Twimbit X

Own the workflow, rent the model

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Join 6000+ industry executives who trust us.

Written By

Varnika Goel

Co-founder | Research Director

More from Twimbit

Welcome to this week’s edition of Twimbit X Insights.

AI infrastructure is attracting investment at remarkable scale, and the models inside your team's tools will keep changing as a result. The workflow around them is yours to keep.

Markets do the AI math

Two weeks ago, stocks tied to AI data centres had an unusually volatile day, with Nvidia down 9.3% and AMD down 17.3%. Investors were weighing how quickly new data centres will get built and how much it costs to borrow for them.

Morgan Stanley and UBS both called the selling overdone, adding that Nvidia's quarterly results would matter more for the sector.

Most sales and marketing teams have no data centres on their roadmap, so why would any of this matter to them?

The billions beneath your tools

Picture an outreach assistant that scores leads and drafts first emails, whether built in-house or bought inside a vendor's product. Behind it sits one model, with prompts tuned to that model's habits and a contract written around one vendor's terms.

Someone has to fund the data centres that model runs in, and the biggest builders are spending heavily. Amazon spent $77.3 billion on property and equipment in its latest quarter, alongside $42.4 billion of operating cash flow, and says the increase in its spending primarily reflects investment in AI.

Oracle reports that its capital spending rose primarily due to data centre expansion, more than tripling to $28.5 billion.

The demand is there too, with AWS sales up 39%, Oracle's revenue up 30% and Nvidia reporting on Tuesday that revenue rose 103% to $115.8 billion.

With this much capital behind it, expect the models underneath your tools to keep evolving, along with their pricing and terms.

The parts worth owning

Market sentiment will keep shifting, and most of it will never reach your team. What does reach it is the steady evolution of the models underneath, including their prices, usage limits and terms.

So the useful question for any revenue leader running AI is how quickly a workflow could move to a different model, and the answer depends on which parts of it are yours.

Your team owns the mapped workflow, the clean customer data and the definition of a good outcome. Those travel to any model, which makes them the assets worth investing in first.

The model is the part you rent and teams that treat it as a choice can move the day a better or cheaper option appears.

If a better model arrived next Monday, how quickly could your outreach assistant use it?

Before we wrap up

With Twimbit X, we build tools that help teams expand what they're capable of, not just how fast they move. The goal is to expand what your team can credibly handle. If this sparked an idea, let’s explore it together. Reach out to see how Twimbit X can help your team raise its own ceiling.

Book a call