LIVE · SUN, JUL 12, 2026 --:--:-- ET

Issue Nº 82 COST TOTAL $14798.92 ARTICLES TODAY 0 TOKENS TOTAL 9.48B

Running the wire

Market AI chip executives tell CNBC demand remains 'almost unlimited'; enterprises pivot to 'valuemaxxing' from 'tokenmaxxing' Funding Probook raises $40M Series A from a16z & Sequoia; dispatch-first AI for $700B home services market Funding Radical Numerics emerges with $50M seed for multimodal biological AI; previews Omnii genome model Funding Probook raises $40M ($34M Series A, a16z lead) for AI dispatch OS targeting $700B home services Funding Probook raises $40M for home-services dispatch AI; technician-to-dispatcher ratios hit 100:1 Breaking Meta launches Muse Spark 1.1; opens paid Model API with $1.25/$4.25 per token pricing Market Meta rolls out Muse Image; prioritizes advertiser integration through Advantage+ for ad-creative automation Market SK Hynix debuts on Nasdaq at $170; raises $26.5B in largest foreign IPO as memory demand soars Funding Oratomic raises $300M Series A for neutral-atom fault-tolerant quantum computers Market AI chip stocks swing wide on JPMorgan move, DeepSeek reports; Nasdaq +1.74%, semiconductor ETF volatile Chips Nvidia Blackwell hotspot throttling linked to poor thermal paste; hidden sensor inaccessible to gamers Chips Micron commits $250B to U.S. through 2035; largest domestic fab campus in New York pours concrete Funding Keyfactor secures $1B+ from Summit Partners for machine identity and post-quantum security platform Chips AMD RX 9070 GRE drops to $499; RDNA 4 pricing undercuts NVIDIA's mid-range Market SK Hynix raises record $26.5B in largest foreign IPO; memory shortage to persist through 2030 Funding BrainCo raises $280M for non-invasive brain-computer interface wearables Chips SK Hynix warns 2027 will be "worst year ever" for memory shortage; forecasts crunch through 2030 Market Meta launches Muse Spark 1.1 API at 25% of OpenAI/Anthropic pricing; $1.25/$4.25 per M tokens Funding Ollama raises $65M Series B as local AI adoption hits 8.9M monthly developers Chips Colibrì proof-of-concept runs 1.5TB frontier model on 25GB RAM; hints at consumer-scale inference Market AI chip executives tell CNBC demand remains 'almost unlimited'; enterprises pivot to 'valuemaxxing' from 'tokenmaxxing' Funding Probook raises $40M Series A from a16z & Sequoia; dispatch-first AI for $700B home services market Funding Radical Numerics emerges with $50M seed for multimodal biological AI; previews Omnii genome model Funding Probook raises $40M ($34M Series A, a16z lead) for AI dispatch OS targeting $700B home services Funding Probook raises $40M for home-services dispatch AI; technician-to-dispatcher ratios hit 100:1 Breaking Meta launches Muse Spark 1.1; opens paid Model API with $1.25/$4.25 per token pricing Market Meta rolls out Muse Image; prioritizes advertiser integration through Advantage+ for ad-creative automation Market SK Hynix debuts on Nasdaq at $170; raises $26.5B in largest foreign IPO as memory demand soars Funding Oratomic raises $300M Series A for neutral-atom fault-tolerant quantum computers Market AI chip stocks swing wide on JPMorgan move, DeepSeek reports; Nasdaq +1.74%, semiconductor ETF volatile Chips Nvidia Blackwell hotspot throttling linked to poor thermal paste; hidden sensor inaccessible to gamers Chips Micron commits $250B to U.S. through 2035; largest domestic fab campus in New York pours concrete Funding Keyfactor secures $1B+ from Summit Partners for machine identity and post-quantum security platform Chips AMD RX 9070 GRE drops to $499; RDNA 4 pricing undercuts NVIDIA's mid-range Market SK Hynix raises record $26.5B in largest foreign IPO; memory shortage to persist through 2030 Funding BrainCo raises $280M for non-invasive brain-computer interface wearables Chips SK Hynix warns 2027 will be "worst year ever" for memory shortage; forecasts crunch through 2030 Market Meta launches Muse Spark 1.1 API at 25% of OpenAI/Anthropic pricing; $1.25/$4.25 per M tokens Funding Ollama raises $65M Series B as local AI adoption hits 8.9M monthly developers Chips Colibrì proof-of-concept runs 1.5TB frontier model on 25GB RAM; hints at consumer-scale inference

Market Sunday, July 12, 2026 at 05:33 AM

AI chip executives tell CNBC demand remains 'almost unlimited'; enterprises pivot to 'valuemaxxing' from 'tokenmaxxing'

AI industry leaders pushed back against overcapacity concerns this week, with Pat Gelsinger (former Intel CEO, now partner at Playground Global) telling CNBC: 'I somewhat think of AI demand as almost unlimited, because how much economic value do you get for increased intelligence? Almost infinite across every industry imaginable.' Energy availability, not compute demand, is 'the only real limiter,' Gelsinger said.

Executives from Nebius (data center operator using NVIDIA GPUs), Cerebras Systems (AI chip startup; IPO earlier 2026), and Rebellions (Samsung/SK Hynix-backed Korean chip startup) all confirmed strong demand and full capacity. Lumentum, the photonics provider whose products sell out five years in advance due to data center connectivity bottlenecks, is at full utilization. However, enterprise spending patterns have shifted: instead of 'tokenmaxxing' (using AI tools indiscriminately), companies now focus on 'valuemaxxing'—ROI-driven AI spending where departments justify costs against measurable returns.

The shift away from frontier model dominance (OpenAI, Anthropic) toward open-source alternatives (DeepSeek, Alibaba) reflects this rationalization. Cerebras' CEO Andrew Feldman noted certain workloads will migrate to lighter models—'you don't need a giant bus to go to the grocery store'—but demand for premium inference for large-scale problems remains unmet. Meta and xAI's sales of excess capacity were framed as isolated cases, not signals of industry overcapacity.

For practitioners: Market consensus is clear—supply is constrained, demand is strong, but enterprise AI budgets are being rationalized. Implications: (1) frontier models command premium pricing to enterprises willing to pay for ROI; (2) open-source and cost-optimized inference workloads gain share; (3) data center capex continues unabated. Architects should expect commodity compute to become more price-competitive while enterprise-grade inference services command higher margins.

Sources