Your AI Intelligence Briefing — Friday, May 1, 2026
Welcome to May. If there's a single thread running through this week's most important AI stories, it's this: the race has decisively moved from the model layer to the infrastructure and influence layers. Big Tech just confirmed it will collectively spend over $700 billion on AI capex this year alone — a near-doubling from 2025 — while SoftBank is racing to IPO a robot-driven data center company at a $100 billion valuation and Huawei is projecting a 60% surge in AI chip revenue as China builds its own hardware ecosystem in response to U.S. export controls. The compute arms race is no longer theoretical; it is the defining capital allocation story of our era.
At the same time, the week showed that the battle for AI's cultural narrative is heating up alongside the hardware fight. OpenAI's acquisition of the TBPN talk show and Novo Nordisk's sweeping deal to embed AI across its entire drug pipeline both signal something important: enterprises and AI labs alike are no longer just building capabilities — they are fighting for the story told about those capabilities, and for the workflows through which AI actually reaches the real world. If this trajectory continues, the most valuable AI companies of 2028 may not be those with the best models, but those that most effectively control both the physical infrastructure and the public conversation surrounding the technology.
Dig deeper into past issues →Fortune / Bloomberg
Following a wave of Q1 2026 earnings reports, Alphabet, Amazon, Meta, and Microsoft collectively confirmed capital expenditure plans that now top $700 billion for the year — nearly double last year's figure — with nearly all of it directed at AI data centers and compute infrastructure. Alphabet's cloud business surged and its shares climbed over 6%, while Meta fell sharply after investors balked at the scale of its AI spending acceleration. The divergence matters: Wall Street is no longer rewarding AI ambition uniformly — it is demanding demonstrated revenue conversion, and companies unable to show that translation are feeling the heat in real time.
CNBC / Financial Times
SoftBank is creating a new entity called "Roze" that will use autonomous robots to construct and operate AI data centers at scale, with a U.S. IPO targeted as early as the second half of this year at a valuation that some executives are pitching at $100 billion. The venture would bundle SoftBank's existing land, energy, and infrastructure assets together with ABB Robotics — the industrial automation leader that SoftBank agreed to acquire last year — creating what Masayoshi Son envisions as the physical construction layer of the AI stack. The concept is striking but carries real risk: SoftBank's past history with capital-intensive construction bets has been mixed, and deploying autonomous robots on live, unstructured data center job sites is a substantially harder engineering challenge than warehouse automation.
Financial Times / The Deep Dive
Huawei is projecting a 60% jump in AI chip revenue to roughly $12 billion in 2026, driven by a flood of orders from ByteDance, Alibaba, and Tencent for its Ascend 950PR processor — with ByteDance alone committing more than $5.6 billion in purchases this year. A critical accelerant was DeepSeek's decision to optimize its V4 model specifically for Huawei hardware, creating immediate enterprise demand and — crucially — shipping a CUDA-compatible software stack that removes the single biggest barrier to switching away from Nvidia. Bernstein analysts now estimate Huawei could capture up to 50% of the Chinese AI chip market under current export restrictions, formalizing what many observers have long anticipated: a permanent bifurcation of the global AI hardware ecosystem into two separate and largely incompatible stacks.
TechCrunch / NPR / CNBC
OpenAI acquired TBPN — a daily tech talk show with a devoted but small audience of founders, investors, and operators — for a price reported to be in the "low hundreds of millions," marking the company's first purchase of a media property. The show will sit within OpenAI's strategy organization and report to chief global affairs officer Chris Lehane, though the company has promised editorial independence for the hosts. Analysts are split: some see it as a savvy move to shape AI's public narrative ahead of an anticipated IPO, while critics question whether a media acquisition is the right use of capital for a company still burning through billions in infrastructure costs — and whether any editorial outlet can truly remain independent once acquired by the company it covers most.
CNBC / BioPharm International / Novo Nordisk
Novo Nordisk announced a comprehensive partnership with OpenAI to embed AI capabilities across its entire operation — from early-stage target identification and clinical dataset analysis through manufacturing, supply chain logistics, and commercial execution — with full deployment targeted by the end of 2026. The deal is structured with human oversight requirements and data governance guardrails, and includes an OpenAI-led upskilling program for Novo's global workforce. The announcement is significant beyond its headline figures: Novo is locked in an existential competitive battle with Eli Lilly in the obesity and diabetes drug market, and with Lilly having already signed 16 AI-related deals since 2025, the partnership reads less like a tech experiment and more like a strategic survival bet on whether AI can compress R&D timelines enough to close a widening competitive gap.