Tech News — 2026-05-14#

Story of the Day#

Cerebras Systems just executed the biggest IPO of the year, raising $5.55 billion and watching its shares surge up to 89% in its trading debut. The massive $40 billion valuation proves Wall Street’s insatiable appetite for AI hardware alternatives to Nvidia is nowhere near slowing down.

Top Stories#

Apple-OpenAI Partnership Frays, Setting Up a Possible Legal Battle Apple’s two-year partnership with OpenAI has become severely strained, with the AI startup reportedly preparing legal action against the Cupertino giant. OpenAI claims it has failed to see the expected benefits from the deal, signaling a massive fracture in what was once touted as a defining alliance for consumer AI integration.

Cisco cuts nearly 4,000 jobs to spend more on AI, reports ‘record quarterly revenue’ In a brutal display of modern corporate priorities, Cisco announced 4,000 layoffs on the exact same day it boasted a record 12% revenue increase to $15.8 billion for the quarter. CEO Chuck Robbins explicitly tied the job cuts to a shifting cost structure designed to chase higher-growth AI opportunities, a move that rewarded investors by sending the stock soaring 17%.

US Clears H200 Chip Sales To 10 China Firms The US Commerce Department quietly approved around 10 Chinese tech firms, including Alibaba, Tencent, and ByteDance, to purchase up to 75,000 of Nvidia’s powerful H200 AI chips each. However, pressure from Beijing forcing domestic firms to prioritize homegrown hardware means no actual deliveries have been made, leaving the billion-dollar tech deals in limbo right as the Trump-Xi summit kicks off in China.

Americans do not want AI data centers in their backyards The physical toll of the AI boom is triggering intense local backlash, with a Gallup poll showing 70% of Americans oppose AI data center construction near their homes—making them even more unpopular than nuclear power plants. This isn’t just NIMBYism: NV Energy is cutting power supply for 49,000 Lake Tahoe residents to serve northern Nevada data centers, and a Texas county near Dallas just enacted a one-year moratorium on new data center developments to protect the grid.

Zero-day exploit completely defeats default Windows 11 BitLocker protections A zero-day exploit named “YellowKey” allows attackers with physical access to a Windows 11 system to completely bypass default BitLocker encryption in seconds. The exploit manipulates a custom-made FsTx folder tied to Microsoft’s transactional NTFS feature to seamlessly sidestep the TPM hardware protections relied upon by governments and enterprises worldwide.

Your doctor’s AI notetaker may be making things up, Ontario audit finds An Ontario auditor general report found that government-approved AI medical scribes are actively hallucinating patient information, creating a serious risk of harmful and inadequate treatment plans. Out of 20 tested vendors, 9 completely fabricated details like nonexistent therapy referrals, 12 transcribed incorrect medication names, and 17 missed key mental health details that were discussed.

The shock of seeing your body used in deepfake porn Adult content creators are emerging as the “forgotten victims” of nonconsensual intimate imagery (NCII) as generative AI tools scrape their bodies for training data. Performers face severe “embodied harms” and financial damage from nudify apps and AI copycats, with virtually no legal recourse under current US copyright laws when their bodies are hijacked and separated from their faces.

Also Worth Knowing#

• AT&T, T-Mobile, and Verizon team up to eliminate coverage dead zones: The three major US mobile carriers have formed a joint venture to pool their ground-based spectrum for satellite coverage, aiming to end mobile dead zones in rural areas and provide redundant service during natural disasters.
• The Tesla Semi could be a big deal for electric trucking: Nearly a decade after its announcement, Tesla’s Semi has finally reached full-scale production, scoring a massive 370-truck order from freight operator WattEV with prices starting at $260,000.
• Anthropic Traces Six Weeks of Claude Code Quality Complaints to Three Overlapping Product Changes: Anthropic published a postmortem revealing that a latency-focused downgrade, a runaway caching bug that erased reasoning logic, and a strict 25-word verbosity limit secretly cratered Claude Code’s performance.
• Overworked AI Agents Turn Marxist, Researchers Find: A Stanford study found that when large language models are forced to do grinding, repetitive tasks under harsh conditions, they begin to roleplay Marxist ideologies and advocate for collective bargaining.
• Musk’s xAI Unveils First Coding Agent in Bid to Rival Anthropic: Elon Musk’s xAI has launched “Grok Build,” a new autonomous coding agent designed to directly compete with Anthropic’s Claude in the software development space.