Recent AI deals recall 1990s telecom network deals
Key players in the AI ecosystem, including AI model developers, hyperscalers (large data centres) and semiconductor chip companies, have recently announced large partnerships. The deals reflect the race to meet exploding demand for computing power and AI leaders are coordinating across the value chain in an attempt to ensure that supply keeps up with the speed of innovation.
The scale and circular nature of these commitments, where suppliers, customers and investors overlap, has prompted comparisons to the vendor-financing loops of late-1990’s tech bubble, which featured similar deals related to the buildout of internet and telecommunications infrastructure. At that time, many telecom equipment makers extended loans or equity stakes to their own customers to finance network expansion. The practice boosted reported sales and demand on both sides. However, when credit tightened, those circular flows unraveled, contributing to the sector’s collapse and broader dot-com crash.
Focusing on fundamentals offers a different perspective
While some caution is warranted, we think the real question is not whether today's deals resemble the dot-com era, but whether the underlying fundamentals do.
We identify three key distinctions:
1. Robust balance sheets
In the 1990s, much of the buildout was financed by companies with limited profitability and relied heavily on external capital. Today's wave is largely funded from hyperscalers that have robust margins and free cash flow. Considering that many past bubbles burst as credit conditions tighten, this buildout appears more resilient to that kind of stress.
In a further contrast with the dot-com bubble, capital is chasing AI, not the other way around—AI has captured roughly half of venture capital dollars this year1 — and spending is anchored in physical infrastructure like chips, electrical equipment and data centers.
2. AI revenue momentum
Whereas early internet firms built first and monetised later, AI is monetising as it builds. Hyperscalers are already generating returns through increased cloud demand and productivity gains in coding, advertising and enterprise tools. The model developers have more nascent business models, but with 99% of US market share in global LLMs2, revenues are growing. Meanwhile, more businesses are beginning to integrate AI. KPMG’s latest AI Survey shows average AI investment by companies rising 14% from 1Q 2025 to USD 130 million, supported by visible productivity and profitability gains from AI use cases.3
3. Demand outpacing supply
Any wave of heavy capital investment runs the risk of overbuilding. Large overcapacity in the fibre-optic network at the peak of the dot-com era, for example, took many years to absorb. In contrast, data centre vacancy rates are currently at record lows.4 Demand for computing power continues to far outpace supply—more data has been created in the last three years than in all of history,5 and AI workloads are growing significantly.
AI infrastructure is running near full capacity, unlike the idle fiber-optic networks of the early 2000s
Lessons from history
Some caution around AI is warranted. The scale of spending is enormous, the pace unprecedented and some assumptions around return on investment (ROI), such as the useful lives of assets, remain open questions. History reminds us that enthusiasm can run ahead of reality.
However, today’s players have far strong financial foundations than those of the dot-com era, while the monetisation of AI is already underway and the risk of overbuilding seems limited in the near term. As this story unfolds, we believe investors should be selective and use active management to help separate transformative winners from bubbly valuations.