Google Releases Gemma 4: The Most Capable Open Models Yet
Google announced Gemma 4 on May 28, its most capable open-weight model family to date, released under the commercially permissive Apache 2.0 license. Built from the same research and technology as Gemini 3, Gemma 4 brings multimodal capabilities — text, image, audio, and video — to models you can run on your own hardware.
The family comes in four sizes (1B, 4B, 12B, and 27B parameters), with multimodal support starting at the 4B tier. The smaller models activate an effective 2B–4B parameter footprint during inference, preserving RAM and battery life. In collaboration with Google Pixel, Qualcomm, and MediaTek, these models run entirely offline with near-zero latency on edge devices like phones, Raspberry Pi, and NVIDIA Jetson Orin Nano.
Since the first Gemma launch, developers have downloaded the models over 400 million times, spawning more than 100,000 community variants. Gemma 4 is purpose-built for advanced reasoning and agentic workflows, delivering what Google calls "unprecedented intelligence-per-parameter." The Apache 2.0 license removes the restrictive terms that limited some enterprise use of earlier Gemma releases.
Source: Google Blog
Dell's AI Server Revenue Surges 757%, Stock Hits Record High
Dell Technologies reported its fiscal Q1 earnings on May 29, and the numbers stunned Wall Street. AI server revenue hit $16.13 billion, a staggering 757% year-over-year increase, as the agentic AI boom drives unprecedented demand for compute infrastructure.
Total revenue came in at $43.8 billion, crushing analyst estimates of $35.5 billion. Non-GAAP earnings per share hit $4.86 — up 214% year-over-year and well above the $2.95 consensus. Dell raised its full-year revenue guidance to approximately $167 billion, including $60 billion from AI server sales alone, up from a prior outlook of $140 billion.
The stock surged 32% on Friday, its best single-day performance ever, briefly touching $441 in after-hours trading. The results underscore how the AI infrastructure buildout is accelerating, not plateauing — a signal that enterprise AI deployment has moved decisively from pilot projects to production-scale investment.
DuckDuckGo Installs Surge 30% as Users Reject Google's AI Search
Google's aggressive push toward AI-powered search is driving users to alternatives. DuckDuckGo reported a 30% spike in US app installs during the week of May 20–25, with iOS installs peaking at a 69.9% week-over-week increase on May 25.
The surge followed Google's announcement that AI agents would be deeply integrated into Search, including always-on "information agents" that monitor the web continuously. Visits to DuckDuckGo's dedicated AI-free search page (noai.duckduckgo.com) averaged 22.7% week-over-week growth, suggesting a meaningful segment of users actively wants a search experience without AI intermediation.
The timing is notable: DuckDuckGo typically sees a traffic dip over US holiday weekends, but Memorial Day weekend 2026 bucked that trend. The growth rate in the US was "multiple times bigger" than international rates, pointing directly at Google's US-centric AI Search rollout as the catalyst.
Source: TechCrunch
California's 30 AI Bills Clear Crossover Deadline
Nearly all of California's 30 active AI-related bills were approved by their chamber of origin ahead of the May 29 crossover deadline, setting the stage for an intense four-week sprint before the July 2 summer adjournment.
Key bills advancing include SB 813, which would establish a California AI Standards and Safety Commission; SB 867, which would ban companion chatbots in children's toys; AB 1542, which would prohibit businesses from selling or sharing sensitive personal information entirely (replacing the current opt-out model); and SB 574, establishing AI use standards for attorneys.
California continues to be the de facto AI regulatory laboratory for the United States. Governor Newsom also signed an executive order on May 21 aimed at preparing workers and businesses for AI-driven job displacement, adding an executive-branch dimension to the legislative push.
Source: Transparency Coalition
Quick Hits
Anthropic's revenue run rate tops $30 billion — up from approximately $9 billion at the end of 2025, reflecting explosive growth in enterprise Claude adoption. The company also announced plans to open an office in Milan. (Anthropic)
OpenAI ships GPT-5.5 Instant and rolls out Codex computer use to Mac users, with Windows environment support improving in recent Codex releases. Goal Mode is now enabled by default in Codex CLI. (OpenAI)