What to Buy and When to Buy & Sell Stocks?
PNN Terminal, a new compass of investment, finds directions to elevate asset allocation
GS & Citi project 12-month asset price targets, we project & test daily/weekly/monthly PTs
Our projections help manage fear & panic due to market dips (e.g., tariff shock, black swan)
Reference: Goldman Sachs (projection), Bloomberg (Terminal), J.P.Morgan ($2B AI)
AutoScore precisions of options-based projection lines using automated unbiased t-tests & charts (New Gold Standard of Stock Predictive Analytics)
Stock Resilience & Alpha
Generate YesRatio = ƒ(SharpeRatio, ResilienceMetric) to improve alpha with VUSS (validated uptrend steep smooth) projection lines (Innovation of Investment)
Agentic AI Integration
Generate Buffett & Munger agentic joint advice by RAG multi-agents collaboration with self-hosted LLMs on CUDA GPU (Innovation of Investment Engineering)
(NVDA close price CPMA projection 9-26: 178.39; actual 9-26: 178.19)
(SPY actuals matched projections - confirmed by an automated t-test. Projection precisions automatically measured vary on stocks/metrics/projection periods)
(Sharpe-Yes ratio table outperformed the market - SPY Sharpe ratio 0.85)
We deliver concrete benefits to PNN fund managers & CEO, among them
Elevate Portfolio Performance: Leverage quantitative forward Sharpe-Yes ratio table and VUSS projection lines for asset allocation to outperform the market.
Improve Risk Management: Leverage qualitative joint advice of Buffett & Munger multi-agents using projection datasets. Forecast market corrections with PNN charts & precision metrics.
Improve Investment & Risk Strategies: Implement 'Buy the Dip'/'Sell the Rip'/'Do Nothing' strategies & manage fear/panic due to market correction, using V/IV-shape projection lines.
For CEOs – Capital Decision Timing: Elevate market cap through timely capital decisions (buyback, partnership, CEO capital agents) guided by the projection line of your stock.
Scalable self-hosted agentic AI platform for financial and government institutions
MIT reports 95% AI pilots fail, Athena improves your success rate with example agents/LLMs
Contact cto@pnncapitalus.com
Key Features of the Platform
Scalable Cloud Deployment
Fast deployment with efficient scaling. Sizable use cases.
Custom LLM Creation
Tailored large language models with proprietary data. Agentic RAG & LLM API.
Base for PNN Products
Provide self-hosted LLMs for agentic products of PNN Capital
First-principle Question 1: When is the right time to buy and sell stocks? Use PNN Terminal's future-dated projection lines (incrementally validated) to gain crucial insight.
2: How to implement Agentic RAG AI for my project? Launch Athena AI platform from AWS Athena, then use PNN Terminal as the example to design and develop your target state.
Products of PNN Capital
PNN Terminal
Agentic Investment Analytics (AIA) of asset allocation for investors & funds
Athena MAAC
Agentic collaborations displace ~25% human meetings by text & voice agents
1. PNN Terminal for Alpha Elevation. Implementation of USPTO # 19/245,322. Reference: Bloomberg Terminal
First Principle Problem: When is the right time to buy and sell stocks? Precise market timing is very hard but not impossible. Only a cadre of sophisticated asset managers may succeed at market timing; most retail investors and many fund managers are better off using timing-removed passive strategies (GPT/Gemini LLMs 2025)
Innovative Solutions to Improve Alpha: PNN Terminal's Future-dated Projections by Options Sentiment, YesRatio, Behavioral Investment Experiments, Agentic AI Integration, Agentic Technical Indicators (Ye 2011 2025, Black-Scholes 1997)
Asset Pricing Theories for Asset Allocation: paradigm shift from traditional asset pricing theories (Sharpe 1990) to options-based asset pricing & projection (Ye 2011 2025). The traditional asset pricing theories cannot solve the market timing problem because they have missed options data, technical indicators, and agentic AI integration.
Technical Indicators: provide high-level trend directions; PNN Terminal generates specific future-dated price points and projection lines.
Patent (Ye 2025): Implementation of USPTO # 19/245,322: System and Method for Agentic Investment Analysis Using Options Sentiment and Resilience Metrics
Stock Resilience: YesRatio = ƒ(SharpeRatio, ResilienceMetric) to improve stock picking.
Validity: PNN Terminal's future-dated projection lines & algos are incrementally validated with live automated unbiased back & forward tests.
How to Choose a Projection Line? Use CCA projection line if actuals self-validate CCA projections by live t-tests & charts. Corroborate with asset trends by TIs.
Get Evidence for Precise Projection Lines: projected/actual t-test results are generated automatically and incrementally (daily), accessible at a tab of PNN Terminal (Ye 2025).
Alpha strategy by PNN Terminal: pick asset based on validated projection line - uptrend steeper smooth | higer return lower beta (Ye 2025, Carhart 1997 momentum, Shiller 2013)
Beta strategy by PNN Terminal: pick asset based on validated projection line - less fluctuation | less covariance (Ye 2025)
Buy the V-shape Dip: projection line shows V-shape with future-dated higher prices after the dip
Reallocate Assets: switch flat or downtrend projection lines to VUSS projection lines
Improved CCA algo: Projection line of CCA algo is significantly closer to actuals than projection lines of the algos of USPTO#19/245,322
CPMA algo/projection line: Among 4 future-dated projection lines, the CPMA line is the closest to the actual prices based on the CPMA t-test means and p-value
LLM training on CUDA GPU: Tuning a new LLM on CUDA GPU for Buffett & Munger AI Agents of PNN Terminal with the 4th generation of LLM training and fine-tuning technology
2. Agentic Collaboration by Text and Voice Agents - USPTO # 19/216,814: System and method for AI agent-based meeting simulation and collaboration
3. LLM Cloud Security Posture - USPTO # 19/218,399: Cloud Large Language Model Security Posture Assessment System and Method
4. Agentic AI Video - USPTO # 63/780,599: System and method for video large language model (LLM) agentic collaboration
5. PNN CEO for Elevating Brand Market Cap - Implementation of USPTO # 19/321,197: System and Method for Elevating Corporate Market Capitalization Using an Agentic AI Framework on a Custom-tuned Large Language Model
