IBM builds a dedicated OpenAI practice for enterprise AI
Two enterprise moves set the tone today: IBM is putting tens of thousands of consultants behind OpenAI's models, and a startup founded by a former OpenAI researcher just banked several hundred million yuan to chase physical AI.
IBM announced Thursday it is building a dedicated OpenAI practice inside IBM Consulting, training tens of thousands of consultants to deploy OpenAI's models for the world's largest companies. Under the deal, terms of which were not disclosed, IBM and OpenAI will jointly market AI offerings and build industry-specific solutions for financial services, government, telecom, and retail. IBM will fold GPT-5.6, Codex, and ChatGPT Work into IBM Consulting Advantage, retrain tens of thousands of existing consultants on OpenAI's credentials over the next several months, and field a group of "Forward Deployed Experts" through OpenAI's Partner Network.
The announcement comes less than a year after IBM struck a similar alliance with Anthropic, and it expands the June OpenAI Daybreak cyber partnership to IBM's Autonomous Security product. The enterprise AI fight has moved from model quality to distribution, and OpenAI is running the systems-integrator playbook — Infosys, TCS, now IBM — that won the last generation of enterprise software. IBM gets a hedge: it now carries frontier models from both leading labs alongside its own Granite family, a model-agnostic stance that looks wiser after the company cut its 2026 revenue forecast last month. Whether the Switzerland-of-frontier-AI position pays better than picking a side is IBM's bet to prove.
Light Origins, the embodied-AI startup founded by former OpenAI researcher Jiang Xu, closed a Pre-A round of several hundred million yuan (tens of millions of dollars) led by Guoke Capital. China Merchants Group Ventures and Xianghe Capital also participated. The money funds large-scale embodied-intelligence model training, multimodal data infrastructure, and a full-stack robotics software and hardware platform, the company said.
Jiang spent the InstructGPT-to-GPT-4 era at OpenAI, and Chinese media reported he was cited eight times in the GPT-4 technical report's contributor list — tied with OpenAI co-founder Greg Brockman for the most mentions. His company, founded in late 2024 with more than 100 staff (over 90 percent of them in R&D), is pursuing what it calls a scaled-pretraining, scaled-alignment, scaled-deployment paradigm, and promises technical reports and cross-embodiment demos next. Embodied AI is still waiting for its ChatGPT moment, and Chinese capital is placing early, large bets on the labs most likely to produce it.
What to watch: whether other global consultancies follow IBM into dedicated OpenAI practices — and whether Light Origins' demos back up the physical-AI pitch.
IBM is betting it can sell every frontier lab's models without clients wondering which one is best. Can a consultancy really stay neutral? Tell us in the comments.
Sources: TechCrunch · IBTimes Singapore · Jiemian (界面新闻) · China Business Journal (中国经营报) · Sina Finance (新浪财经)
