Microsoft plans 80 billion dollars of AI data-center spending for fiscal 2025
Microsoft said in January 2025 it was on track to invest about 80 billion dollars in AI-enabled data centers in fiscal 2025, over half in the US.
MatterGen generates new materials to order
Microsoft published MatterGen in 2025, a generative AI that designs new crystals from target properties, with one compound made in the lab.
The Stargate Project
OpenAI, SoftBank, Oracle, and MGX announced Stargate in 2025, a company intending to invest 500 billion dollars over four years in US AI infrastructure.
DeepSeek releases DeepSeek-R1
DeepSeek releases DeepSeek-R1, an open reasoning model trained largely via reinforcement learning, intensifying global LLM competition.
OpenAI launches Operator and the Computer-Using Agent
OpenAI's Operator, launched January 2025, used a vision-and-reasoning model to click and type inside a real web browser for you.
Trump Revokes the Biden AI Executive Order (2025)
President Trump signs an executive order revoking Biden's AI order and directing an action plan to secure American AI dominance.
US Executive Order 14179: Removing Barriers to American Leadership in AI (2025)
Trump signed Executive Order 14179 on 23 January 2025, the affirmative AI directive accompanying EO 14110's revocation and calling for a national AI strategy.
Copyright Office AI Report Part 2: Copyrightability
The US Copyright Office reaffirmed that purely AI-generated output is not copyrightable and that prompts alone do not make a human the author.
Meta guides to 60-65 billion dollars of capex for 2025
In its 2024 results Meta reported 39 billion dollars of capital spending and guided to 60-65 billion for 2025, almost all driven by AI infrastructure.
The first International AI Safety Report (2025)
Chaired by Yoshua Bengio and mandated by over 30 countries, the first International AI Safety Report (2025) assessed the risks of advanced AI.
Augment Code launches enterprise coding platform with Context Engine and Cosmos orchestration
Augment Code launched its enterprise AI coding platform with a 500K-file Context Engine for deep codebase understanding, autonomous Remote Agents, and the Cosmos multi-agent SDLC orchestration platform — backed by $252M in funding.
OpenAI launches Deep Research
OpenAI's Deep Research, launched February 2025 on a version of o3, runs multi-step web research and writes analyst-grade reports.
The Anthropic Economic Index
In 2025 Anthropic began measuring how people use Claude across occupations, reporting 57% augmentation versus 43% automation in its first release.
FTC settles with DoNotPay over 'robot lawyer' claims
The FTC finalized an order against DoNotPay, the self-styled 'world's first robot lawyer,' for deceptive AI claims, with $193,000 in relief.
The Paris AI Action Summit (2025)
At the February 2025 Paris AI Action Summit, around sixty countries signed a statement on inclusive AI; the US and UK declined to sign.
Thomson Reuters v. Ross: the first US AI-training fair-use ruling
A federal judge ruled that a legal-AI startup infringed Westlaw's copyrights by training on its headnotes and rejected its fair-use defense.
Harvey raises a $300M Series D at a $3B valuation
Legal-AI startup Harvey raised 300 million dollars in February 2025 at a 3 billion dollar valuation, with Sequoia leading and customers up to 235 firms.
Google's AI co-scientist
A 2025 multi-agent system built on Gemini that generates and refines research hypotheses, with some predictions confirmed in the lab.
Anthropic introduces Claude Code and the agentic coding era
Anthropic introduces Claude Code, its first agentic coding tool, alongside Claude 3.7 Sonnet as a limited research preview.
Anthropic releases Claude 3.7 Sonnet, its first hybrid reasoning model
Anthropic ships Claude 3.7 Sonnet, billed as the first hybrid reasoning model, with a toggleable extended thinking mode, and previews Claude Code.
ECMWF puts an AI weather model into operations
On 25 February 2025 ECMWF made AIFS its first operational machine-learning forecast, running about 1,000 times more energy-efficiently than physics models.
Charmbracelet launches Crush terminal coding agent with LSP integration
Charmbracelet released Crush, a Go-based TUI terminal coding agent with LSP integration, MCP support, and cross-platform reach including Android and BSD systems.
Potpie launches knowledge-graph coding agent for large codebase navigation
Potpie released an open-source agentic platform that converts a codebase into a Neo4j knowledge graph and runs specialised agents — debugging, code generation, test planning — with structural graph-based understanding rather than keyword or embedding search.
Sutton and Barto win the 2024 Turing Award for reinforcement learning
ACM gave Andrew Barto and Richard Sutton the 2024 Turing Award for the conceptual and algorithmic foundations of reinforcement learning.
MIT and OpenAI study how emotional ChatGPT use affects wellbeing
A 2025 MIT-OpenAI study ran a four-week trial with 981 users and found heavier ChatGPT use linked to more loneliness and emotional dependence.
NIST publishes an Adversarial Machine Learning taxonomy (NIST AI 100-2)
NIST's AI 100-2 gave the field a common taxonomy of adversarial ML attacks: evasion, poisoning, privacy, and abuse, with mitigations.
Google releases Gemini 2.5 Pro, a thinking model
Google launches Gemini 2.5 Pro, a reasoning model that tops the LMArena leaderboard and leads math and science benchmarks like GPQA and AIME 2025.
Therabot: the first randomized trial of a generative-AI therapy chatbot
Dartmouth's 2025 NEJM AI trial was the first RCT of a generative-AI therapy chatbot, reporting large drops in depression and anxiety symptoms.
CoreWeave's IPO, the first big AI-cloud listing
CoreWeave priced its IPO at $40 a share in March 2025, raising about $1.5 billion in the first major public listing of an AI cloud provider.
Anthropic launches Claude for Education with a Learning mode
On April 2, 2025, Anthropic launched Claude for Education with a Socratic Learning mode and campus-wide deals at Northeastern, LSE, and Champlain.
Meta releases the Llama 4 mixture-of-experts models
Meta ships Llama 4 Scout and Maverick, its first natively multimodal mixture-of-experts models, with Scout offering a 10-million-token context window.
Google announces the Agent2Agent (A2A) protocol
Google launched A2A in April 2025, an open standard letting AI agents from different vendors discover and delegate tasks to each other.
Google's Ironwood TPU is built for the age of inference
Google's seventh-generation Ironwood TPU, unveiled in 2025, scales to 9,216 chips and 42.5 exaflops, the first TPU designed primarily for inference.
MICrONS maps a cubic millimeter of mouse brain
In 2025 the MICrONS project released the largest combined wiring-and-activity map of brain tissue, 200,000 cells in a cubic millimeter of mouse cortex.
Google, PJM, and Tapestry Apply AI to the Electric Grid
Google partnered with grid operator PJM and Alphabet moonshot Tapestry to use AI for speeding generator interconnection and grid planning.
IEA Energy and AI Report
The IEA projected data centre electricity use would more than double to about 945 TWh by 2030, driven largely by AI.
OpenAI Preparedness Framework
OpenAI's policy tracking frontier capabilities that could cause severe harm, with High and Critical thresholds gating deployment.
Alibaba releases Qwen3 with switchable thinking modes
Alibaba's Qwen team ships Qwen3, a family of dense and mixture-of-experts models that can switch between step-by-step thinking and fast direct answers.
Copyright Office AI Report Part 3: Generative AI Training
The US Copyright Office said some AI training is fair use but using pirated works to make market-competing output likely is not.
GitHub Copilot Coding Agent reaches general availability at Microsoft Build 2025
GitHub Copilot Coding Agent reached general availability at Microsoft Build, becoming the first native GitHub product to autonomously accept a GitHub issue, write code on a branch, run tests via GitHub Actions, and open a draft pull request with no developer involvement during execution.
AlphaEvolve, a Gemini coding agent that designs algorithms
DeepMind's AlphaEvolve used Gemini to evolve code, improving matrix multiplication and recovering 0.7 percent of Google's compute.
OpenAI Launches the Codex Cloud Coding Agent
In May 2025 OpenAI launched Codex, a cloud agent that works on coding tasks in parallel sandboxes and opens pull requests.
Anthropic releases Claude Opus 4 and Sonnet 4
Anthropic releases Claude Opus 4 and Sonnet 4, models built to sustain focused work over thousands of steps, marking the shift to long-running agentic coding.
Alibaba launches Qwen Code as Apache-licensed terminal coding agent
Alibaba's QwenLM team released Qwen Code, an Apache 2.0 CLI coding agent forked from Gemini CLI and optimised for Qwen3-Coder models, gaining 25,000 GitHub stars rapidly.
ByteDance SE Lab's Trae Agent reaches top SWE-bench score with multi-model ensemble
ByteDance Software Engineering Research Lab released Trae Agent, achieving 75.2% on SWE-bench Verified through a multi-model ensemble of Claude, Gemini, and o4-mini.
mini-SWE-agent achieves >74% SWE-bench Verified with 100-line agent core
The SWE-agent team released mini-SWE-agent, achieving over 74% on SWE-bench Verified with approximately 100 lines of core Python and only bash commands, pip-installable in seconds.
Meta signs a 20-year nuclear deal for the Clinton plant
Meta's June 2025 deal bought all 1,121 MW of Illinois' Clinton nuclear plant for 20 years, keeping it open past its subsidy to power AI data centers.
Anthropic launches Claude Gov for national security
In June 2025 Anthropic released Claude Gov, custom models built for US national-security customers working in classified environments.