AI Business with Fexingo: Artificial Intelligence Companies, Models, and Enterprise Adoption

Episode 105 cuts through the hype to examine a specific operational challenge every AI company is grappling with: the cost of running inference at scale. Lucas and Luna break down why NVIDIA's five-day surge of 7.9 percent to $211 reflects not just training demand but the hidden economics of serving models. They walk through a concrete example from a mid-size SaaS company that switched from GPT-4-...

In this episode, Lucas and Luna explore the growing trend of AI companies constructing their own data centers instead of renting from cloud providers. They discuss the economics behind this shift, using the recent SK Hynix IPO as a lens into the massive capital expenditure required. The hosts analyze how companies like Microsoft and Google are investing billions to secure compute capacity, and wha...

In this episode, Lucas and Luna explore why major AI companies are now building their own robotics hardware. They dive into the case of Amazon's Digit robot from Agility Robotics, which is already being deployed in warehouses, and examine how companies like Tesla and Google are investing in humanoid and task-specific robots. The hosts discuss how owning the robot body gives AI firms control over t...

Episode 102 of AI Business with Fexingo dives into one of the hottest trends in artificial intelligence: AI companies designing their own custom silicon. Lucas and Luna break down why major players like OpenAI, Google, and Amazon are moving away from off-the-shelf chips from NVIDIA and AMD, and what this means for the broader industry. We look at the specific economics driving this shift — includi...

In this episode of AI Business with Fexingo, Lucas and Luna explore the growing trend of AI companies training models on synthetic video data — entirely computer-generated footage that simulates real-world scenes. With real-world video collection expensive, privacy-invasive, and hard to scale, firms like Waymo and NVIDIA are turning to engines that generate infinite training miles in virtual envir...

In this episode of AI Business with Fexingo, Lucas and Luna explore why AI companies are turning to video games as a rich source of training data. They discuss how game engines provide structured, interactive environments with perfect labels, and how this approach helps models learn physics, strategy, and multi-agent coordination. Lucas shares a striking example: a startup using game clips to trai...

Episode 99 of AI Business with Fexingo. Lucas and Luna explore a fresh angle: how AI companies are borrowing econometric tools from central banks to forecast model revenue. They discuss why Anthropic, Mistral, and others now hire PhD economists to build demand curves for tokens, and how SambaNova's new $1B raise at an $11B valuation ties into the trend. Specific numbers like ARM's 15% drop and MET...

Lucas and Luna dive into the growing practice among AI companies of using voice recordings — from smart speakers, customer service calls, and voice assistants — to train their models. They discuss the privacy trade-offs, the regulatory landscape in mid-2026, and what this means for consumers. Specific references include recent moves by Amazon and Google, and the 4.9% dip in Apple's stock after a p...

In this episode of AI Business with Fexingo, Lucas and Luna explore the growing trend of AI companies building their own robots from scratch. They discuss why firms like Tesla and NVIDIA are shifting from software-only to integrated hardware-robotics platforms, referencing Tesla's Optimus robot and the financial incentives behind vertical integration. The conversation touches on NVIDIA's recent ro...

Episode 96 of AI Business with Fexingo: Lucas and Luna dive into how major AI companies are using public user data — including Google searches, social media posts, and forum comments — to train their models. They discuss the recent TechCrunch article about opting out of Google's AI training, the Microsoft layoffs citing AI efficiency, and broader industry trends. Topics include the scale of data c...

Episode 95 of AI Business with Fexingo explores a growing trend: artificial intelligence companies poaching top talent from the video game industry. Lucas and Luna discuss how former gaming executives are bringing expertise in real-time rendering, GPU optimization, and synthetic environments to AI firms. They examine specific examples like NVIDIA hiring ex-Blizzard engineers for simulation trainin...

Episode 94 of AI Business with Fexingo digs into a quiet shift in cloud computing: major AI companies are moving inference workloads from NVIDIA GPUs to AWS's custom Graviton processors. With Amazon's Mechanical Turk shutting down and data labeling costs rising, the economics are shifting. Lucas and Luna break down how Graviton's ARM-based architecture delivers up to 40% cost savings for certain A...

In this episode of AI Business with Fexingo, Lucas and Luna dive into the growing trend of AI companies using synthetic data to train autonomous vehicles. They explore why companies like Waymo and Tesla are turning to simulated environments to generate millions of miles of driving data, the role of game engines like Unreal and Unity, and the economics behind synthetic vs. real-world data collectio...

Lucas and Luna explore the growing trend of AI companies building custom inference chips. With Palantir up 14.5% in five days and Super Micro Computer down 11.1%, they discuss the economics of inference versus training. They break down why companies like Amazon, Google, and Microsoft are developing their own silicon, and why this is reshaping the AI hardware landscape. Specific examples include Am...

In this episode, Lucas and Luna explore why AI companies are bypassing traditional utilities and building their own power plants in the Permian Basin. With NVIDIA up 1.2% in the past five days and Super Micro Computer down 11.1%, the conversation drills into the specific economics behind the trend: the cost of building a 1-gigawatt natural-gas plant versus buying from the grid, the strategic value...

Episode 90 of AI Business with Fexingo examines why major AI companies are investing billions into developing proprietary foundation models rather than relying on third-party APIs. Lucas and Luna break down the economics, using NVIDIA's 5-day gain and Palantir's surprising 14.5% rally as context. They explore the strategic calculus behind vertical integration, from control over training data to ma...

In this episode of AI Business with Fexingo, Lucas and Luna explore a quiet shift in how AI companies set prices: instead of product managers or sales VPs, firms like OpenAI and Anthropic are hiring PhD economists to design pricing models. We break down why economic theory is replacing gut feel in the age of inference costs, usage-based tiers, and margin pressure. Specific examples include OpenAI'...

Episode 88 of AI Business with Fexingo dives into why major AI companies—including Meta and Anthropic—are moving away from relying solely on NVIDIA and AMD GPUs for inference workloads. Lucas and Luna discuss the economics of building custom inference chips, including how Meta's vibe-coded gaming app Pocket reveals a shift toward cheaper, in-house silicon for serving models. They reference recent...

Lucas and Luna explore how AI companies are turning to economists—not engineers—to solve the pricing puzzle of inference. With model costs dropping and enterprise customers balking at per-token fees, firms like OpenAI, Anthropic, and even Palantir are hiring behavioral economists and industrial organization PhDs to design pricing tiers that maximize revenue without scaring off users. Lucas breaks...

Lucas and Luna dig into the surprising trend of AI companies buying natural-gas-fired power plants—specifically in the Permian Basin—to secure electricity for their data centers. They discuss the economic rationale, the role of stranded gas assets, and what this means for the grid. Lucas references the recent $800 million raise by neocloud Together AI and the 14% five-day rally in META as context...

Episode 85 of AI Business with Fexingo dives into the emerging trend of AI companies acquiring natural-gas-fired power plants to secure energy for data centers. Lucas and Luna examine a specific case: a mid-sized AI startup that bought a 200-megawatt plant in the Permian Basin for $180 million, with a 15-year gas supply locked in at $2.85 per million BTU. They discuss the economics behind the move...

As AI model training and inference demand skyrockets, major tech firms are moving beyond data center leases and cloud contracts to a more radical solution: buying their own power plants. In this episode, Lucas and Luna explore why Microsoft, Amazon, and others are signing direct power purchase agreements and even acquiring generation assets to secure cheap, reliable electricity. They discuss the e...

Forget A/B testing with real humans — a growing wave of AI startups is turning to 'synthetic users' to simulate how people interact with their products. In this episode, Lucas and Luna dig into the numbers behind this trend, including how one startup used synthetic users to cut testing time by 70 percent. They also explore the tension between speed and validity: can a simulated user really predict...

Episode 82 of AI Business with Fexingo explores the accelerating trend of AI startups designing custom chips. Lucas and Luna break down how companies like Cerebras and Groq are challenging NVIDIA's dominance, with a focus on the recent $550 billion South Korean memory investment to ease 'RAMageddon'. They discuss the economics of in-house silicon, citing NVIDIA's 3.1% dip to $193.90 and AMD's 3.1%...

Amid a broad AI stock selloff in late June 2026, memory maker Micron has bucked the trend, with Wall Street calling it 'the next Nvidia.' In this episode, Lucas and Luna explore why high-bandwidth memory is becoming the bottleneck for AI inference and training, how Micron's HBM4E chips fit into the picture, and what the shift from compute-bound to memory-bound AI workloads means for enterprise ado...