iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Cheaper AI models could change which chips and data-centre resources are in demand, but the available statements from Infosys CEO Salil Parekh do not establish that they are reducing capacity utilization. In a July 2026 call, Parekh discussed lower running costs and more efficient token use for Infosys’s smaller language models, alongside client attention to token and memory-chip costs. Those remarks point to pressure for efficiency—not a confirmed decline in data-centre usage.
What did Infosys CEO Salil Parekh say?
On Infosys’s Q1 FY27 media call in July 2026, Parekh said the company’s small language models for banking and IT operations were gaining traction. He described their smaller parameter sets, lower running costs and more efficient token use: “The parameter set is smaller, the cost of running it is smaller, the token usage is more efficient, and so on.” The statement describes Infosys offerings; it is not an independent benchmark of small models against larger ones.
Parekh also said clients were paying considerable attention to token costs and data sovereignty. On infrastructure, he mentioned memory-chip costs and the continuing need for efficiency. His comments identify costs and efficiency as matters clients are considering, but do not say that cheaper models are causing data centres to be used less.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
The exact headline framing—that high data-centre and chip costs combined with cheaper AI models pose a risk to capacity usage—cannot be verified from these remarks. It should not be treated as a direct Parekh quote or an established conclusion about data-centre utilization.
How could cheaper models affect data-centre demand?
A lower-cost model can make AI economical for more tasks, while also reducing the compute needed for a particular task. Those effects can pull in opposite directions: greater efficiency may reduce resources per request, but lower costs may encourage more requests or new workloads. The net effect on infrastructure use depends on how demand grows, which models and chips are selected, and how workloads are deployed.
Infosys’s 2026 AI imperatives analysis frames inference economics around cost per task, workflow and autonomous loop. It also says the business effects of recent model-size innovations remain uncertain. That is a useful way to assess the issue: a cheaper token alone does not reveal whether a business uses less total compute, because the number and complexity of tasks may change too. Infosys Knowledge Institute’s AI imperatives analysis
What the chip outlook says—and what it does not
Infosys Knowledge Institute’s Semiconductor Industry Outlook 2026 describes continuing demand for AI servers and memory, alongside rising costs in parts of the chip supply chain. It projects that cloud service providers’ ASIC-based AI server shipments will grow approximately 45% in 2026, compared with 16% growth for GPU-based AI server shipments. These are report projections, not final shipment results or measurements of utilization.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →- The report states that DRAM revenue tripled from $50 billion in 2023 to more than $150 billion in 2025. This is the report’s historical estimate.
- It projects HBM’s share of DRAM to rise from 5% in 2023 to 50% in 2028.
- It discusses growth in GPU- and ASIC-based AI server shipments, AI-related memory demand, and higher costs in parts of the chip supply chain.
Together, those figures describe a changing and potentially expensive supply environment. They do not show that installed data-centre capacity is idle, nor do they establish that adoption of smaller models is causing a utilization decline. Infosys Knowledge Institute’s Semiconductor Industry Outlook 2026
Capacity utilization is not the same as staffing utilization
Infosys’s Q4 FY26 press-call transcript discusses AI services, business opportunities, competitive intensity and productivity. A separate transcript reproducer records the CFO saying staffing depends partly on business volume and utilization. That is a comment about workforce and company operations, not data-centre capacity. It cannot be used as evidence that servers or facilities are being underused. The SEC-hosted transcript is available in Infosys’s Q4 FY26 press-call filing.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What would establish a real capacity-usage risk?
To determine whether cheaper models and high infrastructure costs are creating a utilization mismatch, readers would need evidence that connects installed capacity to actual workload demand over time. Useful indicators would include utilization rates for the relevant facilities or accelerators, the period and geography covered, workload growth, and whether capacity is GPU-based, ASIC-based or otherwise allocated. The statements and projections cited above do not provide that measurement.
Model and hardware choices also cannot be ranked from the available material: it supplies no like-for-like benchmark. A meaningful comparison would need to account for workload fit, cost per task or token, inference efficiency, software and systems integration, data-residency requirements, and supply constraints—not just model size or chip price.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsQuick Recap
Best Value
- Family farms not data design for people against AI server farms, data center expansion, rural land buyouts, corporate agriculture, and industrial tech development replacing farmland and open space. Rural conservation and anti data center message.
- AI protest design for farmers, land conservation supporters, anti AI activists, sustainability groups, environmental advocates, rural communities, and people opposing server farm construction, power grid strain, and farmland destruction.
- Lightweight, Classic fit, Double-needle sleeve and bottom hem
What readers can conclude
- Parekh described smaller Infosys models as having smaller parameter sets, lower running costs and more efficient token use.
- He said clients were attentive to token costs, data sovereignty, memory-chip costs and efficiency.
- Infosys’s 2026 semiconductor outlook projects changing AI-server demand and reports strong memory-market growth, but these figures do not measure capacity usage.
- The available statements support a question about cost, efficiency and changing infrastructure needs—not a verified claim that cheaper models are lowering data-centre utilization.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
