed meaningful differences. The US$2.90 spread in opening price estimates (US$31.00 to US$33.90) reflects a genuine limitation in short-term forecasting. Models that incorporated pre-market data (ChatGPT, Copilot, Grok) anchored lower, while those relying primarily on prior-close extrapolation (Gemini, Perplexity) anchored higher. In volatile stocks with overnight catalysts, this distinction is material.

Fourth, Perplexity's epistemological honesty stands out. By explicitly refusing to "fabricate" numbers without proper data inputs and instead providing a statistical framework, Perplexity demonstrated a commitment to intellectual integrity that, while potentially less satisfying for readers seeking definitive predictions, represents arguably the most responsible approach to short-term price forecasting.

Methodology

Each AI chatbot was given an identical prompt requesting a five-day stock price forecast. The models used their own web-search and data-retrieval capabilities; no proprietary data was provided. Responses were collected without modification. Variations in depth, format, and analytical approach reflect each platform's native capabilities.