IDC zvyšuje na rok 2029 prognózu výdavkov na infraštruktúru uznanú OCP o 2,7-násobok na 800 miliárd dolárov

IDC projects global spending on OCP-recognized IT equipment will reach $800 billion in 2029, about 2.7 times its prior forecast of $295 billion, and $892 billion in 2030. The 2030 total includes approximately $814 billion for servers and $628 billion in AI-related infrastructure spending; OCP-recognized equipment is forecast to account for about 51% of covered infrastructure categories. IDC attributes the acceleration to expanding AI infrastructure investment and adoption of open designs.

Analysis

The investable signal is not the headline spending curve; it is the expected mix shift toward power-dense compute and the infrastructure around it. If deployments materialize, bottlenecks in power distribution, liquid cooling, networking and grid interconnection can support pricing and backlog for suppliers such as Vertiv, Eaton and Schneider Electric, while GPU and networking demand benefits Nvidia, AMD, Broadcom and Arista Networks. These are conditional beneficiaries, not a direct read-through from OCP recognition to company revenue.

The key caveat is that this is an OCP-commissioned study, and the sharp revision versus last year makes forecast methodology and category scope more important than the point estimate. “OCP-recognized” spending is not necessarily incremental spending, nor does it establish market share gains for any named vendor. The concentration in servers also leaves the outlook unusually exposed to GPU availability, hyperscaler capex discipline and the economics of AI workloads; weaker utilization or a shift to custom accelerators could reduce system value even if compute deployments continue.

Near term, the release may reinforce AI-infrastructure positioning, but it is not independently verified earnings evidence. Over 1–3 months, track hyperscaler capex guidance, supplier backlog/conversion, data-center power availability and liquid-cooling orders. Over 6–18 months, grid constraints and AI return-on-investment are the main swing factors. Contrarian risk: investors may capitalize a long-range infrastructure forecast as if it were near-term revenue, while underweighting power and interconnection delays.

Key Decisions for Investors

• Avoid buying broad AI hardware exposure solely on this forecast. Treat it as a demand indicator, not an earnings revision; require corroboration in hyperscaler capex and supplier orders.
• Watch a relative-value expression: favor power and cooling exposure such as Vertiv, Eaton or Schneider Electric over undifferentiated server assembly if order growth and backlog conversion remain stronger. Do not initiate until company disclosures confirm the exposure; this release alone does not.
• For the next 1–3 months, monitor hyperscaler capex guidance, lead times and backlog conversion at data-center infrastructure vendors, alongside grid-interconnection and power availability updates. These determine whether projected demand can become shipments.
• Falsify the bullish infrastructure thesis if major cloud providers cut capex plans, relevant suppliers report slowing orders or backlog conversion, or power constraints materially defer deployments. A broad AI-capex slowdown would also weaken GPU, networking and facility-equipment beneficiaries.