UBS Global Technology and AI Conference: AI demand is "insatiable"! AMD proclaims this is only the second year of a "ten-year super cycle."

According to the latest summary from the UBS Global Technology and AI Conference, leading companies believe that the AI industry is currently in the second year of a ten-year super cycle, with demand exhibiting "greedy and continuously growing" characteristics. The main issue now lies with supply-side bottlenecks, rather than insufficient demand.

According to Wind Chaser Trading Desk, on December 4th, UBS stated in the conference notes that for investors, this indicates that the strong growth in the AI sector is far from over. From chip design to cloud computing infrastructure and data center construction, the entire industry chain is experiencing several years of sustained high prosperity. Frontline manufacturers generally believe there is no "bubble" in the market, but are striving to fulfill long-term orders that have already been placed.

The outlooks shared by key companies at the conference—AMD, CoreWeave, and Nebius—especially validated this trend, showcasing astonishing growth expectations, billion- and even ten-billion-dollar cooperation orders, and strong confidence in explosive demand continuing over the coming years.

Emerging cloud service (Neocloud) and infrastructure experts pointed out that, as model sizes expand, post-training applications become widespread and inference tasks consume more compute resources, demand for AI is surging rapidly.

The visibility of data center connection demand has expanded from the traditional one quarter to one to three years, highlighting that the whole industry is preparing for long-term growth. Traditional server demand is also accelerating due to the rise of AI agent workloads. Overall, the market consensus is that the current challenge is not finding demand, but rather how to overcome supply constraints—from power and labor to advanced components—to deliver computing power faster.

AMD: We Are in the Second Year of a Ten-Year Super Cycle

AMD CEO Lisa Su expressed extreme optimism about the AI cycle at the conference. She clearly stated that the company does not see a "bubble" currently, but instead views AI as the "second year of a ten-year computing super cycle."

Ambitious Growth Targets: AMD recently raised its compound annual growth rate (CAGR) forecast from over 50% in prior years to over 60%. The company expects its total addressable market (TAM) to reach $1 trillion by 2030, and aims to capture more than 10% market share. Besides AI GPUs, AMD’s revenue share in the server CPU market has grown to 40% and continues to rise.

Huge Validation from OpenAI Partnership: AMD’s multi-generational 6GW partnership agreement with OpenAI is strong evidence of its product competitiveness. Lisa Su explained that each GW of compute collaboration equates to "tens of billions of dollars in sales." This partnership not only brings AMD huge order volumes, but also deeply involves OpenAI in AMD’s technological success, providing powerful endorsement to other AI-native customers and hyperscale cloud providers.

Advancing Full-Stack Solutions: Through the acquisition of ZT, AMD is building a full system-level solution. The company plans to launch the MI450 rack in 2026, delivering full-stack capability from chip to system, and is dedicated to building an open ecosystem.

Strong Confidence in Supply: AMD emphasized that it has established deep partnerships with TSMC, memory and packaging partners, and is confident it can meet strong demand and achieve growth targets.

CoreWeave: Demand Is "Greedy", The Challenge Is Faster Delivery

As a representative of emerging AI cloud providers, CoreWeave describes current market demand as "greedy" and "relentless".

Unstoppable Demand Growth: The company believes demand surge is driven by three factors: 1) model scaling laws remain effective; 2) explosive growth of compute-intensive post-training applications; 3) mainstream adoption of more compute-hungry inference models. CoreWeave currently has a backlog of $55 billion in contracted orders.Firm Denial of "Overbuilding": CoreWeave strongly refutes market concerns of overbuilding AI infrastructure. The company stresses they are building based on customers’ committed five-year usage, essentially “producing to order” and still racing to catch demand. As concrete evidence, one cluster with over 10,000 H100 GPUs renewed contracts for a longer term at prices only 5% lower than a few years ago, proving the lasting value and returns of premium compute power.Nvidia Ecosystem Dominance: Despite market hopes for diversification, CoreWeave admitted that demand on its platform “overwhelmingly points to Nvidia GPUs.” If client demand changes, it may consider other options, but for now all signals still point to greater demand for Nvidia GPUs.Robust Financing Model: CoreWeave funds its expansion through asset-level term loans, with 60% of current customer orders coming from investment-grade counterparties, greatly enhancing its financial stability.

Nebius: We Are Not Just GPU Lessors, But the "Hyperscale Cloud Provider" of the AI Era

Nebius, spun off from Yandex, displays its ambition to become a full-stack AI “hyperscale cloud provider”, not just a provider of GPU-as-a-service.

Stunning Demand Growth Rate: Nebius observes some client demands doubling every 6-8 weeks. Its Q3 2025 business pipeline grew 70% quarter-on-quarter, creating $4 billion in new business. This strongly indicates the industry is not overbuilt, and inference demand is growing even faster than capital expenditures.

Massive Microsoft Partnership: Nebius signed a five-year $19 billion agreement with Microsoft, primarily for supporting Microsoft’s CoPilot business and new model development, rather than competing with Azure’s core cloud business. This provides top-tier endorsement for Nebius’s scale and technological strength.

Clear Path to Profitability: Nebius is focused not just on growth, but also on profit. The company expects long-term EBIT margins to reach 20–30%. Its path includes: self-developed racks and full-stack optimizations to achieve 20% cost savings, providing high-premium cloud services, and leveraging its engineering team for operational scale. For accounting, Nebius uses a four-year depreciation schedule for Hopper GPUs, which is more conservative than peers’ six-year schedules.

Flexible Capacity Expansion: Nebius has raised $8.5 billion to expand data center scale. Its strength lies in flexible project management, enabling simultaneous advancement of multiple data center projects, effectively solving bottlenecks in delivering AI computing power.

~~~~~~~~~~~~~~~~~~~~~~~~

The above exciting content comes from Wind Chaser Trading Desk.

For more detailed analysis, including real-time interpretation and frontline research, please join the [Wind Chaser Trading Desk · Annual Membership]

Risk Warning and DisclaimerThe market carries risks, investment requires caution. This article does not constitute personal investment advice, nor does it take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article fit their situations. Investment based on this is at your own risk.