- Amazon's Trainium has crossed a $10B+ run rate, growing triple digits annually.
- Google TPU, Microsoft Maia, and Meta MTIA all scaling alongside continued NVIDIA Blackwell/Rubin procurement.
- NVIDIA data-center revenue tracking to ~$197B for the year.
- US AI infrastructure strategy now explicitly framed as a counterweight to China's open-source push.
Snapshot — May 2, 2026
27 stories
Claude Opus 4.7 is now generally available, with Anthropic positioning the release as a meaningful step up from 4.6 specifically on advanced software engineering tasks. The update reinforces Anthropic's coding-focused positioning as enterprise adoption of Claude for workflow automation accelerates.
- The ARC Prize Foundation analyzed 160 game runs of OpenAI's GPT-5.5 and Anthropic's Opus 4.7 on the ARC-AGI-3 benchmark, identifying three systematic error patterns that explain why both models score below 1% on the benchmark.
- The analysis suggests current frontier models share structural reasoning blind spots rather than simply lacking scale.
Meta raised its 2026 capex guidance to $125–145B, up from a prior $115B. The increase reflects sustained infrastructure commitment from the hyperscaler tier — and continues to validate the structural Nvidia thesis even as AMD gains share (data-center revenue up 39% YoY to $5.4B last quarter).
Eighteen months after a CFIUS-stalled filing, Cerebras has returned with a Nasdaq IPO targeting up to $4B at a ~$40B valuation — roughly 5× its September 2025 private mark. The wafer-scale challenger comes to market backed by a $10B OpenAI compute commitment and a separate $1B AWS arrangement, framing it as the first credible public-market alternative to Nvidia.
- OpenAI has quietly enabled marketing cookies and ad-targeting tracking by default for free ChatGPT users in ad-enabled markets.
- Paying subscribers are unaffected.
- The move signals a structural shift in OpenAI's consumer monetization strategy — as AI compute costs remain enormous, the company is exploring advertising-supported tiers to subsidize free access.
- Databricks announced AppKit, a new accelerator for enterprise apps, and a private-beta connector that lets Replit Agent build, query, and deploy directly against governed Databricks workspaces.
- Bain & Company is cited as an early reference customer that has displaced Tableau and Power BI internally.
- The deal is the most significant AI-IDE-meets-data-platform integration to date.
- A Harvard study found an AI system delivered more accurate emergency-room diagnoses than two human physicians it was benchmarked against.
- The finding adds to mounting evidence that frontier models, properly conditioned on medical reasoning, are crossing parity thresholds in narrow clinical-decision tasks.
A Mayo Clinic AI system identified pancreatic cancer signals in patient data an average of three years before clinicians made the actual diagnosis — a potentially transformative result for one of oncology's deadliest cancers, where late detection drives most of the mortality.
The Pentagon signed agreements with AWS, Google, Microsoft, OpenAI, NVIDIA, SpaceX, Reflection AI, and (added later the same day) Oracle to deploy on Impact Level 6 and 7 networks. Defense Secretary Pete Hegseth told senators Anthropic refused the department's "terms of service," comparing the position to "Boeing telling us who we can shoot at." The move ends Claude's prior role as the only frontier model on the Pentagon's classified network.
- Researchers published work proposing a human-in-the-loop AI framework for monitoring and control of advanced nuclear reactors, positioning AI as a key enabler for next-generation clean energy infrastructure.
- The system is designed to augment human operator decision-making rather than replace it, addressing both reliability requirements and the regulatory need for human oversight in critical safety systems.
- Microsoft Agent 365 became generally available, extending AI agent governance capabilities across enterprise Microsoft 365 environments.
- The platform lets organizations deploy, manage, monitor, and audit AI agents at scale while establishing guardrails for agentic workflows inside workplace software.
- The launch positions Microsoft as the enterprise control plane for agentic AI — relevant as customers evaluate how to govern autonomous agents running across productivity tools. 📊 Industry News & Deals
- Week one of the Musk vs.
- OpenAI trial concluded with Musk on the stand in Oakland, calling himself a "fool" for investing $38 million in an organization that became an $800 billion enterprise, warning of a "Terminator"-like AI future, and admitting that xAI has used OpenAI's models in its own AI training pipeline — a striking admission given the adversarial nature of the suit.
Microsoft's first new enterprise tier since E5 (2015) bundles E5, Copilot, Microsoft Entra Suite, and the new Agent 365 control plane for governing AI agents. Pricing at $99/user/month makes AI agents a licensable seat tier — a notable structural shift in how Microsoft monetizes Copilot in the enterprise.
Mistral released Medium 3.5 — a 128B dense model with a 256k context window, 77.6% on SWE-Bench Verified, and pricing of $1.50/$7.50 per million input/output tokens under a modified MIT license. Bundled alongside is a new "Vibe" remote-agent runtime and Le Chat Work Mode, marking the lab's most enterprise-grade open-weight push yet.
- Nvidia CEO Jensen Huang publicly criticized tech industry leaders for what he called a "god complex" — specifically, prominent figures who make sweeping predictions that AI will imminently eliminate most human jobs.
- Huang argued that such claims do real, measurable harm by discouraging young people from pursuing AI-adjacent careers and creating unnecessary societal panic.
- A WSJ profile of OpenAI CFO Sarah Friar reveals she privately counseled waiting until 2027 for the company's IPO, even as market pressure and investor expectations mount.
- Friar is credited with playing a pivotal behind-the-scenes role in preserving the Microsoft cloud partnership through its recent restructuring.
In an extensive interview, Replit CEO Amjad Masad addresses the Cursor partnership, the company's open conflict with Apple over App Store policy, and why Replit has rebuffed acquisition interest. He also flags rising "AI bloat" — non-technical users generating code volumes that drive up infrastructure burn rates.
A widely-shared technical analysis from Simon Willison concludes that DeepSeek V4 closes much of the gap to Western frontier models, particularly in long-context reasoning and code synthesis — while remaining materially cheaper to run. The piece is being read inside enterprise AI teams as a serious signal on cost-of-intelligence trajectories.
- Stanford HAI's 2026 AI Index confirms that AI capability continues to accelerate rather than plateau, with industry producing over 90% of notable frontier models in 2025.
- Several top models now meet or exceed human baselines on PhD-level science questions, multimodal reasoning, and competition mathematics.
The Academy Awards formally barred AI-generated acting performances and scripts from Oscar eligibility. The decision sets a precedent that other guilds and award bodies are likely to mirror, and tightens the labor-and-IP guardrails around generative video as Sora 2 and competitors enter the production workflow.
- A widely-shared technical analysis from Simon Willison concludes that DeepSeek V4 — released April 24 with 1M-token context, MoE architecture, and open weights — is "almost on the frontier." The post drew 577 points on Hacker News and is reshaping how Western practitioners benchmark Chinese open models.
- A Hangzhou court issued what is being described as the first major Chinese ruling holding that AI displacement alone is not lawful grounds for termination.
- The decision is likely to influence how Chinese employers structure AI-driven workforce transitions and will be closely read by HR and legal teams globally.
Meta acquired Assured Robot Intelligence — founded a year ago by Xiaolong Wang — with the full team joining Meta Superintelligence Labs to train physical agents using human-experience data. The deal signals Meta's intent to compete directly in the humanoid-robotics race alongside Tesla, Figure, and Apptronik.
A placeholder app in the latest Grok iOS build confirms imminent CarPlay support for Grok Voice mode. Grok will join ChatGPT and Perplexity as third-party AI assistants in CarPlay — meaningful distribution leverage as xAI extends beyond Tesla into the broader vehicle market.
- xAI released Grok 4.3 today, featuring significant price reductions and a new "Imagine" agent mode designed for creative and multimedia projects.
- The model shows benchmark gains on practical tasks compared to its predecessor, but independent reviewers note it continues to trail the top-tier offerings from OpenAI and Anthropic on reasoning and coding benchmarks.
- xAI introduced "Custom Voices," allowing developers to create a usable voice clone from just one minute of recorded speech.
- The feature builds on xAI's recently launched Grok Speech-to-Text and Text-to-Speech APIs and is intended for use in developer applications.
- The low sample-length requirement sets a new bar for accessibility in voice cloning, though it also raises fresh concerns around synthetic voice misuse and identity fraud that safety researchers are already flagging.