Broadcom just announced VMware AI Factory at VMware Explore, and I want to talk about why it matters — and about the bigger story it's part of.
VMware AI Factory solves a real, hard problem: getting from "we want to run AI on our own infrastructure" to "we are running AI on our own infrastructure" that used to take weeks or months. Provisioning bare metal, standing up the software stack, wiring GPUs, choosing and deploying a model — all of it, manually, slowly. AI Factory automates that entire path. Enterprises can now go from bare metal to a served model in hours, choose their own hardware, and run more than 150 open source and commercial models, including newer, lower-cost options like Qwen, on infrastructure they fully control.
That last point matters more than it might look at first. If you can run a capable open model on your own private cloud, your options increase and the economics of a lot of AI work change in your favor. Not every job to be done, needs the most expensive model available. Some do. Most probably don't. VMware AI Factory is what makes that choice available in the first place — and it's exactly the kind of infrastructure decision that needs a financial answer sitting right next to it.
That's where the rest of the story comes in, and it's a story we're building together, as one Broadcom, not two separate efforts.
I've said before that a token, at its core, is a prediction of the next best word — how accurately, efficiently, and effectively that prediction happens is the basis of artificial comprehension itself. That's true whether the token comes from a frontier model in the public cloud or an open model running on your own hardware in your own data center. VMware AI Factory gives you control over where and how that prediction happens. ValueOps AI Tokenomics is what tells you what it was worth once it does.
VMware AI Factory already gives you real telemetry through its AI Gateway — token throughput, latency, compute and memory utilization. That's genuinely valuable data, and it's exactly the kind of foundation the rest of the picture gets built on. ValueOps AI Tokenomics takes that same consumption and connects it back to the initiative it was funded to support, so you know not just what happened and how much it consumed/cost, but why it mattered, and whether it's eligible to be capitalized instead of expensed.
This is exactly why we've been saying, again and again, that this isn't a spending problem — it's an alignment problem. VMware AI Factory just made it dramatically easier and less expensive to run AI on your own terms. That's a real accomplishment, and it raises the stakes on the alignment question in the best possible way: more of the enterprise is about to be running more AI, faster, than ever before, on infrastructure built for exactly that.
Where VMware AI Factory gets you from metal to model, ValueOps AI Tokenomics gets you from model to meaning. Together, ValueOps AI Tokenomics with VCF AI Factory enables organizations to be financially savvy about AI investments and consumption and govern and steer your business better, taking into account security,data sovereignty and data gravity. That's the complete arc: infrastructure to value, one Broadcom story from the ground up.
Get in touch with our AI experts to learn more!
It automates deploying AI on private infrastructure, taking the process from bare metal to a served model in hours instead of weeks.
Running capable open models on private cloud infrastructure increases options, improves economics, and allows for control over security and data sovereignty.
VMware AI Factory provides the infrastructure and telemetry, while ValueOps AI Tokenomics connects that data to business meaning and financial value.
Through its AI Gateway, it provides data on token throughput, latency, compute, and memory utilization. Why track AI consumption?
Connecting consumption to funded initiatives helps determine if the investment was worthwhile and whether it can be capitalized instead of expensed.