Is Data Centre Infrastructure the Next Tech Bubble?

Infrastructure Leads the Charge in AI Tech …

Artificial intelligence (AI) has triggered one of the largest infrastructure investment cycles since the emergence of cloud computing. Around the world, billions of pounds are being committed to building AI-ready data centres, GPU clusters, high-speed networking and advanced semiconductor manufacturing capacity. Technology giants such as Microsoft, Amazon, Google and Meta continue to announce record capital expenditure, while a new generation of specialist infrastructure companies is racing to build the computing backbone that will power the next decade of AI innovation. For technology startups, this investment boom creates enormous opportunities. AI companies require unprecedented amounts of compute power to train and deploy increasingly sophisticated models, and infrastructure providers have become some of the fastest-growing businesses in the market.

Why AI Infrastructure is Growing So Quickly

Unlike previous software revolutions, generative AI demands vast amounts of physical infrastructure.

Every new large language model requires:

• Thousands of high-performance GPUs
• Massive quantities of high-bandwidth memory (HBM)
• Low-latency networking
• Advanced storage systems
• Reliable, low-cost power
• Purpose-built AI data centres
• Unending supply of clean water for cooling

This has created an entirely new ecosystem of infrastructure providers supplying the core layers of the ecosystem necessary to run the AI revolution.

Category Opportunities to Run Heavy AI

Many analysts continue to forecast exceptionally strong growth throughout the remainder of the decade, driven by enterprise AI adoption, sovereign AI initiatives and the emergence of AI agents that require substantially greater inference capacity than today’s applications. Several analysts expect AI infrastructure spending to continue expanding through at least 2029, supported by annual market growth rates exceeding 50% in some segments.

The core layers of the infrastructure ecosystem are:

• Semiconductors & Custom Silicon: Companies design advanced processing units, high-bandwidth memory, and electronic design tools that allow chips to execute parallel workloads.

• Power & Electrical Equipment: Grid components, heavy-duty transformers, substations, and precision power management systems feed electricity to high-density facilities.

• Thermal Management & Cooling: Liquid cooling (clean water) and thermal architectures dissipate the intense heat generated by high-performance server racks.

• Networking & Connectivity: High-speed copper and optical networking gear connect massive clusters of accelerators across data centres.

• Physical Construction: Specialised engineering firms construct the industrial real estate and data centre shells for major cloud providers.

The Companies Leading the Charge

Micron Technology

Unlike many AI infrastructure businesses, Micron sits further upstream in the supply chain. Its High Bandwidth Memory (HBM) products have become essential components within NVIDIA’s latest AI accelerators.

Demand for HBM currently exceeds supply, allowing Micron to benefit from:

• Higher selling prices
• Improved margins
• Long-term supply agreements
• Strong visibility of future demand

Provided AI investment continues, Micron appears well positioned to benefit from sustained demand for advanced memory technologies.

CoreWeave

CoreWeave has become one of the most closely watched AI infrastructure companies. Rather than building AI models itself, it provides specialised GPU cloud infrastructure used by AI developers, hyperscalers and enterprise customers. The company has secured large multi-year contracts and continues to expand aggressively into new international markets. Analysts remain broadly optimistic about its medium-term prospects despite recognising significant execution risk associated with its enormous capital expenditure programme. Recent analyst targets have implied meaningful upside from current trading levels, with several firms maintaining Buy or Overweight recommendations. However, investors should recognise that CoreWeave is also one of the most leveraged businesses within the sector, requiring substantial ongoing financing to support its rapid expansion. S&P Global expects continued exceptional revenue growth but also forecasts extremely high capital expenditure requirements through 2027.

Nebius Group

Nebius has emerged as another specialist AI cloud provider targeting organisations requiring large-scale AI compute. Analysts have highlighted its software layer, distributed GPU architecture and rapidly expanding data centre footprint as important competitive advantages. Bank of America and other firms have suggested Nebius could become a significant market share winner as enterprise AI adoption accelerates. Although smaller than some competitors, Nebius is viewed by many analysts as having considerable growth potential if it can continue executing its expansion strategy.

Applied Digital

Applied Digital occupies a slightly different position within the market. Rather than operating purely as an AI cloud provider, it develops high-performance data centre infrastructure designed specifically for AI workloads.

The company’s long-term success depends largely upon:

• Securing long-term customer contracts
• Expanding available power capacity
• Maintaining access to financing
• Delivering new facilities on time

As AI infrastructure demand grows, companies capable of delivering power-ready AI campuses may become increasingly valuable.

Core Scientific

Originally known for cryptocurrency mining infrastructure, Core Scientific has repositioned itself towards AI and high-performance computing. Its existing data centre footprint and electrical capacity provide a foundation that can potentially be repurposed for AI compute services. This strategic transition reflects a broader trend across the industry, where existing infrastructure operators are seeking to capitalise on the explosive demand for AI processing power.

What Analysts Expect by September 2027

While individual price targets continue to evolve with market conditions, there is a broad consensus among many analysts that AI infrastructure spending remains in its early stages.

Several common themes have emerged:

• AI compute demand is expected to remain supply constrained.
• Enterprise AI adoption should continue accelerating.
• Sovereign AI investment is increasing globally.
• GPU demand continues to exceed available capacity.
• Data centre construction pipelines remain exceptionally strong.

For companies such as CoreWeave and Nebius, analysts continue to forecast significant revenue expansion over the next 12–24 months, supported by multi-year customer contracts and substantial remaining order backlogs. Micron is expected to benefit from sustained HBM demand, while Applied Digital and Core Scientific are viewed as beneficiaries of continued AI data centre expansion. Taken together, the sector still appears to offer meaningful upside through approximately September 2027, assuming current AI investment trends continue.

Could This Become the Next Tech Bubble?

History teaches us that every major technology revolution will experience periods of excessive optimism. We have seen this cycle so many times that the pattern is unmistakable:

• The railway boom
• The dot-com bubble
• Solar energy
• Electric vehicles
• Cryptocurrency

In each case, the underlying technology ultimately proved transformative, but investor expectations temporarily became detached from commercial reality. The AI infrastructure market could eventually follow a similar pattern.

What Could Trigger a Slowdown?

Several risks deserve consideration.

• Oversupply: If infrastructure capacity begins expanding faster than demand, pricing power could weaken rapidly.

• Falling AI Compute Costs: Advances in AI model efficiency may reduce the amount of computing power required for training and inference. Better algorithms often reduce hardware requirements over time.

• Hyperscaler Self-Sufficiency: Companies such as Microsoft, Google, Amazon and Meta continue investing heavily in their own infrastructure. As these organisations become increasingly self-sufficient, reliance on specialist AI cloud providers could diminish.

• Capital Intensity: Perhaps the biggest risk is financial. Many AI infrastructure companies are investing tens of billions of dollars before generating corresponding cash flows. Should financing conditions tighten or customer demand weaken, heavily leveraged businesses may come under significant pressure.

When Might the Market Turn?

Predicting market cycles is almost impossible, but history suggests infrastructure booms rarely continue indefinitely. Current analyst expectations generally support continued strong growth through 2027, with some forecasts extending into 2028–2029.

Beyond that period, several scenarios become increasingly plausible:

• Growth begins normalising as supply catches up with demand.
• AI infrastructure becomes more commoditised.
• Profit margins compress.
• Investors shift their focus from revenue growth towards sustainable profitability.
• Equity valuations become increasingly sensitive to earnings rather than future potential.

In other words, the sector may not experience a dramatic collapse, but it could transition from today’s high-growth environment into a more mature and selective market.

Lessons for Tech Startups

For startups building AI applications, today’s infrastructure investment represents an extraordinary opportunity. Access to computing power is improving rapidly, new AI-native cloud providers are creating greater competition and innovation continues at remarkable speed. However, founders should avoid assuming current market conditions will last forever, because every technology cycle eventually matures. Businesses that build sustainable competitive advantages, rather than relying solely on favourable market sentiment, are those most likely to survive when investment cycles inevitably slow.

The End of the Beginning of a New Tech Cycle

Whether AI infrastructure proves to be the next technology bubble remains uncertain. The underlying demand drivers appear genuine, supported by accelerating enterprise adoption, sovereign investment and the growing computational requirements of modern AI systems. Some key infrastructure companies are all positioned to benefit if this investment cycle continues through 2027 and beyond. At the same time, investors should distinguish between the long-term success of AI and the valuations attached to individual infrastructure providers.

History shows that transformative technologies often outlive speculative investment booms. Even if AI continues reshaping every industry, not every company participating in today’s build-out will emerge as a long-term winner. For technology startups, the message is clear: take advantage of the unprecedented AI infrastructure now being built, but remember that sustainable businesses are created through sound commercial fundamentals, not by assuming every technology boom will last forever.

You may want to read: “Startups Must Understand Buyer Personas.”