AI coding upstart Factory seeks $150M at a $1.5B valuation
Factory, which builds autonomous coding agents, is in talks to raise a new round led by Khosla Ventures, with Keith Rabois set to join the board — squaring up against Anthropic, OpenAI, and Cursor.
One-Line Summary
AI money keeps chasing infrastructure and agentic tools while Big Tech pushes AI deeper into everyday work and regulated industries.
Big Tech
OpenAI lobbies for broader AI use in life sciences
OpenAI shares a new policy-and-research report arguing that AI can help scientists move faster, connect siloed knowledge, and design treatments more efficiently — and it wants policy changes to expand AI’s role in life sciences. The report highlights ideas like faster drug discovery, repurposing FDA-approved drugs, and automating labs, while acknowledging that expert judgment and real-world validation still matter. 1
At the same time, Novo Nordisk announces a strategic partnership with OpenAI that spans drug discovery, manufacturing, supply chain, and commercial operations, with pilots starting immediately and full integration aimed by end of 2026. The collaboration includes workforce upskilling and strict data governance, and builds on Novo’s Gefion AI supercomputer. Analysts note investor optimism but warn that clinical results, not announcements, will be the real test. 2
For teams in healthcare and pharma-adjacent fields, this signals that AI will be used not only for R&D but also to optimize operations like supply chain and commercial analytics. It also shows a push for better data access and national-scale AI resources — a backdrop that could expand AI projects inside regulated enterprises if guardrails and governance are clear. 3
Industry & Biz
Factory: autonomous coding agents startup chases $1.5B valuation
Factory builds autonomous AI bots (agents) that write code, and it is in talks to raise $150 million at a $1.5 billion valuation, with Khosla Ventures leading and Keith Rabois joining the board. Other reported participants include Sequoia Capital, Insight Partners, and Blackstone. The company positions itself against players like Anthropic, OpenAI, and Cursor in the crowded AI coding market. 4
For non-developers, “autonomous coding” means software agents that can take a task (like fixing a bug or scaffolding a feature) and handle multiple steps with less human oversight. If such tools keep improving, product and design teams could see faster iteration cycles and lower handoff friction — but success depends on accuracy, review workflows, and integration into existing tooling. 4
The investor lineup underscores how quickly capital is flooding “agentic” software that promises time savings for engineering teams. The open question for buyers is whether these agents truly reduce cycle time without increasing rework — a calculus that will be settled by side-by-side metrics (issues resolved per week, PR rework rates) rather than demos. 4
Upscale AI, Fluidstack, Accel: capital fuels AI infrastructure and late-stage bets
Upscale AI, an infrastructure startup focused on custom chips and the systems that connect them, is reportedly in talks to raise $180–$200 million at about a $2 billion valuation — despite not having launched a product yet. Investors include Tiger Global, Xora Innovation, and Premji Invest, and the pitch is a full-stack, open-standards path to scalable AI infrastructure. 5
Meanwhile, AI data center builder Fluidstack is said to be discussing a $1 billion round at an $18 billion valuation, after a reported $50 billion multi-site deal with Anthropic to build custom AI data centers in Texas and New York. Fluidstack’s customer list also includes Meta, Poolside, and Black Forest Labs, and the company relocated its HQ to New York to pursue U.S. opportunities. 6
On the venture side, Accel raises $5 billion with $4 billion earmarked for a late-stage Leaders Fund, planning roughly 20 checks averaging $200 million each, focused on AI-powered software, hardware, robotics, defense tech, and data center infrastructure. Accel’s portfolio includes Anthropic, Perplexity, and Lovable — signaling ample late-stage dry powder chasing AI winners. 7
New Tools
Google AI Mode: side-by-side browsing and tab search in Chrome
Google upgrades AI Mode in Chrome so when you click a result, the webpage opens side-by-side with the AI panel — letting you compare sources and ask follow-ups without losing context. You can also add recent tabs, images, or files (including PDFs) directly into an AI Mode query using the new plus menu. The updates are available now in the U.S., with broader rollout planned. 8
Early testers report that side-by-side keeps focus during deep reads and long videos. Coverage from The Verge and TechCrunch notes this reduces “tab hopping” and that AI Mode can answer questions grounded in the page you’re viewing. For many users, this can turn research, shopping, and study flows into a single, continuous workspace. 9
In parallel, Chrome adds Skills — one-click, reusable AI prompts you can run across pages or selected tabs — plus a library of prebuilt Skills for tasks like comparing specs or scanning documents. Skills roll out on desktop (English-US) with privacy safeguards and confirmation prompts for sensitive actions. 10
Canva AI 2.0: an agentic design assistant that calls tools for you
Canva’s updated AI assistant lets you describe what you want and then automatically calls the right design tools, building multi-layered, editable outputs from a single prompt. It adds connectors to Slack, Gmail, Google Drive, Calendar, and Zoom so the assistant can build context from messages, files, and meetings if you allow it. A web research skill and task scheduling (drafts for review) are included. 11
Executives say the platform is shifting from “design with AI” to an “AI platform with design tools,” aiming to be where small businesses finish editing, collaboration, and deployment. The research preview launches now, with broader access in coming weeks. For marketers and designers, this could compress campaign production times — if brand controls and review flows are set up. 12
Canva also claims efficiency gains in its models (for example, its Lucid Origin image-generation model and image-to-video model) and is expanding coding, spreadsheet, and email assistants — nudging more day-to-day content operations into a single AI-first workspace. 11
What This Means for You
Capital continues to flood into AI infrastructure and “agentic” tools, which means more options will hit your stack quickly — but picking winners should start with measurable workflow gains. For coding or design agents, pilot them on a narrow, recurring task and track cycle time, revision counts, and error rates before expanding. 6
If you’re in healthcare or regulated industries, OpenAI’s push and Novo Nordisk’s partnership suggest AI will move deeper into R&D, manufacturing, and commercial ops. The opportunity is in data readiness: inventory what data you can legally and ethically use, define governance, and identify high-friction processes where AI-supported synthesis or forecasting could help. 2
On everyday productivity, Google’s AI Mode turns Chrome into a canvass for research and decision-making. Side-by-side browsing and tab search can simplify tasks like vendor comparisons, requirement gathering, or studying — but ensure you sanity-check AI summaries against original sources for critical decisions. 8
Marketing and design teams can test Canva’s agentic workflows to auto-generate campaign variants, repurpose assets, and prep channel-specific formats. Configure brand kits and approval steps so speed doesn’t compromise consistency — and start with internal or low-risk content before moving to flagship materials. 11
Action Items
- Try Chrome’s AI Mode side-by-side: On Chrome desktop (U.S.), open AI Mode and click through results to compare pages and ask follow-ups without switching tabs; add recent tabs and PDFs with the plus menu for a focused research sprint.
- Save a reusable Skill in Chrome: Turn one of your best prompts (e.g., “summarize key takeaways on any open page”) into a one-click Skill and test it across a few sites to standardize your workflow.
- Pilot Canva’s AI assistant on a real asset: In the research preview, brief it to produce 3–5 on-brand variants of a social post or newsletter header, then refine layers and export the best version.
- Skim OpenAI’s life-sciences report intro: Read the executive summary to understand how AI synthesis and lab automation claims are framed; note any ideas you can adapt to your team’s evidence-gathering.
- Define a small agentic trial in engineering or ops: Choose one repetitive task (bug triage summary, weekly report, or template design refresh) and run an agent/tool for one week with before/after metrics.
Comments (0)