Technology Trends 2020: IT, Hardware, Servers and Lessons

What changed in technology in 2020

2020 was a turning point not because brand‑new computers or a “new internet” appeared, but because the conditions changed abruptly. People and companies had to move work, education and services online quickly. IT shifted from being “support” to being the backbone of continuity.

For everyday users the changes were simple: a laptop and a stable connection became more important than ever. Home workstations were upgraded en masse, and requirements for video calls, audio quality and security grew. Many saw for the first time that a home router, a password and system updates affect work as much as an office PC.

For companies the changes ran deeper. Procurement and infrastructure planning stopped being calm: delivery times shifted, shortages became real, and demand for support and spare parts increased. At the same time the need for manageability grew: centralized deployment, device inventory, access control and fast incident response. Where supply chain transparency and local service matter, interest grew in predictable deliveries and on‑site support—not just the lowest price.

The purpose of this article is practical: what became critical, where mistakes were common, and which solutions passed the stress test.

Below are four layers that moved most in 2020: workflows and remote work, infrastructure (clouds, hybrid, networking), hardware, and security as daily discipline.

Remote work and the rapid rise in demand for digital services

Remote work was the main accelerator of the trends we later called the technological trends of 2020. Companies that had postponed online processes for years did the transition in weeks. The main goal was not “how pretty it looks,” but “it has to work every day.”

At first demand rose for video conferencing, corporate chats and collaborative document editing. Simple things quickly ran into practical details: employees have different devices, home internet can be weak, and some teams weren’t used to documenting decisions. Where short face‑to‑face chats used to help, clear communication rules and unified channels became necessary.

The biggest technical hit fell on VPNs and corporate networks. Hundreds of people connected at once, and what was previously considered “just in case” became critical infrastructure. Support loads increased: password resets, access setup, headset and camera problems, and security questions.

Budgetwise, 2020 was not just about subscriptions. Companies had to buy devices and peripherals quickly, expand VPN and network capacity, strengthen the service desk and publish simple user guides. In practice the most common purchases were laptops or compact PCs for employees who had only worked in the office, webcams and headsets, additional monitors and collaboration licenses.

A typical example: the accounting team moves to remote work and it turns out half the people are on personal computers and the VPN collapses at peak hours. Practical solutions won: standardize workstations, configure role‑based access and assign clear support procedures.

Cloud and hybrid infrastructure: where the focus shifted

In 2020 cloud moved beyond “someday” discussions. Companies addressed specific questions: how to quickly add capacity, how to provide remote access, and how to keep running under increased load. The emphasis shifted from ideal architectures to speed of change and resilience.

Organizations moved to the cloud what was easiest and fastest to show benefit: corporate email and collaboration, backups, test environments, websites and seasonal services. Critical databases and core accounting systems often stayed on‑prem: migration takes time, requires reworking integrations and a clear risk management plan.

The hybrid model became a compromise: some services in the cloud, some on company servers. This lets you keep sensitive data and latency‑sensitive systems local, while using cloud resources for customer portals, analytics, development and rapid scaling. For many government organizations and large enterprises this also meant clearer control over where data is stored, who administers it and how regulations are met.

Containers and microservices were widely discussed in 2020 because they helped ship changes faster and move applications between environments. Simply put: a container packages an app with everything it needs to run the same way on different servers. Microservices break a large product into smaller parts so they can be updated independently. But this didn’t fit everywhere: small teams sometimes found it easier to improve a monolith than maintain dozens of components.

Observability became more important. When part of the system runs in the cloud and part in your data center, and users are remote, you can’t tell what’s failing by sight. Metrics (what’s happening with load and latency), logs (why an error occurred) and alerts (when to act), plus unified dashboards, became the norm.

Example: a company keeps its accounting system and domain on its own servers and puts the customer portal and dev environment in the cloud. Without metrics and logs incidents turn into finger‑pointing. With observability you can see whether the bottleneck is the link, the database, the app or an overloaded server.

Computer hardware in 2020: demand, shortages and practical choices

In 2020 it became clear that hardware is not just a consumable but part of resilience. When thousands of employees moved home at once, demand for basic items soared and supply chains faltered. So the technology trends of 2020 looked less like new gadgets and more like a fight for stock and reliable lead times.

The year’s main pain point was shortages and unpredictable deliveries. Companies that used to refresh equipment quarterly began planning six months out and keeping spares. Winners were often those who kept several model alternatives and didn’t lock themselves to a single brand or configuration.

Laptops and monitors were updated most because they determined whether a person could work a full day at home. Next came peripherals and networking: headsets, keyboards and mice, plus routers and access points. Even “small” items like webcams and decent microphones became more important than incremental CPU speed.

Purchases prioritized practical parameters: a decent camera and mic for calls, reliable Wi‑Fi (often module quality and 5 GHz mattered more than advertised top speeds), battery life, comfortable keyboard and screen, plus service and spare parts availability.

Workplace needs changed too. Home required basics: a second monitor, a comfortable chair, good lighting, and sometimes a compact PC instead of a personal laptop. Offices were rethought: more shared meeting kits and cameras for video conferencing and less dependence on a dedicated workstation.

Servers and data centers: what business and government demanded

Server loads rose sharply and sometimes unexpectedly in 2020. Video conferencing, VPNs, remote work, electronic queues, citizen and client portals—all became primary channels rather than extras. Many discovered bottlenecks weren’t always in internet capacity: older storage, insufficient memory and weak virtualization CPUs were common limits.

Configurations shifted toward “more of what helps parallelism”: more cores, more RAM and faster disks. NVMe was especially valued where many small operations occurred (VMs, databases, mail). Even without increasing server count, role separation (dedicated nodes for virtualization, others for storage and backups) delivered large gains.

Virtualization made it faster to allocate resources to teams. When a department needed a service “yesterday,” spinning up a VM or container was easier than waiting for procurement and physical installation. In 2020 deployment speed often mattered more than a perfect architecture.

Reliability stopped being optional. Business and government increasingly required services that run without interruption and manageable failures. The minimum that became standard: redundant power and network paths, regular backups with a clear recovery plan, monitoring (to spot issues before users complain), and accessible support and spare parts.

A typical scenario: a customer portal slows down every morning. CPU is OK, but virtualization hosts are memory‑constrained and the disk subsystem can’t handle the peaks. Servers with larger RAM pools, fast disks for the busiest databases and a standby node often resolved the issue without a full cloud migration.

Cybersecurity in 2020: attacks, remote work and basic hygiene

Security became a louder conversation not out of fashion but because remote work expanded attack surface quickly. Companies opened remote access to internal systems in weeks, which broadened attack vectors more than any planned change.

The most common remote‑work problems were mundane, not exotic: weak or reused passwords, phishing emails about “urgent updates,” personal devices without protection and home routers left at factory defaults.

Zero Trust in plain terms is the “never trust by default” approach. Previously, once a user was inside the network they were often trusted widely. In 2020 it became clear every access to a service or data should be verified, even if the user is already “inside.”

The good news: the first improvements don’t require complex architectures or big budgets. Quick gains usually came from getting basics in order:

• MFA for email, VPN, admin panels and critical cloud services.
• Regular OS and application updates, especially browsers and office suites.
• Backups with verified restores (not just a checkbox that a backup exists).
• Privilege limitation: fewer local admins, separate admin accounts.
• Short anti‑phishing training with real examples.

Practical scenario: an accountant working from home gets a fake “IT” email and enters credentials on a spoofed page. With MFA enabled and role‑based restrictions on financial systems, damage is often contained. If backups are tested, recovery from ransomware is possible without weeks of downtime.

The core lesson was simple: security is not a one‑off project but daily discipline.

AI and analytics: where real demand appeared and where there was noise

In 2020 many companies used AI not as an experiment but to solve urgent tasks. Real demand was strongest where it reduced human load and enabled data‑driven decisions: chatbots for routine queries, support request analysis, demand and inventory forecasting, and basic automation of document checks.

There was more hype around a universal AI that would replace entire processes. In practice simple, well‑measured scenarios won: reducing response time, lowering call center load, preventing stockouts, or quickly spotting anomalies in data.

The main limiter was data, not model choice. If requests come in different formats, there’s no single customer registry, and events aren’t logged, even a trendy neural network will give poor results. Many followed a clear sequence: first bring order to data sources, then automate.

A separate topic is accelerators and GPUs. They are necessary for heavy training (for example, computer vision) and intense analytics loads. But for many 2020 tasks CPUs and smart optimization were sufficient: request classification, simple forecasts, anomaly detection.

To avoid months of work, a practical approach prevailed: a 4–8 week pilot with clear rules and metrics. Define success up front (response time, forecast accuracy, share of automated handling), data quality requirements, total cost of ownership and a clear rollout plan.

How to apply 2020 lessons: a step‑by‑step plan for IT leaders

2020 showed winners were those with clear priorities and buffers in people, capacity and security. It helps to act not by gut feeling but with a short plan.

Five quick steps to bring order

1. Map critical processes. List 10–15 services without which work stops (email, accounting, file access, telephony, key databases). Note an owner, acceptable downtime and dependencies (internet, VPN, server, licenses).

2. Assess real load. How many users work from home, which apps are “heavy,” when are peaks (month‑end, citizen services, exams). A common 2020 mistake was to plan by average instead of by maximum.

3. Choose a hosting model per service, not one model for everything. Systems that need strict control and predictable latency are often easier to keep on‑prem. Services that must scale quickly are better in the cloud. Many organizations ended up with a hybrid approach.

4. Define a basic security and backup minimum that works for remote work: MFA, updates, least‑privilege, backups with restore checks, and separate admin accounts.

5. Create a 12–18 month refresh and procurement plan for PCs and servers. Prioritize what to replace first (workstations for video calls, servers for virtualization, network) and include lead times.

A small example

If remote workstation load has grown, it makes sense to update part of the PC fleet and recalculate server capacity simultaneously. In Kazakhstan this is often combined with requirements for local content and clear service to avoid long lead times and random configurations.

Common 2020 mistakes that were costly

2020 showed problems often came not from complex technologies but from small things that were postponed for years. Firefighting responses led to downtime, overspending and data loss.

The most common story was procurement. Laptops, monitors, webcams and even switches became scarce and lead times lengthened. Organizations that didn’t verify actual load and keep spares ended up with a “zoo” of models, different chargers, mixed performance and expensive support.

Five frequent mistakes:

• Buying PCs and servers without a simple calculation: how many users, what tasks, what headroom for capacity and delivery times.
• Postponing VPN and MFA, then closing gaps after an incident.
• Migrating systems to the cloud without fully understanding data, latency, integrations and access rules.
• Skimping on backups and never testing restores until a failure or ransomware event.
• Choosing suppliers based only on price without assessing service, spare parts and realistic support.

A telling scenario: accounting and sales moved to remote work, access was given through a single shared VPN without MFA. A month later an account leaked, an attacker accessed the network and touched a shared file resource. A backup existed but restore had never been tested and some archives couldn’t be opened. The result: downtime, manual reconciliation and urgent purchases.

Another lesson of 2020: a service network and clear support became part of resilience, not an optional extra.

Quick checklist: company readiness after 2020

IT isn’t sustained by admin heroics but by basic settings and clear processes. If you treat the 2020 trends as lessons, start with a self‑check.

Check the basics

Answer five questions honestly. If you’re unsure about at least two, you have risk.

• Is there an up‑to‑date inventory: laptops/PCs, servers, network gear, licenses, user accounts and access rights?
• Is MFA enabled where possible and are OS and app updates enforced and verifiable on workstations and servers?
• Are backups in place with tested restores and clear RPO/RTO objectives?
• Have peak‑time bottlenecks been checked: internet link, Wi‑Fi, VPN, storage, performance monitoring?
• Is there a procurement and repair plan with delivery buffers and a list of “what’s critical” vs “what can be temporarily replaced”?

If time is short

Practical order: fix access and MFA first, then updates, then backups with a test restore, and only after that expand channels and procurement.

Example: if VPN is failing, people send files via messengers and backups haven’t been checked for months, closing basic security and backup holes can be cheaper and faster than buying new hardware immediately.

A real example: how a mid‑sized organization survived 2020

An organization with 350 employees: a head office, a couple of branches, a small server cluster and a 6‑person IT team. Before 2020 only a few specialists worked remotely; most processes relied on office networks and standard PCs.

Within two weeks it became clear the problem wasn’t video calls but access and bottlenecks. VPN collapsed under connection load, some staff used home PCs, and there weren’t enough corporate laptops. Support was overwhelmed with repetitive requests and the business demanded “like before, but from home.”

They made a practical 1–2 month plan. Standardized remote workstations (same software set and password policies, a short instruction). Unloaded access: split traffic, tightened rules on who connects to what, and added monitoring. Scaled server resources for file services, mail and accounting with some headroom.

Results were measured by key service uptime, support response time, number of security incidents and share of standardized workstations.

The main conclusion from 2020 trends: those who simplified and standardized quickly won. The rule became: fix basic resilience first (access, security, monitoring, capacity), then experiment.

Next steps: turn 2020 trends into a clear plan

To avoid endless discussions, fix a 90‑day plan and assign owners.

Focus on three priorities: workstations (where downtime occurs), infrastructure (network, servers, storage and backups) and security (access, updates, endpoint protection and user training).

90‑day mini plan

• Approve 3–5 user scenarios (office, remote, engineer, manager, operator).
• Assign standard PC and all‑in‑one configurations per scenario, and 1–2 server profiles for your workloads.
• Define access rules: MFA where possible, separate admin accounts, ban shared passwords.
• Verify backups: what is backed up, how often, and has a test restore been performed?
• Introduce basic update and inventory control: what is installed and who is responsible for patches.

After standards, plan support. In 2020 many budgets went to urgent purchases, but without a defined service level and repair readiness downtime became more expensive than the equipment. Decide in advance which components must have spares (for example, drives and power supplies) and who services regional sites.

If local content and supply chain transparency matter, include it in requirements. In Kazakhstan some organizations rely on domestic manufacturers and system integrators with local production and service networks, for example GSE.kz, to close needs for workstations and server infrastructure faster.