A hard operating-system deadline and the rapid rise of on-device AI are fueling the biggest enterprise PC refresh cycle India has seen since the pandemic. Commercial shipments jumped 9.5% year-on-year in the second quarter of 2025, with the enterprise segment soaring 21.2%, according to IDC. Canalys projects further commercial PC growth of 7.9% in 2025 and 9.8% in 2026. Behind the numbers lie two interlocking forces: Microsoft’s looming end of support for Windows 10 on October 14, 2025, and a surge in demand for AI-equipped laptops that bundle neural processing units (NPUs) as standard features.

The upgrade push is not being driven by any one technology but by a convergence of factors that turn a routine refresh into a strategic modernisation event. Organisations are spending not just to replace aging hardware but to embed AI readiness, improve security postures, and future-proof their fleets.

Market Numbers: Commercial Shipments Surge

IDC’s latest tracker shows India’s commercial PC segment expanded 9.5% YoY in the April–June quarter. Enterprise demand—the slice coming from large private-sector organisations—leapt 21.2% YoY in the same period. Those figures are echoed by Canalys, which forecasts commercial PC growth of 7.9% in 2025 and 9.8% in 2026, underlining that the engines of expansion are not consumers but businesses, governments, and educational institutes.

Consumer demand, by contrast, is softening or flat. The value of shipments is rising even as unit growth moderates, because organisations are buying premium-configured, AI-capable devices. “Enterprise investment in high-performance devices continues, supporting hybrid work and cloud adoption,” said Prabhu Ram, Vice President, Industry Research Group at CyberMedia Research (CMR). “SMBs and larger enterprises remain central to this momentum, propelled by large-scale device refresh cycles tied to Windows 10 end-of-life.”

Windows 10 Deadline Looms: A Hard October 14 Stop

Microsoft will halt mainstream security updates for Windows 10 on October 14, 2025. After that date, any device still running the operating system will be exposed to unpatched vulnerabilities unless the organisation buys Extended Security Updates (ESU)—a costly, temporary fix that does not preserve driver support or long-term compatibility.

That single date has become a forcing function. Large enterprises typically refresh hardware every 3 to 3.5 years, and many of the machines bought during the 2020–21 remote-work spike are now due for replacement. The calendar alignment of refresh cycles with the Windows 10 end of support is creating a concentrated buying window that vendors and channel partners are aggressively pursuing. “We saw a big spike in PC procurement in 2020-21 when remote working became the norm,” said Bharath Shenoy, Principal Analyst at IDC India. “Usually, for enterprises 3 to 3.5 years is the upgrade cycle time. So this refresh cycle is falling at the same time as Windows 10 support ends so they are buying new PCs.”

IT teams facing the migration are choosing among three paths: upgrade existing machines to Windows 11 when hardware compatibility permits (TPM 2.0, supported CPU), buy new Windows 11 or Copilot+ PCs, or enroll in ESU for a small subset of devices that cannot be modernised in time. Most large organisations are blending all three approaches.

AI Laptops: From Niche to Near-Standard

The other major driver is the rapid normalisation of AI laptops—notebooks that integrate a dedicated NPU alongside the CPU and GPU. NPU performance is measured in trillions of operations per second (TOPS). Basic AI notebooks with 10 TOPS and above accounted for 88.1% of AI notebook shipments in India during the first half of 2025, IDC reports. Although advanced AI laptops with 40 TOPS and above still represent a small slice, the trend is unmistakable: entry-level NPUs are becoming table stakes in commercial SKUs.

On-device AI appeals to enterprises for four reasons: latency and privacy, since inference runs locally without cloud round-trips; cost control, because high-frequency features don’t accumulate cloud inference bills; productivity gains from real-time summarisation, transcription, and email triage; and a stronger security posture when hardware-backed protections like TPM and Secure Boot ship alongside the NPU.

Many procurement teams now specify AI-capable hardware as a requirement rather than an option. “Enterprise share is on the higher side, because bigger enterprises such as global MNCs and Indian IT, ITES majors who have bigger spending capacity and want future-ready devices, are procuring AI PCs,” Shenoy noted.

Yet marketing numbers can mislead. TOPS is a simplified metric that says little about real-world performance on enterprise LLMs or multi-stream inference tasks. Workload characteristics, model precision, memory bandwidth, and software maturity vary widely across platforms. “Many companies view AI as a value-add rather than a core requirement, with security, reliability, and overall performance continuing to drive buying decisions,” CMR’s Ram cautioned.

Silicon diversity is exploding: Apple’s M-series Neural Engine, AMD’s Ryzen AI (high-end parts exceeding 40–50 TOPS), Intel’s Core Ultra line, and Qualcomm’s Snapdragon X Elite/X Plus all compete in the AI laptop space. Each brings its own NPU architecture, developer toolchain, and integration complexity.

Vendor Dynamics: HP Leads, Apple Gains Ground

Traditional Windows OEMs dominate enterprise volume, with HP holding a commanding 35% market share in India’s commercial PC segment in Q2 2025, according to IDC. Lenovo, Dell, and Acer also have deep channel relationships and supply contracts with large IT services firms and Indian multinationals.

But Apple is carving out a growing enterprise footprint. Canalys data shows Apple shipped 236,000 macOS PCs in India in 2024 at 23% YoY growth, with more than 65% of those driven by enterprise demand. MacBooks with in‑built AI chips are being procured in higher volumes by large organisations and Global Capability Centres (GCCs). Canalys estimates Apple could double its enterprise MacBook business by the end of 2025 and is on track to ship close to 283,000 MacBooks to India’s commercial segment by 2027.

Aggressive pricing on older M‑series models is accelerating adoption. Zoho, the SaaS company, reported that 65% of its employees prefer MacBooks, that its IT team receives 50% fewer support requests for Macs than Windows PCs, and that developer build speeds have improved. Such case studies are compelling, but IT leaders must validate them in their own environments.

Simultaneously, government “Make in India” incentives are pushing OEMs to assemble more models locally, which can affect price competitiveness in government tenders and large enterprise deals that favour local sourcing.

Procurement Strategies: Refresh, Upgrade, or Extend?

Fleet replacement decisions now involve more than hardware specs. Enterprises weigh:

  • Upgrade existing devices: For machines that meet Windows 11 requirements, this preserves capex but leaves AI capabilities on the table.
  • Purchase new Copilot+ PCs: A capex-heavy approach that delivers modern NPUs, Secure Boot, and TPM 2.0 in one stroke.
  • Enroll in ESU for Windows 10: A stopgap that bridges timing gaps but defers migration work and raises total cost over time.

Vendor financing, seasonal sales, and localised discounts are reshaping the arithmetic. In some deals, discounted older‑generation premium devices—like M1 MacBooks—become price‑competitive with new non‑AI Windows laptops, distorting traditional TCO comparisons. IT teams are increasingly factoring in soft benefits such as reduced support tickets and expected cloud‑inference savings, but those benefits are not yet standardised across fleets.

Risks and Missteps: Beyond the TOPS Hype

The buying wave carries its own pitfalls. Over‑reliance on TOPS can lead organisations to acquire hardware that looks powerful on paper but underperforms on their actual AI workloads. Running large numbers of unpatched Windows 10 devices after October 14 represents a significant compliance and security risk, and ESU programs do not solve underlying compatibility or driver gaps.

Multiple NPU architectures create a testing burden for ISVs and SRE teams. Standardising on one platform simplifies operations but risks vendor lock‑in or missing competitive features available on another. Aggressive discounting of older premium models can pull fleets into mixed‑OS environments that increase management overhead.

Then there is the skills gap. On‑device AI adds operational complexity: configuring secure AI model deployment, managing model lifecycles, harmonising endpoint telemetry with observability, and ensuring privacy‑preserving inference are nontrivial. Many IT teams are still building these competencies.

What IT Leaders Should Do Now

The windowsforum community highlights a practical checklist for CIOs:

  • Run targeted proofs of concept on candidate NPU platforms using representative enterprise AI features, not just benchmarks.
  • Conduct an inventory and readiness audit: identify devices that can run Windows 11 and prioritise replacements where TPM, Secure Boot, and modern firmware are missing.
  • Understand AI feature architecture: confirm whether on‑device features run fully locally, partially in the cloud, or as hybrid flows—each model carries different compliance obligations.
  • Validate management and MDM support for any planned macOS expansion; ensure existing tooling can handle mixed fleets at scale.
  • Model total cost of ownership with support‑ticket deltas, expected cloud inference savings, training, and integration costs—capex alone is an incomplete picture.

A Strategic Modernisation Window

The intersection of a mandated OS lifecycle event and the rise of AI‑capable hardware creates a rare inflection point in India’s PC market. Fleets refreshed with modern silicon, built‑in NPUs, and updated security baselines can unlock productivity gains, reduce certain recurring cloud costs, and improve the user experience. Vendors and channels are responding with diverse product mixes and financing options that make upgrades easier to achieve within budget cycles.

Yet the cycle also carries risk: marketing metrics such as TOPS are imperfect proxies for real‑world capability, the skills and governance required to operate AI at the endpoint are still nascent inside many IT organisations, and short‑term discounting is producing mixed TCO signals that could lead to fragmented estates and higher operational complexity.

The prudent path for enterprise CIOs and procurement teams is clear: move decisively, but move with evidence. Validate AI features against actual workflows, prioritise security and manageability, and be prepared to pay attention to lifecycle, integration, and support economics—because the machines bought today will define how businesses use AI and secure data for years to come.

Quick Facts

  • Windows 10 end of support: October 14, 2025. Mainstream security updates cease unless ESU is purchased.
  • Commercial PC growth: IDC Q2 2025: commercial 9.5% YoY, enterprise 21.2% YoY. Canalys forecast: 7.9% in 2025, 9.8% in 2026.
  • AI notebook adoption: 145.2% YoY growth in H1 2025; basic AI notebooks (≥10 TOPS) were 88.1% of that total.
  • TOPS thresholds: Entry‑level AI laptops start at ~10 TOPS; advanced Copilot+ experiences typically require ≥40 TOPS.
  • HP leads commercial: 35% share in Q2 2025. Apple shipped 236,000 Macs in 2024, 65% to enterprises, and aims to double enterprise MacBook business by end‑2025.