Your signal. Your price.
Semiconductor stocks and the Nasdaq fell as investors questioned AI valuations, but the host said tighter financing for risky data centers did not yet show the demand collapse or canceled spending associated with the dot-com bust. Jensen Huang remained bullish, while Cathie Wood argued that fear of a bubble was itself evidence against one.
Wood supports international AI oversight and technical monitoring, including chips designed to restrict model training and systems that track activity. He says monitoring is acceptable only when access is tightly controlled and the monitors themselves face scrutiny.
Iron’s sponsor description says it builds data centers and provides GPU infrastructure powered by renewable energy. The pitch targets customers needing scalable computing capacity for AI.
A hardware hacker found two tampered Trezor wallets bought from unauthorized sellers: a Model T from Facebook Marketplace and a Trezor One from TikTok Shop. Both used replacement chips to generate attacker-known recovery phrases; Trezor said it knew of the counterfeit method and had no evidence of compromise in current devices bought through authorized channels.
Mel Mattison expects demand for AI compute to keep rising even as investors debate which models will win. Marty Bent cited low household adoption and spending, while Mattison favors semiconductor and infrastructure suppliers over trying to pick model-company winners.
David Sacks argued that AI infrastructure demand remains strong: customers seek compute, older GPUs still rent at premium prices, and power, permitting, and construction constrain new supply. He said CoreWeave’s bonds can fall in the secondary market without changing the coupon on existing debt, though future borrowing may cost more.
Barclays analyst Dan Levy says Tesla and SpaceX could announce a merger before Tesla’s shareholder meeting, with a joint TerraFab chip plant as a possible driver. Neither company has filed or hired bankers, and prediction markets put the chance of a year-end deal at 24%.
AI infrastructure debt is drawing scrutiny as CoreWeave bonds fall and revenue growth at major AI firms slows; Ryan compares the funding cycle with repeated railroad booms and busts, while noting Korea’s chip-heavy market decline fed into U.S. equities.
The host says slowing Anthropic revenue growth and widening CoreWeave bond spreads are signs of strain in the AI financing chain, with potential spillovers to data centers, semiconductors, and memory. Quinn says the resulting yield relief could help other sectors that have been shut out of capital.
Jad Naous’s Stanford and MIT research included hardware implementations of SHA-3 and application-security work with Ramesh Chandra. Their startup explored recovering from exploits by rolling back application-related state onto a patched version, but the technology was too early.
Azeem Azhar says demand for AI compute remains strong, with hyperscalers and newer cloud providers struggling to secure enough chips and power. He sees the boom as productive for now because customers are paying for scarce GPU capacity and providers can earn returns quickly.
James Cridland says the new Apple TV adds little beyond a remote-control speaker and a microphone, with a stated US price of $300; he questioned its value against a roughly $600 local price. Sam Sethi says Apple’s device chips and proximity features strengthen its ecosystem, but lock out Android and Windows users.
ASML says chipmaking-tool demand exceeds its capacity for at least two years, despite a previously announced 30% delivery increase. CEO Christophe Fouquet says customers want more tools and some orders extend to 2029 and 2039; CFO Roger Dassen says adding major capacity by 2028 is extremely difficult.
Nvidia CEO Jensen Huang says he would consider stopping chip supplies to an irresponsible AI developer, but only after discussing the issue and within limits on policing customers’ use of general-purpose technology. Huang frames safety as an ongoing engineering task and announced $250 million in computing power for UK safety research over five years.
Atomic Machines emerged from six years in stealth with a $250 million investment and a code-driven manufacturing system that builds micromachines with moving parts and features down to single-digit microns. Its Prime Switch relay for AI data centers reportedly opens in 50 microseconds, 1,000 times faster than conventional approaches.
Jeffrey Morgan says on-desk AI systems fill the gap between laptops and large cloud deployments: enterprises get data locality and privacy with near-data-center performance. He says NVIDIA's DGX Station has almost a terabyte of VRAM and can run a trillion-parameter model.
Jeffrey Morgan points to Microsoft and NVIDIA's new Windows ML stack and RTX Spark hardware as signs that data-center and desktop architectures are converging. He says the announced desktop system delivers a petaflop, while existing server rooms could host fleets of these devices, though power and cooling remain practical constraints.
Nathan's ad reads promote Trezor Safe 7's two secure elements and open-source Tropic01 chip, and Casa's five-key vault with three-of-five signing. The ads describe both as self-custody options, with Casa offering guided setup, recovery tools, and an inheritance feature.
Dr. Matthew McDougal has a passive RFID chip implanted in his hand that can unlock doors and exchange small amounts of data with nearby devices. It has no battery, is coated in biocompatible materials, and could remain implanted for life; newer versions offer more storage.
Laskin says owning AI means operating the full stack: GPUs, cluster-management software, inference, models, harnesses, and sandboxes. Reflection built and tested that infrastructure internally, then began offering it to customers, much as Amazon turned its own infrastructure into AWS.
Nick says Apple raised prices by about 20% across several product lines as DRAM costs surged, and other hardware companies followed. He attributes the squeeze to AI data-center demand colliding with memory supply that takes years to expand.
Brett frames higher consumer-device prices as a way scarce memory gets allocated toward higher-value uses, including AI workloads. He compares consumer electronics to party balloons in a helium shortage: buyers can pay more or reduce demand before critical chip production loses supply.
Brett says memory remains a cyclical business: manufacturers underinvested roughly two years earlier, then rising prices will eventually trigger new capacity and overinvestment. He expects margins to fall when supply catches up, though the cycle could settle at a different price level.
Brett argues that limits on releasing high-performing US models could give invited customers cheaper access while restricting broader use. He says China would continue developing models, and doubts energy alone determines the AI lead because chips and efficient deployment also matter.
Lambert says Chinese labs have talent comparable to US labs but far less compute; one leading Chinese researcher was stunned to hear that an average OpenAI researcher could access 5,000 H100 equivalents.