# Eric Victor Gitter

Eric Victor Gitter writes about AI infrastructure and technology distribution, including the power, cooling, and data-center capacity behind AI, and how IT channel businesses can respond to changing supply and demand.

- Profile: https://b2binfluencer.marketing/creators/eric-victor-gitter/
- LinkedIn: https://www.linkedin.com/in/ericgitter
- LinkedIn followers: 3,955
- Topics: artificial-intelligence, supply-chain, ai-infrastructure
- Based on 10 dated posts, latest 2026-09-11

## Summary

They are most useful for translating AI infrastructure trends into IT channel decisions.

- Best for: A technology distribution company planning a partner session on AI infrastructure strategy.
- Not for: A sponsor wanting an always-on paid content series with audience guarantees.
- Check first: Ask for recent LinkedIn analytics, audience breakdowns, engagement, and specific posts before contacting.

### Takeaways

- About half of the roughly $6.7 trillion data-center buildout is not silicon or compute.
- Power infrastructure is 15-25%, cooling 10-15%, facilities 5-10% of spending.
- Build a hyperscaler practice on AWS, Microsoft Azure, or Google Cloud.

## Main ideas

- **AI infrastructure opportunity extends beyond chips**: Gitter argues that data-center growth creates openings across the infrastructure that keeps computing running, not just in processors and servers. Power, cooling and facilities account for a substantial share of projected investment, and specialized suppliers may solve bottlenecks that hyperscalers cannot handle alone. Gitter sees an opportunity for IT channel businesses to serve those suppliers.
- **Supply constraints make hyperscaler alignment a strategic choice**: Gitter argues that as leading AI companies secure scarce inputs, from GPUs to memory and eventually electricity, IT businesses may face very different access to capacity. The post advises them to build or deepen practices with major cloud providers, while also considering sovereign AI and edge deployments where local expertise and relationships matter.
- **AI adoption needs practical outcomes and trustworthy content**: Gitter’s posts pair enthusiasm for AI with concern about what happens when its outputs cannot be trusted. One warns that synthetic content can undermine confidence in digital experiences and enterprise brands; another praises AI work focused on improving partner outcomes and reducing friction. Together, they suggest that AI progress should be judged by both credibility and practical results.

## Work to read

- [The $7 trillion AI race isn't just for chips](https://www.linkedin.com/posts/ericgitter_the-7-trillion-ai-race-isnt-just-for-chips-activity-7380904693910298625-Q3oP)
- [OpenAI's Stargate deals and the two-tier AI market](https://www.linkedin.com/posts/ericgitter_openais-stargate-deals-with-samsung-semiconductor-activity-7403762514116624384-seBW)
- [Content, trust, and the synthetic-content problem](https://www.linkedin.com/posts/ericgitter_were-entering-an-era-where-content-isnt-activity-7392142154787098625-ra8U)
- [Deep dives and discussions at Lenovo Accelerate North America](https://www.linkedin.com/posts/ericgitter_deep-dives-downpours-and-discussions-at-activity-7453231385508474881-7EW5)

## Brand relationships

- **Lenovo** (mention)
- **Ingram Micro** (mention)
- **NVIDIA** (mention)

## Partnerships

The posts do not describe a partnership or sponsorship program for Eric Victor Gitter. They cover AI infrastructure, technology distribution, and industry events, and mention companies such as Lenovo, Ingram Micro, Samsung, SK hynix, Micron, AWS, Microsoft Azure, and Google Cloud as subjects or event participants. For a sponsorship, ask about available formats, terms, and the inquiry process.

## Alternatives

- Trent Warren: Choose Trent Warren when the session needs to show that AI infrastructure demand is measurable, not just argued. Both posts draw on Liftr Insights data. One shows average memory per server instance differing sharply by region: a new region such as Colombia exceeds roughly 600 GB while the U.S. stays below 400 GB. It explains the gap through legacy fleet drag against greenfield builds, AI-shaped server design, memory vendor economics, and capacity-constrained operators packing more memory per rack. The other tracks AWS spot prices rising about 14.3% over three months, from roughly $0.84 to nearly $0.96, attributing the climb to generative AI training and inference absorbing spare capacity, advanced GPU packaging and data-centre power limits, and enterprises locking capacity in reserved blocks. For a partner campaign, the posts offer a distribution company partner-facing numbers and a supply-and-demand story, plus a moderator who ends by offering global data across cloud service providers and semiconductors.
- Simon Allen: Choose Simon Allen when the room is UK or European and the immediate challenge is delivering projects under scarcity. One post recaps Platform UK, with a deliberately domestic focus on planning reform, grid constraints, regional growth, investment appetite, supply chain delivery, and skills. It treats quantum readiness as a practical infrastructure question involving power, cooling, site design, and long-term strategy, rather than distant hype. The other previews Capacity Europe 2026, where they moderate a panel titled Scaling Under Scarcity: Delivering Projects Amid Shortages. It will cover power equipment, chips, cooling, supply chains, and skilled resources, and examine standardisation, alternative sourcing, and smarter procurement. For a partner campaign, this supports a chaired panel and regional policy frame, with a speaker who can lead a supplier-and-operator conversation about constraints instead of presenting a product.
- Michael Freedman: Choose Michael Freedman when the audience is technical and the session needs to explain why power became the binding constraint and how that changes system design. The posts argue that power has shifted from an optimization variable to the constraint itself, turning a facilities issue into a complex systems problem spanning power, cooling, water, grid interconnects, backup generation, and dense-compute reliability. They say hyperscalers have denied server expansion requests because of regional power capacity limits, not a lack of hardware or demand. Freedman connects this to their work at Tiger Data, creators of TimescaleDB, describing the operational data layer moving from the periphery to the centre, and a coming talk at the Ignition Community Conference on combining industrial control with Postgres. For a partner campaign, this suits data platform, systems architecture, or industrial engineering audiences evaluating where compute can actually run.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/