Good morning. It's Monday, May 11, and we're covering why developers are rethinking how AI-friendly documents should be created, the growing employee backlash against Meta, and the nationwide fight over data centers.

Know someone who'd love this? Forward it their way. First time reading? Sign up here to keep up with the future of tech.

Apple Nears Intel Chip Deal to Cut TSMC Reliance
Apple has reached a preliminary manufacturing agreement with Intel, according to the Wall Street Journal — a notable shift for a company that has relied exclusively on TSMC for over a decade. The deal would help Apple diversify chip supply amid ongoing constraints that the company says held back iPhone and Mac sales in its March quarter. Intel, now under CEO Lip-Bu Tan, is aggressively courting external customers for its foundry business; its stock jumped more than 15% on the news. Bloomberg separately reported Apple is also exploring a similar arrangement with Samsung.

U.S. Communities Push Back on AI Data Centers
Local governments are blocking AI data centers faster than hyperscalers can build them, with moratoriums jumping from eight to 78 in just one year. The U.S. Data Center Moratorium Tracker lists 50 active bans, including four permanent restrictions, as communities push back over rising electricity costs, pollution, and infrastructure strain. Wholesale power prices have climbed as much as 267% over five years while utilities expand grids to support AI demand, prompting President Donald Trump to press tech firms to cover infrastructure costs.

AI Memory Crunch Hits PC Builders
Motherboard makers are cutting shipment forecasts not because boards are scarce, but because DDR5 RAM has become too expensive for many PC buyers to finish new builds. Taiwanese manufacturers including ASUS, ASRock, Gigabyte, and MSI expect motherboard sales to fall 25% to 30% in 2026 as AI data centers absorb global memory supply. The shortages have cascaded into graphics cards, CPUs, SSDs, laptops, and game consoles — and AMD now estimates its gaming revenue could fall by more than 20% in the second half of the year.

Anthropic Turns to Akamai for Compute
Akamai’s largest contract ever comes not from cybersecurity or content delivery, but from Anthropic’s scramble for AI computing power. Anthropic reportedly signed a seven-year, $1.8 billion cloud computing agreement with Akamai Technologies as demand for Claude coding and enterprise tools accelerates. CEO Dario Amodei said Anthropic saw “80x growth” in annualized revenue and usage during the first quarter, pushing the company to secure additional chips and infrastructure from partners including Google and SpaceX.

Higgsfield is the world's first AI-powered streaming platform, dropping original series made entirely by artificial intelligence. No studios, no gatekeepers, no $200M budgets. Just stories that shouldn't exist yet, streaming right now for free.

Stream for Free.

DeepMind (Google) just invested heavily into Eve Online. They want to analyze player behavior and bake human logic into their AI models! This is the PERFECT playground for research to excel in.

A new essay from Anthropic’s Claude Code team argues that HTML is becoming a more effective format than Markdown for working with AI agents on complex projects. The piece says HTML enables richer visualizations, interactive controls, diagrams, annotated code reviews, and shareable reports — all inside a single browser-renderable file.

The author, Thariq Shihipar, says the shift reflects how developers increasingly use AI-generated documents as living specs, prototypes, and research artifacts rather than lightweight notes. The article also highlights tradeoffs: HTML files take longer to generate, consume more tokens, and create noisier version-control diffs, but proponents argue the added readability and interactivity outweigh the costs. → Read the full article here.

Meta employees are pushing back against new internal policies that track how workers use their computers to help train the company’s artificial intelligence models. According to internal messages reviewed by The New York Times, staff criticized the monitoring program as invasive after executives confirmed there was no opt-out on corporate devices.

The changes come as Meta accelerates its AI strategy under CEO Mark Zuckerberg, including mandatory AI adoption programs, internal token-usage dashboards, and plans to cut roughly 10% of its workforce — about 8,000 jobs — later this month. Employees told the Times the combination of layoffs, surveillance concerns, and pressure to use AI tools has created widespread anxiety inside the company as Meta reshapes itself around AI development. → Read the full article here. (Paywall)

Samsung Workers Push AI Profit Share: Samsung offered chip employees a $340,000 bonus to prevent an 18-day strike, but unions want annual profit-sharing as AI memory demand drives soaring semiconductor profits.

Cloudflare Cuts 20% of Staff: Cloudflare will eliminate more than 1,100 jobs as it restructures around “agentic AI” workflows, even after posting stronger-than-expected quarterly revenue and profit growth.

Perplexity Expands AI Desktop Agents: Perplexity opened its “Personal Computer” app to all Mac users, letting AI agents access local files, apps, and web tools to automate multi-step workflows across devices.

Apple Tests Camera-Equipped AirPods (Paywall): Apple’s next-generation AirPods have entered late-stage testing with built-in cameras that analyze surrounding environments, a push toward AI-powered wearables.

Wispr Flow Targets India With Hinglish Voice AI: India is Wispr Flow's fastest-growing market, but the revenue gap tells a more complicated story — the country accounts for 14% of global downloads yet only 2% of in-app purchase revenue.

Use this when: you've finished a draft but suspect the structure is weaker than the writing.

Thanks for reading. See you next time!

— Matthew Berman, Nick Wentz & the Forward Future Team