Het nieuwste rapport van Huawei stelt 10 belangrijke richtingen voor Intelligent World 2035 voor

Huawei released its Intelligent World 2035 and Global Digitalization and Intelligence Index 2026 reports, projecting that global annual AI-token consumption will rise 100,000-fold by 2035, with agents accounting for more than 90% of traffic. The company forecasts AI will generate more than $27 trillion in cumulative economic value over the next five years, while global investment in digital and intelligent infrastructure is expected to grow at a 19.14% CAGR and exceed $4 trillion by 2030. Huawei outlined investment priorities spanning AI compute clusters, networking, cloud platforms, storage, data-center power and thermal capacity, autonomous driving, and agent security.

Analysis

This is a vendor-led roadmap rather than an order, capex, or product disclosure; it is not independently actionable for Huawei-linked revenue. The useful signal is thematic: agentic workloads shift bottlenecks from model training toward inference throughput, memory bandwidth, networking latency, power delivery, and cooling. That favors the AI infrastructure stack more than application software in the next 6-18 months, particularly suppliers with exposure to scale-out networking and data-center electrical systems.

The non-obvious constraint is that token growth does not translate linearly into semiconductor demand if enterprises route simple tasks to smaller models, cache outputs, or impose usage limits. Hyperscalers will optimize cost per inference aggressively; this can compress the value of commoditized compute while increasing the value of proprietary interconnect, memory, power-management, and liquid-cooling components. Broad infrastructure spending expectations also raise the risk of a 2027-28 digestion cycle if utility interconnection queues, transformer availability, or power prices delay deployments.

Near term, this release should not move liquid U.S. names materially. Over 1-3 months, monitor hyperscaler capex guidance and AI-cluster networking attach rates; those are the validation points, not industry forecasts. Over 6-18 months, a sustained migration to autonomous workflows would expand demand for ANET's high-speed switching, VRT's power/thermal equipment, MU's HBM and enterprise memory, and cybersecurity controls from PANW/CRWD as machine identities and agent permissions proliferate.

Contrarian view: the market already prices a broad AI capex supercycle, so generic exposure through SMH has asymmetric multiple-compression risk if inference monetization trails spend. The better relative-value expression is to own scarce physical bottlenecks versus GPU beta, while avoiding a directional trade solely on a promotional report.

Key Decisions for Investors

• No event-driven position on this release; treat it as a watch signal until a hyperscaler reports incremental inference-driven capex or networking attach-rate acceleration.
• Over 6-12 months, build a basket long ANET and VRT versus short SMH in equal dollar risk terms on AI-infrastructure pullbacks; thesis is that network/power bottlenecks retain pricing power while broad semiconductor exposure is more vulnerable to GPU digestion. Exit if ANET backlog growth decelerates for two consecutive quarters or VRT data-center organic orders fail to support guidance.
• Add MU only after confirmation that HBM supply agreements and data-center DRAM pricing remain firm through the next two earnings cycles; target a 12-18 month memory-bandwidth exposure, but avoid chasing if spot DRAM pricing weakens materially or inventory days rise.
• Use PANW or CRWD as a 6-18 month secondary beneficiary watchlist, not an immediate buy: agent deployment should increase identity, authorization, and audit requirements. Upgrade only if net-new ARR or platform adoption shows AI-security monetization rather than merely higher threat rhetoric.
• Monitor U.S. data-center power interconnection timelines, transformer lead times, and electricity forward prices quarterly. A meaningful deterioration is a signal to reduce VRT and AI-infrastructure longs before capex guidance resets.