Episodes

  • Why AI Inference Costs Are Crashing 40 Percent This Year

    Why AI Inference Costs Are Crashing 40 Percent This Year
    Jun 24 2026
    Lucas and Luna unpack the dramatic collapse in AI inference pricing, using the latest numbers from NVIDIA, AMD, and Super Micro to show how hardware competition is reshaping the economics of running large language models. They trace the shift from training to inference, explain why inference costs have dropped roughly 40 percent year-over-year, and discuss what that means for startups, cloud margins, and the next wave of AI applications. Along the way, they touch on the surprising rally in Super Micro Computer and what the 'death of the GPU shortage' means for the entire supply chain. A focused, data-driven conversation for anyone trying to understand where the AI market is heading in mid-2026. #AIInference #InferenceCosts #NVIDIA #AMD #SuperMicro #GPU #LLM #CloudComputing #AIHardware #TechInvesting #Semiconductors #AIPricing #DataCenters #AIStartups #Technology #FexingoBusiness #BusinessPodcast #ChatGPTAndBeyond Keep every episode free: buymeacoffee.com/fexingo
    Show More Show Less
    9 mins
  • Why AI Model Marketplaces Are Becoming the New Operating Systems

    Why AI Model Marketplaces Are Becoming the New Operating Systems
    Jun 23 2026
    Episode 69 explores how AI model marketplaces are evolving from simple app stores into full-fledged operating systems for the enterprise. Lucas and Luna discuss Anthropic's Claude Tag feature, which learns from company Slack messages to automate workflows, and how Menlo Ventures' $3 billion fund signals a bet on this platform shift. They examine the market data showing Palantir down 12.4% in five days while Super Micro Computer jumps 14%, suggesting a rotation toward hardware that powers on-premise AI. The hosts argue that the real winner isn't a model provider but the company that controls the middleware layer between models and business processes. They also touch on the security implications raised by the Klue and LastPass breaches, and how model marketplaces could centralize AI governance. A concrete takeaway: watch for which platform attracts the most enterprise developers building custom agents, as that will own the next decade of productivity. #AI #MachineLearning #ModelMarketplaces #Anthropic #ClaudeTag #MenloVentures #EnterpriseAI #Palantir #SuperMicroComputer #NVIDIA #OpenSource #AIWorkflows #Slack #AIProductivity #FexingoBusiness #Technology #BusinessPodcast #AIPlatforms Keep every episode free: buymeacoffee.com/fexingo
    Show More Show Less
    7 mins
  • Why AI Model Marketplaces Are the New App Stores

    Why AI Model Marketplaces Are the New App Stores
    Jun 23 2026
    Lucas and Luna explore how AI model marketplaces—platforms like Hugging Face, Replicate, and emerging enterprise hubs—are transforming the way companies access and deploy artificial intelligence. They discuss why this shift mirrors the early app store economy, the economic incentives driving creators to publish models, and what it means for enterprise adoption. With NVIDIA down 1.8% in a week and Palantir dropping 11.3%, the hosts ask whether the real value in AI is moving from chips to distribution. #AIModelMarketplaces #AppStores #GenerativeAI #HuggingFace #Replicate #EnterpriseAI #AIAdoption #NVIDIA #Palantir #AIEconomy #Technology #BusinessPodcast #FexingoBusiness #LucasAndLuna #OpenSourceAI #AIDistribution #ModelDeployment #Inference Keep every episode free: buymeacoffee.com/fexingo
    Show More Show Less
    9 mins
  • Why AI Chip Stocks Are Splitting Into Two Camps

    Why AI Chip Stocks Are Splitting Into Two Camps
    Jun 22 2026
    Lucas and Luna explore a structural divergence in the AI chip market. They break down why NVIDIA and AMD are diverging — NVIDIA down 1.8% over five days, AMD up 0.8% — and what that signals about the shift from training to inference. They discuss Groq's $650 million raise after NVIDIA's failed acqui-hire, and how the market is splitting between general-purpose GPUs and specialized inference chips. The episode explains why investors need to look beyond the headline 'AI boom' and pay attention to where the value is flowing next. #AI #Chips #NVIDIA #AMD #Groq #Inference #Training #Semiconductors #Investing #TechStocks #Divergence #MarketSplit #AIHardware #SpecializedChips #FexingoBusiness #BusinessPodcast #Technology #GenerativeAI Keep every episode free: buymeacoffee.com/fexingo
    Show More Show Less
    8 mins
  • Why AI Software Stocks Are Splitting Into Haves and Have-Nots

    Why AI Software Stocks Are Splitting Into Haves and Have-Nots
    Jun 22 2026
    Lucas and Luna break down the growing performance gap between AI infrastructure companies like Broadcom and AI application stocks like Salesforce and ServiceNow. With Broadcom up 7.7% in a week and CRM down 8.5%, they explore why the market is rewarding picks-and-shoves plays over front-end AI tools. Using ServiceNow's 7% drop as a case study, they discuss the shift from 'AI-washing' to real revenue expectations, and what it means for investors trying to navigate the second half of 2026. #AIStocks #SoftwareSplit #Broadcom #ServiceNow #Salesforce #AIInfrastructure #EnterpriseSoftware #MarketDivergence #AVGO #CRM #NOW #AIRevenue #InvestingInAI #TechEarnings #StockMarket #Business #FexingoBusiness #BusinessPodcast Keep every episode free: buymeacoffee.com/fexingo
    Show More Show Less
    8 mins
  • Why AI Chip Stocks Are Splitting Into Two Camps

    Why AI Chip Stocks Are Splitting Into Two Camps
    Jun 21 2026
    Lucas and Luna break down the emerging divide in AI chip stocks as of June 2026. NVIDIA and AMD are up 2.7% and 5% in the last five days, while ARM Holdings surges 15.4%. But not all chip companies are riding the same wave — ASML and Intel tell a different story. The hosts explore why the market is now distinguishing between AI compute providers and traditional semiconductor plays, and what that means for investors. They discuss how hyperscalers like Microsoft and Meta are increasingly designing their own silicon, whether NVIDIA's GPU dominance is truly unassailable, and why ARM's architecture is suddenly a battlefield. This episode drills into the data: what's driving the split, which companies benefit, and where the risk lies. Specific numbers, real market moves, and no hype. Just a smart conversation about where the AI chip market is heading. #AIStocks #ChipStocks #NVIDIA #AMD #ARM #Semiconductors #AIHardware #GPU #TechStocks #MarketSplit #Investing #Technology #BusinessPodcast #FexingoBusiness #FexingoTech #ChipWar #AIComputing #StockMarket Keep every episode free: buymeacoffee.com/fexingo
    Show More Show Less
    8 mins
  • Why AI Founders Are Jumping Ship to Rival Labs

    Why AI Founders Are Jumping Ship to Rival Labs
    Jun 21 2026
    Nobel laureate John Jumper just left DeepMind for Anthropic, and it’s not an isolated event. Lucas and Luna unpack what’s driving top AI researchers to switch labs mid-career, the non-compete loopholes in the UK, and why talent churn is accelerating in a market where model-building expertise is the scarcest resource. They connect the move to recent stock action: Broadcom up 7.7% on AI networking demand, ServiceNow down 7% as enterprise AI buying patterns shift. A concrete look at how the war for AI talent is reshaping company strategy—and listener-funded shows that help you make sense of it. #JohnJumper #DeepMind #Anthropic #AITalent #NobelLaureate #AILabs #BrainDrain #NonCompete #UKTech #ModelBuilding #EnterpriseAI #Broadcom #ServiceNow #AIStocks #Technology #FexingoBusiness #BusinessPodcast #ChatGPTandBeyond Keep every episode free: buymeacoffee.com/fexingo
    Show More Show Less
    7 mins
  • Why AI Inference Startups Are Undercutting the Cloud Giants

    Why AI Inference Startups Are Undercutting the Cloud Giants
    Jun 20 2026
    Lucas and Luna explore how a new wave of AI inference startups—like Groq, Fireworks AI, and Together AI—are offering faster, cheaper model deployment than AWS, Azure, or Google Cloud. With NVIDIA's stock at $210 and AMD climbing 5% in a week, the hardware landscape is shifting, but the real battle is in software: specialized inference engines that cut latency by 80% and cost by half. Lucas walks through the economics: a typical production AI workload running on AWS Inferentia costs roughly $0.30 per hour, while a startup's custom stack can do the same task for under $0.10. Luna questions whether the startups can sustain margins when the big cloud providers inevitably drop prices. The conversation also touches on ARM's surprising 15% weekly gain—partly fueled by inference-optimized chip designs. A concrete look at the infrastructure layer that will determine whether generative AI becomes a utility or a luxury. #AIInference #CloudComputing #NVIDIA #AMD #ARM #Groq #FireworksAI #TogetherAI #AWS #Azure #GoogleCloud #InferenceStartups #TechInfrastructure #GenerativeAI #LLM #CloudPricing #FexingoBusiness #Technology Keep every episode free: buymeacoffee.com/fexingo
    Show More Show Less
    10 mins