InfoHub Logo

Kazakhstan is gearing up for a decisive AI push in 2026, with experts saying nationwide impact hinges on three upgrades: a shift to IPv6, genuine 5G rollout, and better data governance. The stakes are high—without modern networks and clean data, AI remains stuck in pilots rather than powering government, business and healthcare.

This was reported by the Infohub.kz news agency.

2025 laid the groundwork

In 2025, Kazakhstan formalized its AI agenda: a dedicated ministry was established, an AI-related legal framework was adopted, and a domestic supercomputer came online. But IT veteran Shavkat Sabirov, who has more than 30 years of industry experience, argues that true results depend on the infrastructure reset scheduled for 2026.

He notes that today’s AI use is largely generative, while full-scale deployment across the public sector, enterprises, healthcare and mobile services requires next‑gen networks and robust data pipelines.

Why 2026 is pivotal: IPv6, true 5G and bigger data plans

Kazakhstan still largely operates on IPv4, while AI and big data demand the expanded address space of IPv6. According to APNIC, roughly 43–48 percent of countries already use IPv6. The e‑government operator NIT JSC has begun discussing business tariffs, service pricing and access terms for the national supercomputer to accelerate migration. With coordination across state systems, the transition could speed up—where the role of national operator Kazakhtelecom will be decisive.

On mobile networks, Sabirov expects 2026 to bring a real 5G experience. He cautions that a “5G” symbol on a phone doesn’t guarantee full 5G performance; with true 5G, today’s 15–20 GB monthly plans could vanish in a day. He anticipates operators will move toward much larger data bundles, on the order of 256–300 GB per month, alongside significant upgrades to transport networks and core equipment.

From imported tools to local, sector‑specific solutions

AI expert Aigerim Kairzhanova says the initial hype ebbed in 2025, giving way to systematic work: processes were formalized and a profile ministry began coordination. Many deployments still adapt foreign tools, but tailored solutions for industry, manufacturing and energy remain scarce.

Her forecast for 2026–2027: data quality becomes the main battle. That means cleaning registries, normalizing databases, closing gaps and formulating precise tasks. In her view, the model matters less than the data that feeds it.

Compute, energy and the hardware pipeline

Scaling AI is constrained by the cost and availability of infrastructure. Renting GPU/TPU clusters is expensive, and serious AI workloads require data centers, electricity, water and high‑capacity networks. Without this ecosystem, projects stall at pilot stage.

There are promising signs: the Alem Cloud supercomputer (from Alem AI) has launched, strong local teams are emerging, and modern accelerators—such as Nvidia H200, as well as TPU and Habana—are being shipped and integrated with higher‑throughput interconnects. These moves suggest a shift from scattered pilots to a more systematic build‑out.

Timeline to a digital state

Sabirov estimates that with clear KPIs in state programs, Kazakhstan could complete up to 90 percent of infrastructure renewal and fully embrace IPv6 by 2030. He cites Alatau City as a candidate to become a fully digital city within three years.

Kairzhanova is more bullish: by 2028, she believes Kazakhstan could credibly call itself a “digital state.” AI is already in use across banks, government services, call centers and chatbots. Healthcare shows momentum in radiology—algorithmic support for X‑ray analysis could be scaled next. Efficiency gains are also clear in industry and oil and gas.

Headwinds: bureaucracy, fiscal pressure and security

Sabirov points to red tape as the chief drag—slow approvals dilute speed. He also highlights fiscal pressure on businesses and says incentives are concentrated around Astana Hub, leaving small and midsize firms in need of nationwide support.

Kairzhanova notes that stringent cybersecurity and data‑sovereignty rules limit the use of foreign public clouds for state data. Major vendors like Microsoft offer sovereign or dedicated cloud options and assume security obligations, but broader adoption requires trust, legal clarity and mature practice.

Jobs and skills: the new baseline

Some roles will be reduced, Kairzhanova says, but new positions will appear. The bar is rising: professionals will need both domain expertise and fluency with AI tools. Education programs are likely to shift accordingly.

Digital literacy should start in schools, along with training on neural networks, ethics and critical thinking. AI is a partner, not an oracle—users must verify outputs and understand the technology’s limits.

The bottom line for 2026

Expect returns on AI investment to show up first in cost reductions, faster workflows and better decisions. The winners will be organizations that integrate AI systematically across industry, logistics, finance, the public sector and media—and that secure compute, power, cloud infrastructure and talent.

President Kassym‑Jomart Tokayev underscored the priority in an interview with Turkistan, calling digital technologies and AI national imperatives and framing this year as a springboard for faster change.