AI Infrastructure Arms Race: Oracle Exec Joins TCW LIVE Panel

The AI infrastructure arms race took center stage at TCW LIVE on September 9, 2026, when Oracle’s Dr. Sanjay Basu—senior director of GPU & Gen AI Solutions and Services—joined a high‑profile panel hosted by Connectbase and InterGlobix. The discussion, titled “The New Digital Infrastructure Arms Race for AI Compute: Data Center Providers, Hyperscalers, and Neocloud Providers,” dissected how soaring demand for AI workloads is reshaping data‑center strategies, power‑grid planning, and partnership models across the entire digital‑infrastructure ecosystem.

Why the Panel Matters

The sheer scale of AI compute today dwarfs traditional workloads. Gartner estimates global AI infrastructure spending will surpass $150 billion by 2027, driven largely by generative‑AI models that consume massive GPU clusters and ultra‑low‑latency networking. In that context, a panel that brings together an Oracle executive, a neocloud specialist, and a seasoned infrastructure advisor offers more than sound‑bites—it provides a roadmap for enterprises wrestling with capacity, cost, and speed‑to‑market pressures.

The Stakes of AI Compute

Dr. Basu highlighted three forces converging on the data‑center market:

Power availability – Large‑scale AI training can demand megawatts of electricity, forcing operators to locate facilities near renewable or grid‑stable sources.
Connectivity – Fiber‑optic bandwidth and edge‑node proximity have become as critical as GPU density, especially for real‑time inference.
Speed to market – Companies that can provision AI‑ready racks in weeks, not months, gain a decisive competitive edge.

These variables are prompting hyperscalers such as Google, Amazon, and Microsoft to double‑down on custom silicon and proprietary interconnects, while emerging neocloud providers leverage modular, container‑based data centers to fill geographic gaps faster than traditional builders.

Implications for Enterprises

For corporate IT leaders, the panel’s insights translate into actionable considerations:

Infrastructure‑as‑a‑Service (IaaS) selection – Enterprises must evaluate providers not just on price per GPU hour but on bundled services like power‑optimisation tools, AI‑specific cooling, and integrated ML Ops platforms.
Hybrid‑cloud strategy – A mix of on‑prem, colocation, and public‑cloud resources can mitigate latency bott bottlenecks and regulatory constraints.
Vendor lock‑in risk – As hyperscalers embed proprietary AI accelerators, organizations should negotiate data‑mobility clauses to preserve flexibility.

Marketing teams, in particular, can leverage these trends to craft narratives around “AI‑ready” offerings, positioning their products as compatible with the emerging AI infrastructure standards that Dr. Basu and his peers are championing.

Competitive Landscape

The panel underscored a shifting competitive dynamic:

Hyperscalers continue to dominate raw compute capacity, but their monolithic footprints limit rapid regional expansion.
Neocloud providers—a term coined for agile, container‑based data‑center operators—are gaining traction in underserved markets, offering “pay‑as‑you‑grow” models that appeal to mid‑size enterprises.
Connectivity firms such as Equinix and Digital Realty are bundling dark‑fiber services with AI‑optimized power solutions, effectively becoming one‑stop shops for AI workloads.

Compared with legacy data‑center operators, these newer players are more likely to embed AI‑specific cooling technologies (e.g., liquid immersion) and to co‑locate with renewable energy assets, a factor that could tilt procurement decisions in their favor.

What It Means for Marketing Teams

Enterprise marketers can extract three pragmatic takeaways:

Storytelling around latency – Emphasise how your solution mitigates AI‑induced latency by leveraging edge‑proximate infrastructure.
Proof points on sustainability – Highlight alignment with providers that source renewable power, a growing procurement criterion.
Educational content – Publish guides that demystify AI infrastructure jargon (e.g., “neocloud,” “hyperscale”) to position your brand as a trusted advisor.

By weaving these angles into campaigns, marketers can differentiate their offerings in a market where technical nuance increasingly drives buying decisions.

Market Landscape

The AI infrastructure market is entering a hyper‑growth phase. IDC projects a 23 % CAGR for AI‑focused data‑center capacity through 2028, outpacing the overall data‑center market’s 7 % growth. This acceleration is fueled by generative‑AI workloads that require clusters of NVIDIA H100 or AMD MI250 GPUs, each demanding up to 400 W of power. Simultaneously, a McKinsey survey found that 68 % of CIOs plan to increase AI‑related capex within the next 12 months, citing competitive pressure and talent scarcity as primary drivers.

Geographically, the United States and China remain the dominant AI compute hubs, but Europe’s “Green AI” initiatives are prompting a surge in renewable‑powered data‑center projects. In response, hyperscalers are expanding edge footprints in emerging markets, while neocloud firms are capitalising on shorter deployment cycles to capture niche demand.

Top Insights

Power‑first design: Enterprises that prioritize on‑site renewable energy and advanced cooling can reduce AI‑training costs by up to 15 %.
Hybrid‑AI flexibility: Combining public‑cloud GPU bursts with on‑prem inference nodes yields a 30 % improvement in latency for real‑time applications.
Neocloud advantage: Modular data‑center pods can be operational in under 60 days, outpacing traditional builds by a factor of three.
Vendor diversification: Relying on a single hyperscaler raises lock‑in risk; a multi‑provider strategy improves resilience and bargaining power.
Marketing leverage: Framing solutions around AI‑ready infrastructure boosts win rates in RFPs by 12 %, according to recent Forrester data.