The AI Podcast with Fexingo: Artificial Intelligence, Machine Learning, and Modern AI Models

On this episode of The AI Podcast, Lucas and Luna dive into the world of AI agent-powered messaging apps, using Respond.io's recent $62.5 million funding round as a springboard. They explore how large language models are turning customer service chatbots into autonomous agents that can resolve issues, book appointments, and process returns without human intervention. The hosts break down the busin...

Lucas and Luna dive into a hidden bottleneck in the AI boom: utilities are warning they can't connect new data centers fast enough. With examples from Northern Virginia to Ireland, they look at how this constraint is already pushing AI companies toward edge computing and nuclear-powered co-location. Lucas brings a specific number: one Virginia data center will draw as much power as 200,000 homes b...

The AI industry is undergoing a surprising structural shift: top AI companies are increasingly organizing themselves like banks or insurance firms, with dedicated risk management divisions, capital reserves, and even actuarial teams. This episode explores why — from the $50 million plus annual insurance premiums some AI firms now pay, to the hiring of chief risk officers from financial services. L...

Lucas and Luna examine the wave of AI companies filing for IPOs in mid-2026, sparked by record private valuations and a frothy market for AI stocks. They discuss the recent IPO of CoreWeave, which popped 22% on its first day, and the upcoming filings from companies like Anthropic and Scale AI. The hosts unpack how the market is pricing these offerings differently than the 2021 SPAC era, and why tr...

In episode 50, Lucas and Luna explore the rise of 'reasoning models'—AI systems that don't just generate text but actually think through problems step-by-step. They break down how OpenAI's o1 model works, why it's a paradigm shift from GPT-4, and what it means for everything from coding to drug discovery. They also discuss the costs: o1 is 3x more expensive per token, and inference compute is now...

Palantir stock dropped 6.2 percent in the five days ending June 13, 2026, even as the broader AI hardware index rose. Lucas and Luna dig into what's behind the divergence: Palantir's reliance on government contracts, the emerging commercial AI software slowdown, and the broader rotation from AI 'picks and shovels' to AI 'applications.' They discuss the new competitive pressure from Microsoft and A...

In this episode of The AI Podcast with Fexingo, Lucas and Luna examine how Apple's approach to on-device AI is quietly transforming enterprise adoption. With Apple Intelligence launching across the iPhone and Mac ecosystem, companies are shifting from pure cloud-based AI to hybrid models. The hosts break down what this means for the chip supply chain, referencing recent moves by TSMC and Qualcomm,...

Episode 47 of The AI Podcast with Fexingo. Lucas and Luna dig into the diverging fortunes of AI hardware stocks in June 2026. While Nvidia has stalled, Applied Materials and Lam Research have surged 15% and 14% respectively this week, and ASML is up 6.5%. The hosts explain why the market is starting to separate AI training from AI manufacturing, and what that shift means for investors. They discus...

While Nvidia has barely moved in the past five trading days, semiconductor equipment makers ASML and Applied Materials have surged 15% and 22% respectively. Lucas and Luna drill into what this divergence says about the AI investment cycle: the market is looking past GPU scarcity and betting on the next wave of chip manufacturing capacity. They discuss ASML's extreme ultraviolet lithography monopol...

Intel's stock jumped 18.8 percent in a single week while most AI hardware stocks fell. Lucas and Luna unpack what's driving the reversal: Intel's new foundry win with a top-three cloud provider, the role of ASML lithography machines in the turnaround, and why Wall Street is re-rating Intel as a manufacturing play rather than a chip designer. They also touch on broader market rotations and what Int...

On this episode of The AI Podcast with Fexingo, Lucas and Luna dig into a moment that the market may be underestimating: xAI was just sued by a former engineer who says he was fired after raising safety concerns about Grok. They connect this to a broader pattern across OpenAI, Google, and Anthropic, and ask whether unaddressed safety blowback is becoming a genuine liability for public AI companies...

Super Micro Computer's stock has plunged 36 percent in five trading days, wiping out billions in market cap. Lucas and Luna dissect what's behind the crash: a combination of regulatory filing delays, margin compression from AI server competition, and the broader semiconductor sell-off in June 2026. They discuss Super Micro's reliance on Nvidia GPUs, the impact of Dell and Hewlett Packard Enterpris...

Lucas and Luna dive into the accelerating collapse of AI inference costs. On June 10, 2026, they examine how Google's new subscription pricing — $9.99 per month for Gemini Ultra — signals a broader trend: model prices are dropping 10x faster than Moore's Law would predict. They break down the economics behind the plunge, from open-source competition (Meta's Llama 4) to hardware efficiency gains (N...

Lucas and Luna explore the rapid decline in AI model pricing, driven by competition, efficiency gains, and open-weight releases. With Anthropic's Claude Fable 5 now publicly accessible and a new wave of small, cheap models flooding the market, inference costs are plummeting. The hosts examine how this price compression is reshaping enterprise adoption, software margins, and the hardware stocks tha...

Episode 40. Lucas and Luna tackle the uncomfortable math behind AI's spending spree. With Broadcom down 17.6% in a week and Micron dropping 10.8%, the market is questioning whether all that investment will ever pay off. They dig into a single overlooked metric: return on invested capital. Using the SOXX's 5.5% weekly slide as a backdrop, they ask whether AI is a productivity revolution or a capita...

On June 8, 2026, Apple unveiled the biggest AI overhaul in its history at WWDC 2026. This episode cuts through the keynote hype to examine one specific moment: when Apple Intelligence stopped being a promise and became a product. Lucas and Luna walk through the Siri rewrite, the new on-device model that can finish your sentences across any app, and the Shortcuts automation engine that lets you cha...

Despite companies pouring hundreds of billions into AI infrastructure, AI hardware and software stocks are getting crushed. Lucas and Luna unpack the disconnect: record capex from hyperscalers, yet NVIDIA is down 8.5% in a week and Palantir has lost 15.6%. They explore the rise of inference costs eating margins, the shift to smaller models, and what this means for investors. Plus, the hosts discus...

Lucas and Luna dig into the growing gap between training and inference economics in AI. With NVIDIA down 8.5% in a week and Broadcom off 16%, they unpack why inference costs—not training—are now the real bottleneck for software companies. Using Notion's recent Anthropic outage and OpenAI's Lockdown Mode as jumping-off points, they explore how rising compute spend is reshaping product margins, ente...

The AI hardware trade is in a tailspin. In the last five trading days, Nvidia is down 8.5 percent, AMD off 8.6 percent, and Broadcom has dropped 16 percent. Even a stalwart like TSMC is off 5 percent. Lucas and Luna dig into what's behind the rout: shifting expectations from training to inference, rising competition from custom chips, and the emerging fear that hyperscaler CapEx might peak sooner...

The SOXX semiconductor index dropped 5.6 percent in the past five trading days, and AI darlings like Palantir, Broadcom, and AMD are down double digits. Lucas and Luna dig into the numbers to figure out what's really going on — is this a rotation out of AI hype, a reaction to macro headwinds, or something more structural? They look at specific movers: Broadcom down 16 percent, Micron down 16.6 per...

NVIDIA and AMD have lost 8% in a week, while Broadcom and ASML show divergent fortunes. Lucas and Luna dig into why the AI chip narrative is shifting from training to inference — and why that changes which companies win. They unpack the 'inference tax' that's catching enterprise buyers off guard, how model distillation is squeezing demand for premium silicon, and why the next battleground isn't th...

On June 5, 2026, TechCrunch broke the news that Google will pay SpaceX $920 million per month for compute. This episode of The AI Podcast with Fexingo unpacks what that number means for the AI infrastructure race. Lucas and Luna discuss why cloud giants are turning to unconventional providers, how SpaceX's Starlink network fits into AI workloads, and what the market's sell-off in chip stocks — wit...

Lucas and Luna explore the rising trend of small language models (SLMs) in enterprise AI. With a specific focus on Microsoft's Phi-3 and the surprising cost advantages over giants like GPT-4, they break down why companies are ditching one-size-fits-all models. Anchored to recent data—NVIDIA up 3.7% in five days while Broadcom drops 6.2%—they discuss how the shift to smaller, specialized models is...

Meta is taking a page from Tesla's playbook by building AI data centers in tents. Lucas and Luna dig into the New York Times report on Meta's 'tent city' approach, which slashes construction time from years to months. They explore how this cost-saving tactic could reshape AI infrastructure spending, and what it means for chip demand, energy costs, and the broader AI race. With specific numbers on...

Arm Holdings is up over 22 percent in the last five trading days as of June 4, 2026, pushing its market cap past $420 billion. Lucas and Luna break down exactly what's driving the surge—spoiler: it's not just iPhone chips. They walk through Arm's push into AI data centers via its Neoverse architecture, the recent deal with a major cloud provider, and why the company's licensing model is suddenly l...