AI Spending Shifts Beyond Chips
- Enterprise AI budgets are expanding from raw chips into data, storage, services and distribution infrastructure. - Nvidia backed VAST Data in a round valuing it at about $30 billion, while VAST reportedly raised $1 billion in that round. - Companies are committing dedicated capacity, exemplified by a reported $260 million, 36-month Axe Compute deal for 2,304 Nvidia B300 chips (stocktitan.net).
The AI buildout is widening from chips to the rest of the stack: storage, data systems, networking and dedicated capacity contracts. (cnbc.com) VAST Data said on April 22 that it closed a $1 billion Series F round at a $30 billion valuation, with Drive Capital and Access Industries leading and Nvidia joining the round. The company said the financing included both primary and secondary capital and valued it at more than three times its late-2023 Series E mark of $9.1 billion. (vastdata.com) VAST started in high-performance storage and now pitches an “AI Operating System” that manages data, databases and compute orchestration for large model workloads. Chief executive Renen Hallak told CNBC the company is moving beyond storage into software that helps customers run AI infrastructure more efficiently. (cnbc.com) That change lines up with where enterprise budgets are moving. Gartner said on January 15 that worldwide AI spending will reach $2.52 trillion in 2026, up 44% year over year, and said AI infrastructure alone will add $401 billion as providers build out the foundations for deployment. (gartner.com) The money is also showing up in long-term capacity deals instead of short-term cloud rentals. Axe Compute said on April 22 that it signed a 36-month contract worth about $260 million to deliver a dedicated cluster of 2,304 Nvidia B300 graphics processing units, plus high-speed storage, from one U.S. Tier 3 data center. (markets.businessinsider.com) The contract illustrates what buyers are paying for after the chip purchase itself: reserved access, power, storage and a site ready to run training jobs without competing for shared capacity. Axe Compute said the agreement includes renewal options beyond the initial three years. (barchart.com) Industry data points the same way. IDC said AI infrastructure spending reached $89.9 billion in the fourth quarter of 2025 and projected the market will exceed $1 trillion by 2029, with enterprises treating AI capacity as a structural operating cost. (idc.com) Gartner said on April 22 that total worldwide information technology spending will hit $6.31 trillion in 2026, up 13.5%, with data center systems spending rising 55.8% to nearly $788 billion. Gartner attributed that jump to “accelerating momentum in AI infrastructure and advanced memory.” (gartner.com) Nvidia’s presence in VAST’s round fits that broader shift. As AI customers spend on the systems around the graphics processor, chip vendors and infrastructure startups are both trying to capture more of the budget that sits between the model and the data center floor. (cnbc.com)
