Overview
An investment analysis comparing NVIDIA and Intel as long-term holdings, using six years of monthly market data benchmarked against the S&P 500, portfolio optimisation and a review of company fundamentals.
Problem
A long-term investor is choosing between two semiconductor companies with very different recent trajectories. Which offers the better risk-adjusted opportunity, and how could they be combined?
My contribution
I completed this project individually: data collection and cleaning, the full Excel model (returns, volatility, beta, Sharpe ratios), the Solver optimisation and efficient frontier, and the financial statement and valuation comparison.
Approach
Collected monthly adjusted closing prices for NVIDIA, Intel and the S&P 500 from January 2018 to December 2023. Calculated monthly returns and annualised risk/return statistics, estimated beta by regression against the index, and compared Sharpe ratios using a risk-free rate based on the 10-year US Treasury yield.
Analysis
Built a two-asset portfolio model and used Excel Solver to find the weights maximising the Sharpe ratio, then plotted the efficient frontier by varying the target return. Complemented the market analysis with gross and operating margin, ROE, current ratio, debt-to-equity, and P/E and P/B multiples from the annual reports.
Key findings
- NVIDIA delivered a substantially higher average return than both Intel and the S&P 500, with markedly higher volatility.
- Intel underperformed the S&P 500 on a risk-adjusted basis.
- NVIDIA's beta was above 1, indicating higher sensitivity to market movements.
- The maximum-Sharpe portfolio allocated the majority of weight to NVIDIA.
Evidence
- Final report (14 pages)
- Excel model with six analysis sheets
- Efficient frontier chart
Skills demonstrated
- Financial modelling
- Built an Excel model calculating monthly returns, volatility, beta and Sharpe ratios for two stocks and the S&P 500.
- Portfolio optimisation
- Used Excel Solver to find maximum-Sharpe portfolio weights and plotted an efficient frontier.
- Risk and return analysis
- Compared average and annualised returns, standard deviation and beta against the S&P 500.
- Financial statement analysis
- Calculated gross and operating margin, ROE, current ratio and debt-to-equity from annual reports.
- Equity valuation (multiples)
- Compared P/E and P/B ratios across both companies.
- Data preparation
- Collected and cleaned six years of monthly adjusted price data for three series.
What I learned
Risk-adjusted metrics change the conversation: a higher return is only attractive relative to the risk taken. I also learned how sensitive optimisation results are to inputs, which made documenting assumptions as important as the model itself.
Limitations acknowledged
- — Relies on historical returns, which may not predict future performance.
- — A two-asset portfolio ignores diversification across sectors.
- — Results are sensitive to the sample period and risk-free rate chosen.