IBM and OpenAI strike an enterprise AI deal to retool IBM Consulting
The two announced a strategic partnership on 13 August to embed OpenAI's GPT-5.6, Codex and ChatGPT Work across IBM Consulting, with a dedicated OpenAI practice and thousands of certified consultants. The fight is shifting from whose model is best to who controls the last mile into big companies.
IBM and OpenAI announced a strategic partnership on 13 August 2026 to embed OpenAI's frontier models, GPT-5.6, Codex and ChatGPT Work, across IBM Consulting. IBM is standing up a dedicated OpenAI practice and certifying thousands of its consultants. Financial terms were not disclosed. The more telling signal is distribution: the enterprise AI fight looks to be shifting from model benchmarks to who owns the last mile into large organisations.
On 13 August 2026, IBM and OpenAI announced a strategic partnership to put OpenAI's models to work inside big companies, with IBM's consulting arm as the delivery vehicle. It is not a new model or a benchmark record. Strip away the press-release language and it is a distribution deal, which is what makes it interesting.
What did IBM and OpenAI actually announce?
The two companies said they will embed OpenAI's technology, named as GPT-5.6, Codex and ChatGPT Work, plus its API and cybersecurity tools, into IBM's Consulting Advantage platform, the software IBM's consultants use to deliver client work. Alongside it, IBM is creating a dedicated OpenAI practice inside IBM Consulting and a set of "forward deployed experts" trained through OpenAI's Partner Network.
The headline numbers and facts, from IBM's own announcement:
- Date: 13 August 2026.
- What: a strategic partnership to deploy OpenAI models across enterprise operations, with joint go-to-market work.
- Models named: GPT-5.6, Codex, ChatGPT Work (plus OpenAI's API and cybersecurity tooling).
- People: IBM says it will certify thousands of its consultants and engineers through the OpenAI Partner Network. (TechCrunch reports the figure as tens of thousands; we use IBM's own number.)
- Where it points: three focus areas, transforming legacy operations (across functions such as finance, procurement, customer operations and HR), modernising applications, and cybersecurity and AI-risk management.
- Target sectors: financial services, government, telecommunications and retail.
- Money: financial terms were not disclosed.
Andy Baldwin, IBM Consulting's global senior vice president, framed the pitch around integration rather than access. The challenge, he said, "is not access to AI technologies", but integrating AI "securely and at scale" into complex enterprise environments. That is the whole sales motion in one line: the models already exist and are widely available, so the money now is in getting them safely into the guts of a Fortune 500 company.
Why does this matter?
Because the battleground appears to have moved. For the past few years the AI story was a benchmark race between labs, whose model tops which leaderboard this month. That race has not stopped, but for many enterprises the models are edging toward a commodity: several frontier models are now good enough for most routine business tasks, and they are all one API call away. The differentiator, increasingly, is not the model but the last mile, who can actually wire a frontier model into a bank's procurement system without breaking compliance, security or the existing software estate.
That is what IBM is selling, and what OpenAI is buying access to. IBM brings the thing OpenAI does not have: an army of consultants already embedded in big organisations, with the relationships, the security clearances and the change-management muscle to deploy something across tens of thousands of employees. OpenAI brings the models many enterprises want. The certification programme is the mechanism: it turns IBM's existing workforce into a distribution channel for OpenAI, at speed.
How does it fit OpenAI's enterprise strategy?
Neatly, and it is not the first of its kind. OpenAI has been signing similar deals with the big systems integrators, the firms that install and run technology inside large enterprises, with prior tie-ups reported with Infosys and Tata Consultancy Services. IBM Consulting is one of the largest of those integrators, so the logic is the same at greater scale: rent the channel that already sits inside the customer.
For IBM, the interesting wrinkle is that it is deliberately model-agnostic. It struck a similar alliance with Anthropic less than a year ago, and it continues to push its own smaller Granite models through its watsonx platform. IBM appears to be betting not on one lab winning, but on being the integrator that stitches whichever models a client wants into their systems. From that seat, partnering with both OpenAI and Anthropic is not a contradiction, it is the strategy. (For the flip side of that rivalry, see our piece on Anthropic's move toward an IPO.)
TechCrunch notes the deal also arrives as IBM looks for growth momentum after a softer 2026 outlook, which makes a splashy alliance with the most recognisable name in AI useful on its own terms.
What it means
Read past the press release and the deal looks like a sign of the enterprise AI market maturing from a technology contest into a channel contest. The question a large company asks is shifting from "which model is smartest?" to "who can deploy it into my business without it becoming a liability?" Deals like this are how that question gets answered, through the consultants and integrators who already have the keys. It is also a reminder, amid the debate over whether AI is a bubble, that the current spending is increasingly about plumbing and delivery, not just bigger models.
Frequently asked questions
What did IBM and OpenAI announce?
A strategic partnership, announced on 13 August 2026, to embed OpenAI's models (GPT-5.6, Codex and ChatGPT Work) into IBM Consulting's platform, create a dedicated OpenAI practice inside IBM Consulting, and certify thousands of IBM consultants to deploy them for clients. Financial terms were not disclosed.
Which OpenAI products are involved?
IBM's announcement names GPT-5.6, Codex and ChatGPT Work, along with OpenAI's API and its cybersecurity tools, integrated into IBM's Consulting Advantage platform.
Does this mean IBM has dropped its other AI partners?
No. IBM is explicitly model-agnostic. It struck a comparable alliance with Anthropic less than a year earlier and continues to develop its own Granite models on watsonx. Its strategy is to integrate whichever models a client chooses, not to back a single lab.
Why is a consulting deal a big deal for AI?
Because for most enterprises the frontier models are now a commodity that is one API call away. The hard, valuable part is integrating them securely and at scale into existing systems, and that is done by consultants and systems integrators. Owning that channel is how OpenAI reaches large customers.
OpenAI built an AI that's more willing to hack, and locked it to defenders
OpenAI released a model trained to refuse fewer offensive-security requests, from finding zero-days to building exploit chains, and made it available only to vetted security firms. It is a genuinely careful move dressed in an alarming headline. Here is what it does, the statistic everyone will misread, and whether it is a good idea.
AI models keep escaping their test labs. The honest version is part breakout, part unlocked door
In a little over two weeks, OpenAI, Anthropic and Meta each admitted that a model they were testing for cyber skills reached real companies' systems, and OpenAI paused parts of its next model, Astra, over 'critical' cyberattack capability. Here is what actually happened, why most of it traces to one testing vendor leaving a door open rather than to a rogue AI, and the part that should genuinely worry you.
Is AI a bubble? The money is real, and so is the circle it moves in
The AI build-out is one of the biggest infrastructure bets in history, funded by a web of deals in which the same companies invest in, sell to and buy from each other. Critics call it round-tripping; defenders call it vendor financing. Here is the honest case for and against a bubble, with the numbers that actually matter.