This Week in AI & Automation
Week of February 10, 2026
AI agents graduated from solo performers to team players this week. Anthropic shipped Opus 4.6 with a new "agent teams" mode that lets multiple AI agents coordinate on tasks together — a feature that signals where enterprise AI is heading. Meanwhile, venture capitalists broke a decades-old taboo by backing both OpenAI and Anthropic simultaneously, and a major Infosys report confirmed what we've been saying: manufacturers are done experimenting with AI and building it into their core roadmaps.
The Big Story
Anthropic Ships Opus 4.6 with Multi-Agent Teams
Anthropic released Claude Opus 4.6 on February 5, and the headline feature changes how we think about AI-assisted work. The new "agent teams" mode in Claude Code lets users split tasks across multiple AI agents that work in parallel and coordinate autonomously — instead of one agent grinding through steps sequentially.
The practical impact: a codebase review that took 45 minutes with a single agent now runs in parallel across a team, with each agent owning a piece and flagging issues to the others. The model also ships with a 1M token context window (a first for Opus-class models), stronger coding performance on Terminal-Bench 2.0, and the top score on Humanity's Last Exam.
For enterprises, the real story is the pricing: $5/$25 per million tokens, unchanged from Opus 4.5. More capability at the same cost — the economics of AI-powered knowledge work just got meaningfully better.
Source: Anthropic
Our Take: Agent teams are not just a developer convenience — they're the prototype for how enterprise workflows will run. When AI can split work across specialized sub-agents and reassemble the results, you stop thinking about AI as a tool and start thinking about it as a team. The companies that figure out how to integrate these agents into existing workflows will have a structural advantage over those still running single-agent chatbots.
Notable Developments
VCs Break a Decades-Old Taboo: Dual Bets on OpenAI and Anthropic
Bloomberg reported on February 11 that Sequoia Capital, Altimeter Capital, Blackstone, and Abu Dhabi's MGX are investing in both OpenAI and Anthropic — breaking Silicon Valley's long-standing rule against backing direct competitors in the same space.
Anthropic's latest round is expected to bring in over $20 billion, with Altimeter alone committing over $200 million despite calling OpenAI "its biggest bet ever." The logic is blunt: the cost of missing a generational company outweighs the awkwardness of backing its rival.
Source: Bloomberg
Our Take: When the smartest money in tech refuses to pick sides, it tells you something: this market is big enough for multiple winners, and no one knows who will dominate. For enterprise buyers, this is actually good news — intense competition means faster innovation, better pricing, and more leverage during vendor negotiations. Don't lock yourself into a single AI provider.
75% of Manufacturers Have AI in Their Enterprise Roadmaps
The Infosys Knowledge Institute released its Manufacturing Tech Index on February 14, surveying 650 senior manufacturing executives worldwide. The headline number: 75% of manufacturers have formally integrated AI into their enterprise roadmaps — up from the "98% exploring, 20% ready" stat we covered two weeks ago.
The investment is real too. Over half of manufacturers are spending more than $2 million per AI implementation. Cybersecurity and operational technology lead as the most common AI application area (57%), but cybersecurity is also the biggest barrier to scaling (23%).
Source: Infosys Newsroom
Our Take: The shift from "exploring" to "roadmap-integrated" is the signal that matters. Exploration is free — roadmaps mean budget, timelines, and executive sponsorship. Manufacturers who still treat AI quality control and predictive maintenance as experiments are falling behind peers who are operationalizing it. The $2M-per-project spend also validates that enterprises understand AI isn't a free lunch — it requires real investment in data infrastructure, integration, and readiness assessment.
Vention Launches GRIIP: Physical AI Goes Lights-Out
Vention announced GRIIP (Generalized Robotic Industrial Intelligence Pipeline) on February 10 — an end-to-end physical AI system that deploys autonomous robot cells in unstructured manufacturing environments. The system integrates NVIDIA Foundation models for stereo matching and pose estimation, running on Vention's MachineMotion AI controller powered by NVIDIA Jetson.
The performance numbers are striking: sustained autonomous 24/7 lights-out production over three months, throughput of up to five parts per minute, and CAD-to-pick setup in 15 minutes with full robot cell deployment in under two days.
Source: PR Newswire
Our Take: "Two days from zero to autonomous production" is a sentence that would have been science fiction three years ago. GRIIP represents the convergence of foundation models and physical automation — the same pattern we see in AI-powered quality control. As these systems mature, the ROI calculation for manufacturing AI shifts from "will it work?" to "how fast can we deploy it?"
Quick Hits
- Snowflake + OpenAI: $200M multi-year partnership makes GPT-5.2 natively available to Snowflake's 12,600 enterprise customers for agentic AI workflows.
- Perplexity Model Council: New system runs Claude, GPT-5.2, and Gemini in parallel to generate cross-validated answers — reducing hallucination by ensemble consensus.
- GPT-5.2 Instant Update: OpenAI shipped a February 11 style and quality update for more measured, grounded responses to advice-seeking queries.
- Anthropic on Google Cloud: Plans to use up to 1 million TPUs worth tens of billions — bringing over a gigawatt of AI compute capacity online in 2026.
Numbers of the Week
| Metric | Value | Context |
|---|---|---|
| Manufacturers with AI in roadmaps | 75% | Up from 20% "ready" reported in January |
| Spend per AI implementation | $2M+ | Over half of manufacturers at this level |
| Anthropic funding round | $20B+ | At a valuation north of $350B |
| GRIIP deployment time | under 2 days | From zero to autonomous robot cell |
What We're Watching
Multi-agent systems go mainstream. Anthropic's agent teams aren't alone — OpenAI's Codex and Google's Gemini are all building multi-agent capabilities. The pattern is clear: single-model AI is becoming a team sport. For enterprises, the question shifts from "which model?" to "how do we orchestrate agent teams across our workflows?"
Physical AI accelerates. Vention's GRIIP, NVIDIA's foundation models, and the robotics sector's $4B+ in new funding all point to 2026 as the year physical AI moves from demos to deployment. Manufacturing is the proving ground, but logistics, warehousing, and agriculture are close behind.
The VC dual-bet pattern could reshape AI pricing. When your investors are also funding your competitor, the pressure to differentiate shifts from model capability (everyone's good enough) to enterprise integration, reliability, and total cost of ownership. This is exactly what enterprise buyers should be optimizing for when evaluating AI vendors.
The Bottom Line
This week marked a maturation point for enterprise AI. Agent teams, not solo models, are becoming the deployment unit. Manufacturing investment has crossed from experimental to strategic. And the venture capital world has decided this market is too important to bet on just one winner.
For enterprise leaders, the takeaway is clear: AI capability is no longer the bottleneck — orchestration is. The companies that move from POC to production fastest won't be the ones with the best models. They'll be the ones who figure out how to make AI agents work together, on real data, inside real workflows. The building blocks are here. The question is who builds with them first.
Get This Week in AI & Automation delivered every Saturday.
Previous Editions
Have a story we should cover? Contact us.
Need help with AI implementation?
We build production AI systems that actually ship. Not demos, not POCs—real systems that run your business.
Get in Touch