QAOA Portfolio Optimization on Real IBM Quantum Hardware
Muhammad Faryad
Tier-2 IBM Qiskit Advocate, IBM-Certified Qiskit 2.x developer
Run QAOA on real IBM hardware, with real stock data and a free account
Most QAOA tutorials stop at a toy problem on a simulator. This one goes to hardware. You take real US stock prices, build the portfolio problem, and run it on ibm_fez. Then you compare the result with the exact classical optimum.
This is the full recording of my live Quantum Portfolio Optimization Workshop, with every slide and every notebook. Live seats were $220. The next live run is January 2027.
Taught live to 50 participants in three cohorts. Rated 4.9/5. One participant said, "The workshop provided a well-structured end-to-end perspective on QAOA for portfolio optimization, from classical portfolio formulation and QUBO construction to the Ising Hamiltonian, variational circuit, and hardware execution. I particularly appreciated the emphasis on exact classical verification, baseline comparisons, and the practical limitations..."
Hardware results from IBM Fez on six stocks.
The optimal portfolio was the most-measured outcome, four times more likely than a random guess.
What you will be able to do
Turn a portfolio selection problem into a QUBO, then an Ising Hamiltonian
Build and train a QAOA circuit in Qiskit 2.4 with the Sampler primitive
Pick an optimizer that converges under shot noise, and see why COBYLA and L-BFGS-B stall
Use a CVaR objective to get better portfolios from the same circuit
Train on IBM hardware with a batched optimizer built for the free Open Plan
Check every answer against the brute-force optimum
What's inside
2.5 hours of video, from first-principles derivation to the hardware run
Notebook 1: fetch stock data from Yahoo Finance, compute returns and covariance
Notebook 2: Qiskit Primitives, the parts you need for QAOA
Notebook 3: QAOA for 10 stocks on a simulator
Notebook 4: QAOA for 6 stocks on ibm_fez
All PDF slides, with complete derivations. They work as a standalone reference.
Who this is for
Quants and finance professionals who want to judge quantum optimization by running it
Developers and engineers who know Python and want a real hardware workflow
Students and researchers moving from textbook QAOA to actual devices
What you need
Python and basic linear algebra. Some familiarity with quantum gates helps. No finance background. No paid IBM account. The hardware notebook fits within the free Open Plan allowance.
What this is not
This does not claim quantum advantage. At six to ten stocks, a laptop solves the problem exactly. The point is to learn the full workflow and judge the method for yourself.