From Core To Edge: Akamai On Where AI Inference Must Live Next

What if the real AI race in 2026 isn't about building bigger models, but about where decisions are made, how fast they happen, and whether they deliver measurable value? In this episode, I'm joined by John Bradshaw, Director of Cloud Computing Technology and Strategy at Akamai, to unpack his predictions for the next phase of cloud, AI inference, and the economics that will shape enterprise technology over the next 12 months. As organizations move beyond experimentation, John explains why the boardroom conversation has shifted from capability to return on investment, and how spiraling compute demands are forcing leaders to rethink the balance between performance, cost, and innovation. We explore why this new financial scrutiny is not slowing AI adoption, but refining it. John shares how inefficient GPU workflows, centralized inference, and poorly aligned architectures are being challenged by a more disciplined approach that pushes intelligence closer to the edge. This shift is not only about latency and performance. It is about building scalable, value-driven platforms that can support real-time decision-making, agentic workloads, and global user experiences without breaking traditional IT budgets. Trust is another major theme throughout our conversation. From the rise of everyday AI agents that quietly handle routine tasks to the growing importance of secure, resilient inference pipelines, John outlines how low-latency edge infrastructure, local processing, and hybrid cloud models will redefine reliability for both enterprises and consumers. We also discuss the smart home backlash following recent outages, and why the next generation of connected products will be designed to work even when the network does not. The episode also looks at the future of streaming, where consolidation, intelligent content delivery, and AI-driven personalization are reshaping both the user experience and the economics behind the platforms. Behind the scenes, orchestration is emerging as a defining capability, with multiple models and services working together to validate outputs, reduce hallucinations, and create more dependable AI systems. This is a conversation about moving from possibility to production, from experimentation to accountability, and from centralized architectures to distributed intelligence. So as AI becomes embedded in every workflow and every customer interaction, will the winners be the companies with the biggest models, or the ones that know exactly where their AI should live, how it should be orchestrated, and how it proves its value every single day?

What happens when AI moves from a standalone tool to a teammate that works inside the flow of your organization? In this episode, I'm joined by Mick Hodgins, General Manager for EMEA at Notion, to explore how the idea of a connected AI workspace is reshaping the way teams collaborate, make decisions, and measure productivity. With a career that includes more than a decade at Google scaling growth across multiple countries, Mick brings a unique perspective on what it takes to build technology businesses across diverse markets and why this moment in AI feels fundamentally different from previous waves of innovation. We talk about Notion's journey from a flexible, block-based collaboration platform to an AI-native workspace where context is the real differentiator. Mick explains why AI performs better when it understands how work actually happens, and how embedding agents directly into shared workflows allows teams to move from prompting tools to orchestrating outcomes. From automated reporting and knowledge management to self-improving agent loops that learn from their own performance, the conversation brings to life how organizations are already using AI to remove the "work around the work" and focus on higher-value thinking. A major theme throughout the discussion is return on investment. In a world where many companies are still stuck in pilot mode, Mick shares how leaders can reframe ROI around productivity, speed, and the elimination of repetitive tasks rather than treating AI as a single project with a fixed payback period. We also explore how roles, org structures, and hiring priorities are beginning to shift as agents become extensions of team capability rather than experimental add-ons. Because Mick leads the EMEA region, we also dive into the differences in adoption between the US and Europe, from regulatory considerations and cultural attitudes to the growing strength of the European startup ecosystem. It's a balanced view that recognizes both the caution and the creativity emerging across the region. This is ultimately a conversation about friction. What happens to an organization when coordination overhead disappears, when reporting builds itself, and when knowledge stays current without human intervention? So as AI agents move from novelty to infrastructure, are businesses ready to redesign how work gets done, and what becomes possible when teams stop managing tasks and start compounding impact?

*]:pointer-events-auto scroll-mt-[calc(var(--header-height)+min(200px,max(70px,20svh)))]" dir="auto" tabindex="-1" data-turn-id= "request-WEB:927f9cca-7aa3-47be-8fb7-33bf01261dc7-6" data-testid= "conversation-turn-14" data-scroll-anchor="true" data-turn= "assistant">   Is your cloud foundation ready for the explosion of AI workloads, or are you about to scale technical debt at the speed of innovation? In this episode, I'm joined by Apurva Kadakia, Global Head of Cloud and Partnerships at Hexaware, an AI-first transformation company helping enterprises modernize the core systems that will determine whether their AI strategies succeed or stall. With a front-row seat to large-scale cloud programs across industries, Apurva explains why so many organizations that "moved to the cloud" still find themselves unprepared for what comes next, and why modernization-led migration has become a business priority rather than a technology upgrade. We unpack the real warning signs that cloud environments are not fit for AI, from monolithic architectures and spiraling compute costs to hidden integration complexity and security gaps that only surface at scale. Apurva introduces the idea of "clarity before cloud," a structured approach to understanding sprawling application estates, identifying what truly matters to the business, and matching each workload to the right modernization path using the five R's. It's a conversation that moves beyond theory into the practical decisions leaders need to make now if they want to avoid being locked out of future innovation. The role of AI inside the transformation journey is another major theme. Rather than treating AI as a destination, Apurva shares how AI-led and human-perfected assessment models are already accelerating application discovery, classification, and migration planning, completing the majority of the heavy lifting while keeping human judgment firmly in control. We also explore why governance cannot be an afterthought, and how a dedicated Cloud Transformation Office can drive adoption, reskilling, stakeholder alignment, and data readiness without slowing delivery. Looking ahead to a world of agentic systems and rapidly multiplying cloud workloads, this episode offers a clear message. The organizations that win will not be the ones that adopted cloud first, but the ones that modernized with intent. So as AI moves from experimentation to enterprise scale, are your applications, your architecture, and your operating model truly ready to support it, or is now the moment to rethink your path before the next wave hits?      

What if the biggest transformation in hospitality isn't happening in the dining room, but in the kitchen you never see? In this episode, I'm joined by James Pool, Chief Technology and Operations Officer at Middleby, a company quietly powering more than a hundred brands across commercial foodservice and food processing. With more than three decades spent accelerating how food is cooked, prepared, and delivered at scale, James offers a rare look inside the technology, automation, and connected platforms reshaping how some of the world's most recognizable restaurant and retail brands operate. We explore what the connected, IoT-enabled kitchen actually looks like in practice, and why James prefers to think of it as digital automation for the entire restaurant. From front-of-house energy optimization to automated food safety reporting and real-time equipment intelligence in the back, the conversation reveals how data is being used to reduce waste, improve uptime, simplify training, and ultimately increase profitability at the store level. This isn't about adding more screens or more complexity, it's about removing friction from every step of the operation. James also shares how Middleby is bringing together a vast portfolio of technologies, from rapid-cook ovens and ventless kitchens to robotics and AI-driven service insights, into a single harmonized experience. That integration is opening the door to new formats such as ghost kitchens and non-traditional locations, where food can be prepared almost anywhere without the constraints that once defined a commercial kitchen. Along the way, we discuss how brands like Yum! Brands, Dunkin', Domino's, and Kroger are balancing speed, consistency, cost control, and customer experience in an environment where every investment must prove its return. The episode also takes us inside Middleby's Innovation Kitchens around the world, where operators can experiment with layouts, workflows, and equipment in real conditions before committing capital in the field. It's a powerful reminder that the future of hospitality is being prototyped long before it reaches the high street. So as automation, AI, and real-time analytics move from the factory floor into the heart of the restaurant, is the smart kitchen becoming the most important competitive advantage in foodservice, and are brands ready to rethink how their entire operation is designed around it?

In this episode, I'm joined by Ansel Stein, Vice President of Operations at Crisis24, and the leader behind AiiA powered by Palantir, an intelligence platform built to help executives cut through noise and make better calls in uncertain conditions.  Ansel's background spans more than two decades across analysis, diplomacy, and high-stakes advisory work, including supporting U.S. national security priorities. Today, he's applying that same discipline to the private sector, helping organizations turn overwhelming streams of information into judgment leaders can actually use. We talk about what "intelligence" really means in this context, and why it's different from collecting more data or running another monitoring program. Ansel breaks down the thinking behind the AiiA President's Brief, inspired by the kind of concise, high-rigor briefings senior government leaders rely on, and explains how that model translates into business decision-making without losing context or nuance. If you have ever felt buried by alerts, headlines, and competing narratives, this conversation puts language around that problem and offers a practical alternative. We also address the concerns many leaders have about AI, privacy, and the fear of being tracked. Ansel is clear on boundaries, what data AiiA uses, why open-source intelligence matters, and how governance needs to be designed upfront if trust is going to hold. From structured analytic techniques and scenario planning to the idea that risk and opportunity often sit side by side, this episode is a look at how organizations can move from reacting to anticipating, without handing accountability over to a machine. If your team is trying to shorten the time from signal to decision while still protecting trust, what would it look like to treat intelligence as a leadership habit rather than a crisis tool, and are you ready to build that muscle before the next disruption hits?

How did a routine request from the FBI turn into a decade-long legal battle that helped reshape modern privacy law and ultimately inspire a new kind of mobile network? In this episode, I sit down with Nicholas Merrill, founder of Phreeli and one of the most influential yet often under-recognized figures in the fight for digital rights. Long before privacy became a mainstream talking point, Nick was running an internet service provider that powered major global brands. That journey took a dramatic turn in 2004 when he became the first person to challenge the constitutionality of a National Security Letter under the Patriot Act, living under a gag order for years while the case unfolded. What followed was a deeply personal and professional transformation that led him to question whether litigation and legislation alone could ever keep pace with the scale of modern surveillance. We explore how that experience pushed him toward a third path, building privacy directly into technology itself. From launching the Calyx Institute and developing privacy-focused Android software to raising a multi-million-dollar endowment for digital rights, Nick has spent decades turning principles into practical tools. Now, with Phreeli, he is taking that philosophy into one of the most data-hungry industries of all, mobile telecoms, reimagining what a carrier looks like when it is designed to know as little about its customers as possible. Our conversation also tackles the shifting balance of power between governments and corporations in the data economy, and why the distinction between the two is becoming increasingly blurred. Nick explains the trade-offs involved in building a privacy-first operator in a heavily regulated market, the cryptographic thinking behind Phreeli's double-blind architecture, and why he believes consent and personal agency should sit at the center of the digital experience. This is a story about resistance, resilience, and the belief that technology can be used to restore choice rather than quietly remove it. It is also a timely reminder that privacy is not an abstract concept for activists and engineers, but something as familiar as closing the curtains in your own home. So after three decades on the front lines of this debate, what does Nick think most of us still misunderstand about our digital rights, and what single shift in mindset could change how we all approach privacy in the connected world?

*]:pointer-events-auto scroll-mt-[calc(var(--header-height)+min(200px,max(70px,20svh)))]" dir="auto" tabindex="-1" data-turn-id= "request-WEB:46fbdb14-2fd7-425c-ac70-5727a2453611-0" data-testid= "conversation-turn-2" data-scroll-anchor="true" data-turn= "assistant"> Have you noticed how every week brings a new headline about AI driven fraud, yet it still feels hard to tell what is real risk and what is noise? In this Tech Talks Daily episode, I'm joined by Tommy Nicholas, CEO of Alloy, for a candid conversation that cuts through the fear driven commentary and gets into what fraud teams are actually dealing with right now.  We start with a simple but important distinction that gets blurred all the time. Tommy separates classic "fraud," where institutions take the hit, from "scams," where individuals are manipulated into handing over money or access. That framing changes how you think about solutions, accountability, and where AI is making things worse. Tommy also shares why he believes fraud losses are often massively underreported. It is not because people are trying to hide the truth, it is because organizations rarely have a single, clean view of losses across every product line and channel. Add messy labeling, split ownership across teams, and reporting becomes a best effort estimate rather than an objective number. That reality matters if you're building board level narratives, budgets, or risk models on top of survey data. From there, we talk about what organizations are getting right. Tommy argues there is no magical "undetectable" attack that forces teams to give up, but there is a very real breakdown happening in old fallbacks, especially human review of images and video. The bigger shift he sees is banks and fintechs finally pushing for consistent tooling across every channel, web, mobile, branch, call center, support tickets, because fraud does not respect internal org charts. We then get into why Alloy's AI Assistant is an interesting signal for where agentic AI is heading in regulated work. Tommy explains that agents are only useful when they have rigorous context, strong sources of truth, and clear workflows. Otherwise they guess, and "looks good" is not the same as "safe to run in production." He also lays out where agents can genuinely outperform humans, like scaling investigations during sudden surges, while keeping processes auditable and repeatable. We close by looking ahead at agentic commerce, and why Tommy thinks the breakthrough will arrive through weird, emergent behavior rather than a neat protocol roll out.  When you listen back, do you think the next big leap in fraud prevention will come from better models, better data, or better operational discipline, and what would you bet on if your own customers were the ones on the line?      

How do you prepare an entire generation for a world where AI is already shaping how we work, create, and solve problems? In this episode of Tech Talks Daily, I'm joined by Dr. Tara Nattrass, Chief Innovation Strategist for Education at Lenovo, for a grounded and thoughtful conversation about what responsible AI integration really looks like in K–12 classrooms. Tara brings more than 25 years of experience inside school districts, including serving as Assistant Superintendent for Teaching and Learning in Arlington Public Schools, so this isn't a theory-led discussion. It's informed by lived experience. We explore how the conversation has shifted over the past 18 months. AI has been present in schools for years through adaptive software and analytics, but the arrival of generative and now agentic AI tools has accelerated everything. As Tara explains, the debate is no longer about whether AI should be in schools. It's about how to approach it responsibly, strategically, and in ways that genuinely improve learning outcomes. A big theme in our conversation is AI literacy. Tara breaks this down in practical terms, moving beyond technical understanding to include critical thinking, creativity, collaboration, and the ability to evaluate risk and bias. She shares real examples of students designing AI tools to solve problems in their communities, shifting the focus from passive consumption to active creation. We also talk about infrastructure readiness. Many school systems have bold ambitions around AI, but there is often a gap between vision and technical capability. AI-ready devices, intelligent infrastructure, cybersecurity, and data governance all play a role in making innovation sustainable rather than experimental.  Lenovo's approach, as Tara describes it, centers on building education ecosystems rather than simply refreshing hardware. There is also a careful balance to strike between innovation, privacy, and inclusion. From hybrid AI models to questions around where data is stored and who can access it, schools are navigating complex decisions. Tara shares how Lenovo partners with districts, policymakers, and organizations such as ISTE and ASCD to align infrastructure, professional learning, and governance frameworks. Looking ahead, we discuss what will separate school systems that truly benefit from AI from those that simply layer new tools onto old teaching models. Vision, educator upskilling, cybersecurity, and rethinking assessment all feature prominently in her answer. If you are working in education, technology leadership, or policy, this conversation offers a practical view of how AI-ready classrooms are being built today and what still needs to happen next. As always, I'd love to hear your thoughts. How is AI reshaping learning in your organization, and are you ready for what comes next?

*]:pointer-events-auto scroll-mt-[calc(var(--header-height)+min(200px,max(70px,20svh)))]" dir="auto" tabindex="-1" data-turn-id= "request-WEB:b0120523-17fc-4d94-97d7-97b3e135ed14-3" data-testid= "conversation-turn-8" data-scroll-anchor="true" data-turn= "assistant"> What does autonomous IT really look like when you move beyond the slideware and start wiring systems together in the real world? At Dynatrace Perform in Las Vegas, I sat down with Pablo Stern, EVP and GM of Technology Workflow Products at ServiceNow, to unpack exactly that. Pablo leads the teams focused on CIOs and CISOs, building the workflows and security products that sit at the heart of modern IT organizations. From service desks and command centers to risk and asset management, his remit is clear: enable AI to work for people, not the other way around. We began with ServiceNow's deepening multi-year partnership with Dynatrace. While the announcement made headlines, Pablo was quick to point out that the real story starts with customers. This collaboration is rooted in a shared goal of helping joint customers reduce outages, improve SLA adherence, and shrink mean time to resolution. The vision of autonomous IT operations is not about hype. It is about connecting observability data with deterministic workflows so that insight can evolve into coordinated, system-level action. Pablo walked me through the maturity curve he sees emerging. First came AI-powered insight, summarizing data and surfacing signals from noise. Then came task automation, drafting knowledge articles, paging teams, triggering predefined playbooks. The next step, and the one that excites him most, is orchestrated autonomy. That means stitching together skills, agents, and workflows into systems that can drive end-to-end outcomes. It is a journey measured in years, not months, and it depends as much on digitizing process and building trust as it does on technology. We also explored root cause analysis, still one of the biggest time drains in IT. By combining Dynatrace's AI-driven observability with ServiceNow's workflow engine, enterprises can automate forensic steps, correlate events faster, and shorten the time spent on major incident bridges where teams debate ownership. Even incremental improvements in accuracy can save hours when incidents strike. Trust, of course, remains central. Pablo was candid that full self-healing systems are still some distance away. What we will see first is relief automation, controlled failovers, scripted actions suggested by machines but approved by humans. Over time, as confidence grows and processes become fully digitized, the balance will shift. Beyond the technology, a consistent theme ran through our conversation. Outcomes have not changed. Enterprises still want higher availability, faster resolution, better employee experiences. What is changing is the how. ServiceNow is reimagining its platform to deliver those outcomes at a much higher standard, not through incremental tweaks, but through rethinking workflows for an AI-first world. From design partnerships with banks building pre-flight change checks, to internal teams acting as the toughest customers, this was a grounded, practical conversation about where autonomous operations are headed and what it will take to get there. If you are a CIO, CISO, or IT leader wondering how to move from theory to execution, this episode offers a clear-eyed look behind the curtain.      

What happens when nearly half of organizations admit they have no AI-specific security controls, yet AI-driven data leaks are accelerating at the same time? In this episode of Tech Talks Daily, I spoke with Aayush Choudhry, CEO and co-founder of Scrut Automation, about what he sees as a blind spot in the cybersecurity industry. While much of the market continues to design tools for Fortune 500 enterprises with deep pockets and large security teams, Aayush argues that the real existential risk sits with the 99 percent of businesses that cannot survive a serious breach. Aayush brings a founder's perspective shaped by firsthand pain. Before launching Scrut, he and his co-founder experienced the grind of managing compliance and security as a cloud-native startup trying to sell into enterprises. They were outsiders to GRC and security at the time, forced to learn from first principles. That experience became the foundation for Scrut Automation, a modern GRC platform built specifically for small and mid-sized companies that cannot afford six-month implementations, armies of consultants, or half-million-dollar tooling budgets. We explore why treating compliance and security as separate functions increases risk for smaller organizations. In the mid-market, the same small team is often responsible for both. When compliance is handled as a box-ticking exercise and security as a separate technical discipline, gaps emerge. Scrut's approach converges governance, risk, and security signals into a unified layer that translates hundreds of technical alerts into context-aware risks that actually matter to the business. Our conversation also tackles AI complacency. Using the classic confidentiality, integrity, and availability framework, Aayush outlines what minimum viable AI security hygiene looks like in practice. That includes ensuring AI agents are not over-privileged compared to the humans they represent, placing guardrails around sensitive data fed into models, and extending supply chain security thinking to agentic integrations. For resource-constrained teams, these are not theoretical concerns. They are daily realities. Perhaps most compelling is his view that AI can act as a force multiplier for small teams. By embedding accumulated expertise into agents trained on anonymized patterns and edge cases, Scrut aims to democratize security know-how that would otherwise require multiple full-time analysts. The goal is simple but ambitious: make enterprise-grade security outcomes accessible without enterprise-grade headcount. If you are leading a small or mid-sized business and wondering how to balance growth, compliance, and AI risk without breaking the bank, this conversation offers a candid look from the trenches.