7 carefully selected reads across AI, business, and investing.
The AI sector is transitioning from speculative growth toward deeper infrastructure integration and hard-nosed strategic planning. While industry insiders largely acknowledge an AI bubble, confidence remains high that it will not burst in 2026. Simultaneously, major players are moving toward vertical integration and addressing critical physical bottlenecks, from custom silicon to power grid limitations.
A survey of industry insiders at the Cerebral Valley London summit revealed that while 51% believe the AI market is in a bubble, they do not expect it to burst this year. Anthropic emerged as the favorite private unicorn for investment, while OpenAI faced the most significant sentiment for shorting among the surveyed crowd.
Source: NewcomerOpenAI has officially moved toward vertical integration with the announcement of its custom 'Jalapeño' inference chip, developed alongside Broadcom. This move aims to decouple product economics from merchant GPU supply, signaling a broader industry trend where frontier labs increasingly treat custom silicon as table stakes.
Source: Latent SpaceScaling AI infrastructure is shifting focus from raw compute to the back-end network that connects GPUs. Managing the complexity of high-bandwidth protocols like NVLINK and Infiniband has become a primary constraint, prompting increased investment in specialized network management software like Netris to ensure data center reliability.
Source: a16z NewsFigma CEO Dylan Field argues that AI represents a structural tailwind for the design platform, despite market narratives viewing the company as an 'AI loser' post-acquisition attempt. The company is actively positioning its Canvas product at the intersection of design and AI to maintain its market relevance.
Source: StratecheryDespite record-breaking solar and wind capacity, China faces severe power grid bottlenecks that are forcing curtailment of clean energy. The lack of infrastructure is slowing the adoption of renewables and forcing a reliance on thermal power plants to maintain stability, illustrating the physical limits of energy transitions.
Source: China Business SpotlightWaymo’s rollout of a premium subscription tier highlights the slow, iterative pace of Physical AI compared to the rapid hype cycle of software. The transition to scalable autonomous driving remains hampered by real-world environmental complexities and ongoing infrastructure development challenges.
Source: AI SupremacySkeptics suggest that the current parade of trillion-dollar valuation targets in AI may serve as a mechanism for government fiscal management. By driving massive step-ups in valuations, the resulting capital gains could theoretically help address burgeoning sovereign deficits.
Source: Doomberg