Matt and Georgia recap AWS re:Invent 2025 with special guest Michael Walmsley, AWS Serverless Hero and Global Technology Architect at Accenture. Fresh from the Vegas event with 70,000 attendees, they discuss the major announcements, the shift toward AI agents, and Michael's wild experience coding on a bus for a $100K hackathon prize.

Road to re:Invent Hackathon

• 50 developers coded on buses traveling LA to Vegas over 5 hours
• Michael's team built "Lucky Loo.me" - an AI bathroom finder using facial recognition
• Winning team created "Oric" - an IDE that turns 3 lines into 3,000 lines of AI slop
• Prize: $100K split among the winning team

The Big Theme: AI Agents Everywhere

• "Agents" was the dominant word at every booth
• AWS pushing agent capabilities into every service team
• Evolution from general AI (2024) to production agent platforms (2025)

Announcements we covered:

Agent Core Updates

• New policy controls for blocking unauthorized actions
• Evaluation tools for inspecting agent behavior
• Progressive adoption - use pieces without adopting the whole platform

AWS Agent Marketplace

• Vendors can now sell pre-built agents
• Example: Cloud Zero cost management agent

Lambda Updates

• Lambda managed instances
• Durable functions for long-running workflows in code
• Alternative for developers who don't want Step Functions

S3 Vectors (GA)

• Store 20 trillion vectors in one bucket
• 90% cost savings vs traditional vector databases
• Sub-100ms query times for frequent queries
• "S3 is the cheapest database on the planet"

CloudWatch Unified Data Store

• All logs and metrics exposed in S3 Tables
• Cheap, structured SQL querying of observability data

AWS Interconnect ⭐ Biggest Surprise

• High-speed encrypted links between AWS and Google Cloud
• Azure support coming 2026
• Free during preview (pricing TBA)
• Major shift from AWS's anti-multi-cloud stance
• Acknowledges multi-cloud reality in enterprises

Kiro

• Rebranding away from confusing "Amazon Q" umbrella
• Kiro Powers: AI-activated tool modules
• Reduces context bloat in coding agents
• Active hackathon scene with significant prize pools

Michael Walmsley - AWS Serverless Hero, Global Technology Architect at Accenture, specializing in serverless and SaaS architecture. Fourth year attending re:Invent.

AWS is maturing from general AI capabilities to production-ready agent platforms while finally embracing multi-cloud architectures. The focus has shifted to making agents secure, manageable, and practical for enterprise use.

In this episode, Matt and Georgia sit down with Brad Young (Capgemini Invent) and Alistair Adams (Solution Energy) for a fast-moving conversation about AI’s exploding energy appetite and what it means for the future of data centers, power grids, and sustainability. From geopolitical tension to geothermal innovation, this one covers the full energy spectrum.

What We Covered:

- AI’s Energy Crunch AI growth is driving unprecedented demand for power. Hyperscalers like Meta, Google, and Microsoft are signing multi-billion-dollar infrastructure contracts at record pace, stretching grids and reshaping global infrastructure priorities. - The Rise of “Power-First” Google’s “power-first strategy” shows the new reality: build data centers where the power is, not where the people are. Nvidia’s Jensen Huang agrees—co-locating at generation sites may be the future. Reliable, renewable baseload power is now the real competitive edge. - Water: The Silent Crisis Energy gets the headlines, but water is just as critical. Google already uses ~70 billion litres annually for cooling—on track to rise tenfold. Innovations like geothermal heat rejection (e.g., the Pawsey supercomputer in WA) offer promising alternatives. - Renewables: What Actually Works Not all green energy is created equal. Wind and solar can’t deliver the 24/7 baseload those massive GPU clusters require. That leaves geothermal and nuclear as the only scalable clean options—though nuclear remains politically fraught in markets like Australia.

Regional Realities

- Australia: Victoria faces a looming 1.5 GW gap with coal retirement. - UK: Grid constraints limit data center growth. - US: Federal policy is leaning hard into nuclear and geothermal for AI. - Europe: Regulation is reshaping the tech landscape—for better or worse.

Cloud’s Hidden ESG Problem

Most cloud usage sits in companies’ Scope 3 emissions. As ESG rules tighten, lack of transparency from hyperscalers becomes a real compliance exposure. - Social License Becomes Strategy Community pushback is halting billion-dollar projects. The new game: secure energy, protect water, and bring the community with you. “Permission-based infrastructure” is quickly becoming the norm. - AI, Talent & the Enterprise Gap We discuss the widening skills challenge—junior staff struggle to validate AI outputs, and enterprises claiming “we don’t have use cases” are already falling behind. - Greener Compute Through Smart Pricing Dynamic cloud pricing tied to renewable availability is on the horizon—think “off-peak compute,” automatically routing workloads to greener grids.

Standout Insights

- We’re in the “Nokia 3210 era” of AI—25+ years of disruption ahead. - Robotics is still more marketing than reality. - Enterprise AI adoption is early; the real environmental impact is still to come.

Key Takeaways

- Data center location will follow energy, not geography. - Community permission is as critical as capital. - Water use must be part of every sustainability conversation. - Geothermal and nuclear are the only viable clean baseload options. - The next decade will be messy as demand outpaces grid upgrades. - Hyperscalers are accelerating renewable markets—out of necessity. - ESG exposure from opaque cloud emissions is rising fast.

Conclusion

AI’s growth is forcing a complete rethink of how we power digital infrastructure. The winners will be those who can solve the combined puzzle of clean energy, water management, community trust, and transparent reporting—at a speed the grid has never been asked to move before.

After a whirlwind summer break (Georgia was in Australia, the US, Switzerland, France and back to the UK), your hosts return to talk fake spring in Melbourne, big AI news, and the latest progress in our Agentic AI Experiment.

🚀 AI News Highlights Gemini Nano Banana (2.5 Flash): Google’s new multimodal model nails hands (finally) and shines at storyboarding with JSON prompts.

Kimi K2: A front-end coding powerhouse from China’s Moonshot AI — cheaper than Claude Sonnet 4, though backend isn’t its strong suit.

GPT-5: Quietly flexing its ability to augment answers with real-time web searches.

Regulation: Australia looks set to ditch bespoke AI laws — a move we (cautiously) support.

Cloud & Infra: AWS NZ finally opens after a 4-year wait, while Oracle’s $300B OpenAI deal catapults Larry Ellison to the #1 richest spot.

🤖 The Agent Experiment: Content Co-Creator

We update you on our experimental AI system designed to help creators generate social content ideas. The Vision: AI that uses your interests, calendar, and activities to suggest posts, captions, and even storyboards.

The Hurdles:

Social APIs = pricey + restrictive Scraping trending content = messy (lots of “weird” results) TikTok ≠ Instagram: their algorithms play by very different rules Creator Insights: TikTok’s algorithm makes it easier to go viral from scratch — and creators earn more there than on Insta.

❤️ The MLP (Minimum Lovable Product) Instead of chasing APIs, we’re starting simpler: Web app that asks about passions & activities

Optional calendar integration AI-generated content ideas + Nano Banana-powered storyboards Real-world testing on ourselves first And with Instagram’s new “Edits” feature echoing this direction, the market clearly agrees.

🔮 What’s Next Iterating the Content Co-Creator with real feedback Upcoming episodes on renewables + data center power Inviting listeners to weigh in (feedback@cloud-dialogues.com)

This episode blends AI news, social media realities, and product-building tradeoffs — with plenty of laughs along the way.

In the second instalment of our Agentic AI series, Georgia and Matt are joined by returning friend of the pod and AI product expert, Pete Sbarski, to put our AI-generated startup ideas under some very real scrutiny.

But first… the news:

⚖️ Legal & Regulatory Shakeups A US court ruled that training AI on legally purchased books = OK, pirated books? Not so much.

Germany's taking on DeepSeek, pressuring Apple and Google to block the app over privacy concerns.

⚡️ Infrastructure Watch

Amazon's going big in Australia, announcing a massive $13B investment in data centers and solar by 2029.

Cloudflare and Google Cloud had a rough week, with outages reminding us that even hyperscalers aren’t immune to single points of failure.

💡 Back to Business: Can AI Actually Build a Startup?

Last episode, Claude and Gemini pitched six startup ideas.

This week, we brought in Pete to roast them (constructively). Here’s what made the shortlist:

1. 📝 AI RFP Assistant
2. ✅ Compliance Monitor for SMEs
3. 🔍 B2B Data Quality Tool
4. ⚖️ AI Paralegal for Small Firms
5. 💰 Grant Guru for Nonprofits
6. ✍️ Content Copilot for Solo Creators

🧠 Key Takeaways from Pete

• “Just because you can, doesn’t mean you should” — some use cases (like legal AI) come with huge liability baggage.
• Data hygiene does not necessarily mean fancy tools — a lot of orgs want AI, but what they really need is cleaner, more structured data.
• AI isn’t the bottleneck — people are — from resistance to change to unclear ownership, it’s often human systems holding us back.

✅ What We’re Testing Next

We’re moving ahead with two experiments:

The Content Copilot – using AI to repurpose and summarise our own podcast archive.

RFP / Grant Assistant – trialing real-world applications with a friend’s health tech startup.

We’ll be back soon with some early results — and hopefully a winner. 👀 📩

Want early access and exclusive insights from our experiments?

Subscribe to our newsletter for weekly executive takeaways from the edge of AI, cloud and innovation here: cloud-dialogues.com

In this episode, Georgia and Matt serve up your weekly dose of AI and cloud news — plus the launch of something brand new (and a little wild):

*The Agentic Series* — our hands-on experiment to see if AI agents can design, build, and run a business… with minimal human meddling. 📰

Headlines This Week:

Amazon’s AI Workforce Shift: Andy Jassy’s latest shareholder letter teases a leaner, more AI-powered workforce. The hosts dig into what this could *really* mean (and how Amazon Fresh’s “AI stores” were actually powered by humans in a control center in India).

BBC vs. Perplexity: The BBC’s getting legal over AI scraping — accusing Perplexity of publishing full articles without permission. Is AI crossing the fair use line?

Big Tech Chessboard: Rumors swirl of Apple sniffing around Perplexity, while Elon’s XAI Grok cozies up to Microsoft *and* Oracle Cloud.

Game on. 🤖

Launching The Agentic Series What happens when you put Claude Opus, Gemini 2.5 Pro, and ChatGPT 4.5 in a virtual war room and ask them to create a business? We're finding out. 🧪

The Challenge: Create a digital-only, serverless product business — researched, designed, and architected by AI agents. Each LLM acts as a specialist:

Claude & Gemini: product strategist & market analysts

ChatGPT: evaluator and judge

They were tasked with:

• Finding underserved problems
• Sizing the market
• Proposing architecture + GTM strategy
• Recommending pricing and monetization models

📊 The Outcome: Claude’s Top Picks:

1. AI-powered RFP response assistant

2. Continuous compliance monitoring for SMEs

3. Real-time B2B data enrichment

Gemini’s (eventual) Suggestions:

1. AI-powered paralegal for law firms

2. “Grant Guru” for nonprofits

3. Content co-pilot for solo creators

🎯 Winner (according to GPT-4.5): The AI content co-pilot for creators, thanks to high relatability, market size, and strong demo potential.

🔜 What’s Next? We’re moving into solution design and technical architecture in the next episode — with expert guests joining the build.

Expect practical use cases, ethical debates, cost guardrails, and a few surprises as we let the agents do their thing.

💡 Don’t want to miss what AI builds next?

👉 Subscribe to our newsletter for exclusive insights on executive strategy, AI architecture, and emerging tech trends - subscribe here cloud-dialogues.com

🎙️ Cloud Dialogues #34: Plugging AI Into the Real World — Why MCP Matters

In this episode of Cloud Dialogues, Georgia and Matt go beyond the buzzwords and into the wiring of enterprise AI. From AI agents that write code to cloud strategies that make a lick of sense, we unpack the tools and protocols reshaping modern AI — starting with the week’s wild headlines.

• Databricks acquires Neon, strengthening the serverless Postgres story

• OpenAI drops Codex, a multitasking software agent that’s raising a few eyebrows

• Google launches Gemini 2.5 and VO3, stepping further into multimodal territory

• Microsoft open-sources GitHub Copilot, plus introduces a new SRE agent

• Oracle bets big on AI, snapping up $40B in Nvidia chips for U.S. data centers

We also call out the rise of “vibe coding” and the risks of AI-powered startups without adult supervision. Georgia thinks that credentials still matter....

MCP (Model Context Protocol) is an answer to a growing problem that we are all hearing about: LLMs are powerful, but kind of clueless. Without fresh data, structured tools, and the ability to reason in real-time, they hallucinate. A lot.

MCP fixes that — by letting AI agents safely plug into real systems (think APIs, dashboards, HR tools) with structure, permissions, and traceability.

• ✅ Inject real-time data into AI workflows without retraining models

• ✅ Reduce hallucinations and increase trust, auditability & reliablity

• ✅ Enable secure, role-based access to internal systems (HR, finance, ops)

• ✅ Build agents that do more than chat — they act

• Pulling support ticket history during a live customer call

• Executing a compliance check via a secure API

• Asking your AI assistant to summarize live KPIs across platforms

🎧 We wrap up with a challenge: what would it really take to let an AI agent run an entire startup? (Spoiler: we’re not there yet — but we’re closer than you think - eyes on this space 👀)

📩 Want weekly insights like this straight to your inbox? Subscribe to our newsletter for exec-focused takes on AI, cloud, and digital transformation — link in the comments.

#CloudDialogues #MCP #GenAI #AgenticAI #Copilot #CloudStrategy #AIIntegration #Anthropic #Claude #APIs #ModelContextProtocol #TechLeadership #DigitalTransformation

🎙️ Cloud Dialogues: Vegas, AI, and What’s Next with Google

This week on Cloud Dialogues, Georgia and Matt are joined by Anupam Phoghat, fresh off the tarmac from Google Next 2025 in Las Vegas, to break down the biggest announcements, hottest trends, and a few surprises from the cloud world’s glitziest event.

But first… the news roundup was 🔥:

🚨 In the News:

• AWS’s Big Growth Energy: 19% year-over-year growth and strong focus on making AI cheaper and faster—thanks to Tranan 2 chips.
• Data Center Power Drain: By 2030, data centers could be drawing more power than Japan. Edge computing, you’re up.
• OpenAI Eyes Windsurf: Rumored $3B acquisition of a VS Code fork to boost its dev game—and catch up to Claude Code.
• Nvidia’s $5.5B Problem: U.S. export restrictions to China are biting hard.
• Kubernetes is Back (Again): KubeCon London showed signs of a Kubernetes renaissance—maybe AI on the edge is to thank.
• MCP Becomes the Standard: Google jumps on Anthropic’s Model Connection Protocol bandwagon. AI models are finally learning to play nice with data.

🎰 Straight from the Vegas Strip: Google Next 2025 Highlights

Anu gives us the inside scoop from the Strip, where 35,000+ attendees spread across 10 hotels saw Google go all-in on AI everything.

✨ Big Announcements:

• The AI Hyper Computer: Custom hardware to supercharge inference at scale
• GKE Upgrades: Now with inference gateway and full VM support
• Cloud Run & Cloud WAN: Big leaps in serverless and networking
• Vertex AI’s Model Garden: A smorgasbord of foundation models, including Claude and Llama
• Agent Development Kit: Tooling to build your own AI agents (think RAG meets agents IRL)
• Open-source + Interop Love: More cross-platform support, less vendor lock-in

We also dug into: ⚾️ MLB’s use of AI agents to manage network ops 🧠 Why your AI is only as smart as your training data 🚩 Twitter/X as a case study in what not to feed your model

Anu wrapped by sharing the mission behind Nebulla Action—his new venture building practical, industry-focused AI tools for healthcare, finance, and retail.

🎧 Ready to decode Google’s big bets and what they mean for your business?

Tune in now wherever you get your podcasts—or hit the link in our bio.

#CloudDialogues #GoogleNext2025 #AI #CloudComputing #VertexAI #DevOps #OpenSource #CloudStrategy #GKE #AIInfrastructure #Podcast