OpenAI launched GPT-5.5 on April 23, its most intelligent model to date, with dramatically improved agentic coding, research, and multi-tool task completion. The model is rolling out to Plus, Pro, Business, and Enterprise tiers in ChatGPT and Codex, with API access following on April 24. OpenAI says GPT-5.5 is both smarter and more token-efficient than GPT-5.4, matching its per-token latency while delivering a significant intelligence jump.
DeepSeek released preview versions of V4-Pro (1.6T params, 49B activated) and V4-Flash (284B params, 13B activated) with 1M-token context on April 24, rivaling top closed models in math, coding, and reasoning. However, Bloomberg reports the full launch is being delayed as DeepSeek reworks its software stack to optimize for Huawei Ascend chips — a strategic shift toward China's domestic semiconductor ecosystem as US export controls tighten.
Google rebranded Vertex AI as the Gemini Enterprise Agent Platform, launched Workspace Intelligence (a semantic layer powering agentic work across Docs, Sheets, and Gmail), and shipped A2A protocol v1.0 now live at 150 organizations. The announcements mark a fundamental shift from generative AI to agentic AI across Google's enterprise stack, with Workspace Studio, 200+ models in the Model Garden, and managed MCP servers.
The Information reports Google co-founder Sergey Brin has assembled a dedicated group within DeepMind to rapidly improve Gemini's coding capabilities, with a specific mandate to close the gap with Anthropic's Claude. In a memo to DeepMind staff, Brin wrote: "To win the final sprint, we must urgently bridge the gap in agentic execution." The team is led by DeepMind engineer Sebastian Borgeaud and focused on long-horizon, agentic programming tasks.
Meta is laying off 10% of its workforce — roughly 8,000 people — and eliminating 6,000 open roles to "offset investments" in AI, while planning $115–135B in AI infrastructure capex this year. Microsoft simultaneously cut 12,000 positions. The four hyperscalers (Alphabet, Microsoft, Meta, Amazon) are collectively expected to spend nearly $700B on AI infrastructure in 2026, even as they use AI as justification for slashing headcount.
Mira Murati's 14-month-old startup Thinking Machines Lab signed a new multibillion-dollar agreement with Google Cloud for access to Nvidia GB300-powered infrastructure. The deal makes TML one of the first customers on GB300 systems, joining Anthropic and Meta on Google's latest compute. TML raised $2B at a $12B valuation and ships Tinker, its reinforcement-learning-based tool for creating custom frontier models.
A study highlighted by Nature finds that the best AI agents perform only about half as well as PhD-level human experts on complex scientific tasks. Despite rapid adoption — 6–9% of natural science publications now mention AI — agents remain far from autonomous science. The Stanford AI Index 2026 corroborates: benchmarks designed to measure AI are struggling to keep pace with the field, yet fundamental limitations in agentic execution persist.
SpaceX's S-1 filing targets the largest IPO in history at $1.75T, attributing over 90% of its $28.5T total addressable market to AI — driven by its February absorption of Elon Musk's xAI. But the AI unit posted a $6.4B operating loss in 2025, canceling Starlink's profits and pushing SpaceX into the red. The listing is expected ahead of anticipated OpenAI and Anthropic IPOs later this year.
University of Cambridge researchers engineered a nanoelectronic device using modified hafnium oxide that mimics how neurons process and store information simultaneously. The "memristor" approach to neuromorphic computing could cut AI energy consumption by up to 70%, with potential for 2,000x efficiency gains in specific workloads — directly addressing the sustainability concerns around the $700B+ data center buildout now underway.
State lawmakers have introduced over 600 AI bills targeting private entities in 2026 sessions so far, creating an increasingly fragmented regulatory landscape. Indiana, Utah, and Washington enacted laws prohibiting insurers from using AI as the sole basis for denying claims; Tennessee and Delaware passed bills banning AI from being marketed as licensed healthcare professionals. The federal RAISE Act took effect March 19, but a unified national framework remains aspirational.
