MIDS is the online data science master's from Berkeley's School of Information, which focuses on the information side of computing rather than pure CS. It gets about 880 searches a month in the US, though that is down roughly 45 percent on last year. The degree is 27 units. You can take three courses a semester and finish in 12 months, take two a semester and finish in about 20, or drop to one a semester and take up to 32 months.
No GRE is needed. The curriculum covers machine learning and deep learning, generative AI and LLMs, NLP, data engineering, applied statistics and research design, AI ethics and policy, and product thinking and communication. Featured courses include Applied Machine Learning, Behind the Data: Humans and Values, and Fundamentals of Generative AI. It ends with a capstone, and Berkeley publishes past capstone projects, which are worth a look before you apply. There is an in-person immersion too, with a one-time $500 fee, and you pay your own travel and lodging. The format is what makes it different.
Classes are live and small, run online at set times, with recorded material in between. That is much more like a real seminar than the asynchronous model at Georgia Tech or UT Austin, and some people learn far better with a professor and classmates watching. It is also the main reason it costs so much. And it costs a lot. For 2026-27 the program fee is $2,850 per unit, plus a Berkeley campus fee of about $905 a semester and student health insurance unless you can waive it.
Berkeley estimates the total for a Fall 2026 starter at about $82,226 in tuition and fees, not counting insurance. Compare that to $12,000 to $13,000 for Georgia Tech OMSA or about $10,000 plus fees for UT Austin's MSDS. I think MIDS is a well-run program, and the ethics and research design side is better than anything at the cheap options. But the honest question is whether the live format and the Berkeley brand are worth around $60,000 more than a strong alternative. For most self-funded people, I do not think they are.
For someone whose employer covers it, or who has tried asynchronous study and failed at it, MIDS is a sensible choice.