CSO & VP of AI, DigBI (Microsoft Global Solutions Strategy Alumni)
I have over 30 years in the tech industry, including 26 at Microsoft, my final five years leading AI development were the most transformative. During this period, the importance of Edge and Endpoint AI became crystal clear, especially its impact on productivity, AI agents, and the workforce. I had the privilege of incubating OpenAI, Microsoft’s first Azure AI data center and contributing to products like Microsoft Copilot, GitHub, and Dynamics AI Agents, while helping to shape the company’s broader GenAI roadmap. I oversaw the deployment of both LLMs and SLMs and launched multi-agent systems, seamlessly embedding them into workflows across sales, marketing, support, and engineering.
A large part of my work focused on the strategic hardware needed to power this AI revolution. I evaluated hardware strategies for Edge AI, analyzing GPUs, CPUs, NPUs, and LPUs from AMD, Nvidia, and ARM, and even tested Groq’s LPUs and CoreWeave’s GPUaaS solutions. A key highlight was leading Project Athena, where I drove the development of the Maia100 and Cobalt100 ASICs with TSMC. This effort positioned Microsoft as a leader in custom silicon innovation and was critical for supporting the Azure AI Data Center Stargate-1, which I helped align with SMR and Microsoft Quantum Computing to power OpenAI’s GPT training.
I was deeply involved in the OpenAI partnership and what it took to turn frontier models into real enterprise product value. I worked closely across teams to incubate OpenAI inside Microsoft’s ecosystem, align strategy and operating cadence, and translate rapidly evolving model capabilities into production-ready experiences that could actually scale across business workflows. For me, the real work wasn’t the announcement, it was the grind of making the partnership executable, keeping incentives aligned, and making sure what we shipped could stand up to enterprise expectations.
I’ve been tracking the frontier lab arms race very closely… OpenAI, Anthropic, Google, AWS, Mistral, and then across China I’ve been watching DeepSeek, Alibaba Qwen, Tencent Hunyuan, Baidu ERNIE, Moonshot Kimi, Zhipu GLM, 01.AI, ByteDance Doubao, and others, because the competition has clearly moved beyond “who has the best model.” What I’m really watching now is who owns the agent runtime, distribution, developer surface, cloud attach, and unit economics at scale, because that is where real enterprise lock-in starts. On the platform side, AWS and Microsoft are pushing hard on the full stack, cloud distribution plus custom silicon plus orchestration, while Google is staying vertically integrated through TPU and Gemini, and the Chinese labs are moving aggressively by pairing low-cost models, fast iteration, open-weight strategies in some cases, and strong domestic cloud and app ecosystem reach to drive adoption very quickly.
My focus and experience are global, with deep expertise in AI, frontier models, world models, and cloud infrastructure across the US, Europe, Middle East, China, Japan, Korea, and rest of Asia, where I work with clients in all regions. I’ve advised some of the largest companies in the world on sovereign AI and national strategies designed to control the entire AI stack, helping nations and enterprises reduce dependency on US technology. This work has given me a clear view of how regional priorities differ yet intersect, from infrastructure scaling to regulatory pressures. With this perspective, I’m able to connect insights across markets and guide clients in anticipating shifts, benchmarking against global leaders, and navigating the fast-moving AI landscape with clarity and confidence.
I’ve also been closely tracking the recent acceleration of AGI and AGSI developments across the industry. OpenAI has rolled out o3, o3 Pro, and GPT-5, introducing stronger reasoning and tools like Operator and Deep Research for autonomous web interaction. Their partnership with Oracle on a 4.5 GW compute deal signals massive infrastructure ambitions. xAI launched Grok, advancing multi-agent reasoning and native tool use, positioning itself as a top player in the AGSI space. Meanwhile, Google’s Gemini line has expanded with agent features like AlphaEvolve, integrated across search and workspace. I’ve been working with clients to help them prepare their governance and infrastructure strategies as LLMs evolve into persistent, goal-driven agent ecosystems.
Looking ahead, I’m advising executives on five converging technologies that will reshape enterprise architecture: Quantum AI, neuromorphic chips, post-quantum cryptography, 6G networks, and modular nuclear reactors. These will form the backbone of the next generation of multi-agent ecosystems, running on quantum/supercomputer hybrids and processing planetary-scale data.