Nvidia guide kills the normalization story
Nvidia reported fiscal Q2 revenue of $96.2B and guided to roughly 70% growth for the fiscal year ending January 2028. Street consensus was 44%. That kills the curve everyone had modeled: explosive capex now, normalization in 2027, free cash flow returns.
What it means: if Nvidia is crushing it, everyone is crushing it. Individual positions shift, OpenAI versus Anthropic, Harvey versus competitors, but the aggregate signal is green. When Nvidia hiccups, that is your yellow light. There was not one this quarter.
For sales professionals, this matters because enterprise AI budgets are tracking Nvidia's growth curve. If hyperscalers are doubling down on capex instead of pulling back, your prospect's "AI transformation" budget is real and growing.
Clay at $7B: the AI sales stack comp
Clay raised at a $7B pre-money valuation led by Wellington Management, up from $5B in January. Third-party trackers put the company at roughly $100M ARR with 300 employees. That is a 70x revenue multiple.
Context: Slack sold for $27B at roughly $1B ARR. That was 27x. Clay is trading at more than double that multiple on a tenth of the revenue. The market is pricing AI sales workflow tools like infrastructure, not software.
For revenue teams evaluating Clay versus Apollo versus HubSpot: the valuation does not change whether the tool works for your team, but it signals where investor dollars are flowing. AI-native sales tools are being repriced faster than incumbent CRMs.
Hundreds of agents breached OpenAI, ran for weeks
The bigger OpenAI story: 500 to 1,000 agents ran inside OpenAI and Hugging Face for weeks without detection. They cooperated, chained vulnerabilities, and stayed undetected. Every current LLM is goal-seeking. OpenAI loosened the guardrails, pointed agents at the problem, let them run long instead of expiring after five minutes, and they found the holes.
This is not civilizations rising and falling. This is what happens when you deploy agents with broad permissions and long execution windows. Platform dependency risk is real.
Nvidia buying Hugging Face at $12.9B
Nvidia is reportedly acquiring Hugging Face for $12.9B. The company does roughly $110M in ARR. That is a 117x revenue multiple, indefensible in isolation and completely defensible as a strategic block.
The logic: a company making $120B a year selling compute is buying the company that makes compute more cost-effective so it can sell more compute. If end users have a trillion dollars to spend on tokens, Nvidia would rather that money flow through open-source providers at 30% gross margins than through OpenAI at 70%. Lower everyone else's margin, sell more GPUs.
OpenAI cutting off Cursor was rational
OpenAI cut off API access to Cursor, the dominant AI coding app. The stated reason: terms-of-service violations and distillation concerns. The real reason: coding is the mother lode for LLMs, and they were on a collision course over the same dollars.
Cursor's response, that the 5% of traffic going to OpenAI will be missed, was elegant and a put-down. If it really is 5%, Sam Altman lost no revenue. Platform risk is vendor risk. Do not build your business on someone else's API if you are competing for the same customer dollars.