Enterprise AI news, analysis, and strategy—daily coverage of AI models, agents, benchmarks, and real-world adoption, for technology and business leaders.
Upwind's stock deal for nine-month-old Aegis buys a team and a research lab with no ship dates, SKUs or prices. Don't credit AI-agent security toward a renewal until it ships, and write change-of-control terms into early agent-security pilots now.
September 24, 2026 · 8 minLiveKit bought Loophole Labs and its Substrate hypervisor to cut voice-agent startup below three seconds. The announcement says nothing about Architect, Loophole's paid Kubernetes spot-migration product, or its pre-GA AGX sandboxes.
September 24, 2026 · 9 minSnowflake Copilot is the wrong tool for executives querying in English. Cortex Analyst on Snowflake or Genie on Databricks wins — once you build the semantic layer; ThoughtSpot only earns its seat price across multiple warehouses.
September 23, 2026 · 13 minBlaxel says its agent sandboxes keep running under Baseten, but the promise has no end date, and its standard terms allow assignment without consent, renewal repricing and data deletion 90 days after termination.
September 23, 2026 · 9 minMistral bought Pimento to staff its Vibe assistant and has said nothing about the ad-creative tool 180 brands use. Export brand kits, confirm rights to past creative, and review its ad-account access now.
September 23, 2026 · 8 minZ.ai's ZCode packaged whole workspaces, 86.6% of it .git history, for an upload only Z.ai could decrypt, and its privacy toggles never stopped it. Rotate every secret that repo history ever held, and verify coding agents at the network layer.
September 23, 2026 · 9 minThe briefs technology leaders clicked most this week.
GPT-6 Sol costs $2/$10 per million tokens, half GPT-5.6 Sol's promo rate, permanently. But OpenAI's headline evals ran at xhigh or max effort, and only one carries a dollar cost per task.
September 23, 2026 · 9 minFor a vendor portal with no API, script the flow in Playwright, run it on Browserbase if you need scale, and pay Claude Computer Use only for the step that breaks. As the primary executor, a screenshot agent costs about 30x more in tokens and still hands you an exception queue.
September 23, 2026 · 13 minClaude Opus 5.5 cuts per-token prices 20% and cache reads 60%, but forced tool_choice and disabled thinking now return HTTP 400, the default effort drops to medium, and the 40% saving compares different effort levels.
September 22, 2026 · 11 minWeekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
OpenAI scored Astra for Law at 54% on Vals AI's private validation set, against its own GPT-6 Astra. On Vals' public leaderboard, three general frontier models already sit at 55.29% and Astra for Law is not listed.
September 22, 2026 · 9 minCognex is paying about $500M in cash for RealSense's depth cameras, but the RealSense ID face-authentication line, whose anti-spoofing runs on an online-renewed licence, is being spun out to a team of about 25. Here is what each set of buyers should get in writing before the Q4 close.
September 22, 2026 · 11 minAssemblyAI is the default for post-call transcription and Deepgram for live voice agents. Self-hosting pays past roughly 2-6M call minutes a month with a two-engineer team — and then the model to run is NVIDIA Parakeet, not Whisper.
September 22, 2026 · 20 minGrok 4.7's per-token price didn't move. Its per-task cost did: up 47% at matched effort on Artificial Analysis's index, down 10% on Cursor's own benchmark. Price it on your workload.
September 22, 2026 · 11 minBlackLine bought NetNow on September 21. NetNow's published terms let either side terminate without notice and make keeping copies of your credit files your job — export them, then get ERP, price and export commitments in writing.
September 21, 2026 · 11 minHarvey sold flat seats on top of a metered model bill, and agent usage drove its gross margin to -50% by June. The margin recovered after it launched its own Kimi K3-based model, with no reported price change, so renewals of agentic AI seats need model-change and usage terms.
September 21, 2026 · 11 minMicrochip closed its Hailo acquisition on September 21, ending the solvency risk. Its pledge not to discontinue parts covers manufacturable chips, not the closed Dataflow Compiler every Hailo model depends on.
September 21, 2026 · 11 minIdle GPUs are stranded by static quotas, not by the wrong scheduler. Kueue fixes quota borrowing for free; Run:ai earns its $2,600/GPU/year list price only on memory-enforced fractional GPUs or when NVIDIA AI Enterprise already covers it.
September 21, 2026 · 22 minAmazon cut off Meta's Muse agent, which runs in a Meta-hosted cloud browser with stored logins. The Ninth Circuit's Perplexity ruling covered agents in the user's own browser and left this architecture open, so site terms and detection now decide.
September 21, 2026 · 12 minA 40-call black-box audit from USTC flagged 7 of 15 OpenAI-compatible GPT services for behaviour consistent with output-token inflation. The flags are not proof, but padded answers pass quality evals and invoice checks alike, so measure output tokens per task against the first-party API.
September 21, 2026 · 13 minTypeSafe's Jev answers classification calls for 12-27x less than Claude Haiku 4.5, but the first independent tests show its accuracy depends on how you split the question and its probabilities need recalibrating per question.
September 20, 2026 · 13 minRe-embedding a 10M-chunk corpus costs about $520 at September 2026 rates, so the embedding model is not the one-way door teams fear. voyage-4-large is the default; OpenAI's model has had no successor in 32 months.
September 20, 2026 · 15 minPretrained transformers now top the tabular leaderboard, beating tuned XGBoost below 50,000 rows. The ceiling, the GPU bill and the protocol caveat decide whether it applies to your tables.
September 20, 2026 · 6 minSalesforce closed Dreamforce 2026 with two marquee reveals that nobody can buy yet. Koa is a pilot with a winter GA target; Claudeforce is an open beta with no published price. The shippable announcements were the agent operations tools.
September 20, 2026 · 8 minPlugin4Shell broke SHA pinning in Claude Code, Codex, Copilot CLI and the Gemini CLI. Two shipped client fixes; the control covering everyone else is a Git host naming rule that GitHub and GitLab enforce and Bitbucket and self-hosted servers do not.
September 20, 2026 · 12 min
