Key takeaways
- Start cheap: Wharton's US$850 AI for Business or the free Elements of AI tells you within a month whether a leader needs a deeper programme.
- For a first serious credential, MIT Sloan (US$3,850, 6 weeks) and Oxford Saïd (£2,300, 6 weeks) give the best ratio of brand to time and money.
- Residential programmes at Stanford GSB, Kellogg and INSEAD cost US$11,000 to US$18,500 for 4 to 6 days; pay for them when the cohort of peers is the deliverable.
- For engineering teams, a subscription beats a course: DeepLearning.AI for current LLM and agent material, O'Reilly for depth, Maven for live cohorts.
- Agentic AI courses, Chief AI Officer programmes and vendor certifications each have their own ctaio.dev directory; this hub links to them instead of repeating them.
Which AI courses, university programs and subscriptions are worth paying for?
Ranked within each group by how much a technology leader gets out of it. "Checked" says where each number came from and when.
University executive programs
-
US$3,850 · self-paced online, date-based cohorts · 6 weeks, 6 to 8 h/week
For: Managers and executives who need a strategy-level grounding in AI and want the MIT name at under US$4,000.
Run by MIT Sloan with MIT CSAIL and delivered through GetSmarter, so the content is paced but you study on your own schedule inside each week. It is the default first purchase for a leader who has never had formal AI training. The provider page still showed past cohort dates when checked, so confirm the next 2027 intake before budgeting.
Checked: Provider page (MIT Sloan Executive Education), 19 Nov 2026; start dates on page were out of date
-
US$18,500 incl. lodging and meals · in person at Stanford · 6 days (21–26 Mar or 8–13 Aug 2027)
For: C-suite and business-unit leaders who own the AI operating model and want a residential week with peers. No technical background required.
This is the most expensive residential option on the list, and most of the value is the room: senior general managers and functional heads working through how an organisation changes when AI does part of the work. Both 2027 sessions have application deadlines (10 Feb and 25 Jun 2027), so it needs planning a quarter ahead.
Checked: Provider page (Stanford GSB Executive Education), 19 Nov 2026
-
US$850 · self-paced online, enrol any time · about 4 to 6 weeks
For: Managers and directors who want a credible, low-cost primer before they commit to a longer programme.
At US$850 it is the cheapest university-branded item here, and rolling enrolment means no waiting for a cohort. It covers AI and big data for decision-makers, not model building. Treat it as a screening course: if your team finds it too basic, move them up to MIT or Oxford.
Checked: Provider page (Wharton Executive Education), 19 Nov 2026
-
£2,300 · online with GetSmarter · 6 weeks
For: European and UK leaders, plus technical managers, who want an AI foundations course with an explicit ethics strand.
Oxford positions it for managers and technical professionals across functions, covering how AI works, what it can do and its ethical limits. It sits at roughly the same depth as the MIT course for a lower list price. No 2027 start dates were published when checked.
Checked: Provider page (Saïd Business School), 19 Nov 2026; start dates not yet listed
-
US$12,500 excl. accommodation · in person, Chicago campus · 4 days (12–15 Apr 2027)
For: Senior and functional leaders scaling analytics and AI, and data science leads who need to speak the language of the P&L.
Formerly Leading with Advanced Analytics and Artificial Intelligence, so it leans on data-driven decision culture more than generative AI hype. Four days in Chicago is easier to defend to a CFO than a week in California. It differs from Kellogg's online agentic course, which is covered on the agentic AI courses page.
Checked: Provider page (Kellogg Executive Education), 19 Nov 2026
-
€11,600 excl. VAT · in person, Fontainebleau · 5 days (19–23 Apr 2027)
For: Executives based in Europe or Asia who want a residential programme with an international cohort.
The April 2027 session runs in Fontainebleau; a Singapore run (S$16,600) is listed for late November 2026. INSEAD alumni get 15% off and the five days count toward the INSEAD Certificate in Global Management, which matters if you plan to stack programmes.
Checked: Provider page (INSEAD dates and fees), 19 Nov 2026
-
US$28,000 · blended, live online plus Berkeley campus modules · 8 months
For: Senior executives leading enterprise-wide AI adoption who want a long programme without the Chief AI Officer label.
Eight months and US$28,000 put it in the same bracket as the longest CAIO programmes, aimed at leaders without deep technical backgrounds. No cohort dates were listed when checked, so it is a 2027 option only if Berkeley publishes one. Berkeley's shorter agentic AI course appears on the agentic AI courses page.
Checked: Provider page (UC Berkeley Executive Education), 19 Nov 2026; no dates listed
Learning subscriptions
-
Free tier (videos) · Pro adds labs, quizzes and certificates, 7-day trial · Team plan US$25/user/month billed annually
For: Engineering managers and staff engineers who want short, current courses on LLM apps, agents and evaluation.
Andrew Ng's platform lists 150+ courses with new short courses almost weekly, often built with the model vendors themselves. The free tier is enough to watch everything; Pro is for people who will do the labs. The individual Pro price is shown at checkout rather than on the public page, so I cite the Team rate.
Checked: Provider membership and business pricing pages, 19 Nov 2026; individual Pro price not shown publicly
-
US$399/year list (US$239 offer to 30 Nov 2026) or US$59/month · self-paced · unlimited certificates
For: Leaders buying for themselves or a small team who want university and vendor courses under one subscription.
More than 10,000 courses, including Google, IBM, DeepLearning.AI and university specializations, with unlimited certificates. Degrees, MasterTrack and some premium courses are excluded, so check for the Plus badge before you assume a course is covered.
Checked: Provider page (Coursera Plus), 19 Nov 2026
-
Price set per course by the instructor · live cohorts on Zoom · typically 2 to 4 weeks, 3 to 5 h/week
For: Senior ICs and managers who learn better with deadlines, a cohort and a practitioner instructor than with a video library.
Maven hosts live, cohort-based courses taught by practitioners, and many current AI engineering and AI product courses run there. Quality varies by instructor, so judge each course by who teaches it. The business plan starts at 30 seats with a cancellable 3-month pilot.
Checked: Provider page (Maven for Business), 19 Nov 2026
-
US$49/month or US$499/year · books, video, live events and interactive labs
For: Engineering organisations that want one reference library for books, video and live training across AI and the rest of the stack.
The individual plan includes live expert-led events, interactive labs and the O'Reilly AI Academy learning tracks. It is the strongest option when your engineers need to read the definitive book on a topic as well as watch a course. Cloud sandboxes are on higher tiers only.
Checked: Provider pricing page, 19 Nov 2026
-
US$249/month or US$846 for 4 months · self-paced Nanodegrees with reviewed projects
For: Engineers moving into AI roles who need graded projects and human code review more than lecture content.
The subscription opens the full catalogue, with 80+ Nanodegree programs and 300+ shorter courses, and every project gets a personal review. It is the priciest subscription here, so buy it for a focused 2 to 4 month push and pause it afterwards.
Checked: Udacity help centre article, 19 Nov 2026; public pricing page did not render prices
Free
-
Free · self-paced video · 9 lessons of about 90 minutes (Part 1), plus a 30+ hour Part 2
For: Engineers and technical leaders with some coding experience who want to understand how models are actually trained.
Jeremy Howard's course, recorded at the University of Queensland, teaches PyTorch, fastai and Hugging Face from working code first. It assumes no university maths and no special hardware. The Part 1 recording dates from 2022, so pair it with something newer for LLM agents.
Checked: Provider site (course.fast.ai), 19 Nov 2026
-
Free · self-paced online · Introduction to AI (no maths or code) and Building AI (basic Python)
For: Non-technical executives, board members and whole departments that need baseline AI literacy fast.
Over 2 million people have taken it, in most EU languages, and the introduction requires no programming. It is the cheapest way to get a whole leadership team to a common vocabulary before you spend on executive programmes.
Checked: Provider site (elementsofai.com), 19 Nov 2026
Where do the agentic AI, CAIO and certification directories fit?
This page is the front door to the education cluster, and three narrower directories sit behind it. If the question is agentic AI specifically, the agentic AI courses comparison covers 13 programmes from MIT Professional Education at US$3,250 down to free courses from DeepLearning.AI and Hugging Face, split into leadership and builder tracks. I have kept those courses off this list so you are not comparing the same items twice.
If you are aiming at the Chief AI Officer seat, the Chief AI Officer programs directory compares 26 CAIO programmes, from a US$2,700 three-day course to US$28,000 ten-month programmes at Chicago Booth and Duke. Read it alongside how to become a Chief AI Officer, because a certificate rarely substitutes for a track record of shipped AI work.
If what you need is a proctored credential, go to the AI certifications hub. It maps vendor exams from Anthropic, Google, AWS, Microsoft, NVIDIA, GitHub, Salesforce and IAPP to roles, with ten deep-dive guides. A certification proves you passed an exam on one vendor's stack; a university programme proves you spent time on strategy with peers. Hiring managers read them differently, and you should buy them for different reasons.
Which university executive programme should a technology leader pick?
Pick by format before brand. Self-paced online programmes from MIT Sloan, Oxford and Wharton cost US$850 to US$3,850 and fit around a full-time job at 4 to 8 hours a week. Residential programmes at Stanford, Kellogg and INSEAD compress the work into 4 to 6 days and charge for the room, the faculty access and the cohort.
I'd send a newly promoted VP of Engineering or a CIO without AI training to MIT Sloan or Oxford first. Both run six weeks, both cover strategy, capability and ethics at a level a technical executive will find useful, and both cost less than a single conference trip with travel. Wharton's AI for Business is the screening option: if it feels basic, that is useful information too.
Residential programmes pay off in two cases. The first is a leader who will own an enterprise AI operating model and needs to test ideas against other senior general managers, which is what Stanford's The AI-Powered Organization is built for. The second is a leader whose employer funds executive education anyway and who wants a cohort outside their own industry. Berkeley's eight-month programme belongs in the same conversation as the longest CAIO programmes, and it only makes sense if you would otherwise be choosing one of those.
Plan the calendar early. Kellogg's Leading with AI runs once, on 12–15 Apr 2027, INSEAD's Fontainebleau session follows on 19–23 Apr 2027, and Stanford's March session closes applications on 10 Feb 2027. MIT, Oxford and Berkeley had not published 2027 intakes when I checked, so ask the admissions team for dates before you put the fee in a budget line. Fees exclude travel, and INSEAD's excludes VAT.
Are learning subscriptions better value than courses for engineering teams?
For teams that build, yes. A course is a fixed syllabus that ages from the day it is recorded, while a subscription follows the field. DeepLearning.AI adds short courses almost weekly, often co-built with the model vendors, and its free tier already includes every video. That makes it the cheapest way to keep an AI platform team current.
O'Reilly suits organisations where engineers need the definitive book as well as the video, and its US$499 annual plan includes live events and labs. Coursera Plus is the broadest catalogue and the easiest to justify for mixed teams of engineers, analysts and product managers. Maven fits people who stall on self-paced video: live cohorts of 2 to 4 weeks with a practitioner instructor create the deadlines that video libraries lack.
Udacity is the outlier at US$249 a month. Buy it for an engineer moving into an AI engineering role who needs reviewed projects, run it for one Nanodegree, then pause it. For the career paths those engineers are moving into, see the AI engineer role guide.
How should a CTO build a 2027 AI learning budget?
Budget in three layers. The first is literacy for everyone, which can be free: Elements of AI or the free DeepLearning.AI tier gets a whole department to a shared vocabulary, and that is what the EU AI Act literacy requirement is asking for in practice. The second is a subscription for builders, at roughly US$300 to US$500 per engineer per year. The third is one or two executive programmes a year for the people who set AI direction.
Judge the spend by outcomes. A certificate count tells you little; a reduction in time from AI prototype to production, or a governance review that no longer stalls, tells you the training worked. Before you approve a US$12,000 residential week, ask the attendee what decision they will make differently afterwards and check back in a quarter.
Keep learning connected to the rest of the cluster. Pair subscriptions with the AI newsletters and AI YouTube channels worth following, and send the people who need a network rather than a syllabus to the right event from the AI conferences guide.
Listed here, or should be?
If your entry is out of date or you think something belongs on this list, email [email protected]. Entries are editorial and free. A featured, labelled Sponsored slot is available through vendor listings.
Related questions
- Which agentic AI course should I take, from MIT to free?
- Which of the 26 Chief AI Officer programmes is worth the fee?
- Which AI certifications carry weight with hiring managers?
- How do I become a Chief AI Officer?
- Which AI newsletters are worth an executive's inbox?
- Where does all AI education content on ctaio.dev start?
More from AI Education
- AI Education: where technology leaders learn AI in 2027
- AI conferences
- AI YouTube channels
- AI accounts on X
- AI newsletters
- AI podcasts for executives
- AI voices on LinkedIn for executives
- AI books for technology leaders
- AI peer communities for executives
- Agentic engineering resources
- Vibe coding resources
- AI tool stack for CTOs
- The full AI Education hub on We The Flywheel, with niche lists for every role