DXC and Quantiphi signed LangChain in 48 hours; HCLTech chose Anthropic instead
DXC on October 13 and Quantiphi on October 14 signed LangChain's stack, and the releases show DXC licensing LangSmith, the paid observability platform, while LangGraph's open-source runtime carries no license fee. HCLTech picked Anthropic as its rival rail two days earlier, and none of the three deals will say what any of it pays.
Vincent Jiang · 3 min read
LangChain$25M — LangChain, private, latest valuation $25M signed two enterprise IT-services partners in two days this week: DXC Technology+5.08% — DXC Technology, up 5.08 percent today (NYSE: DXC), the Virginia-based operator of mission-critical IT estates, on October 13, 2026 1, and Quantiphi, an AI-native services firm, on October 14, 2026 2. Neither release names a customer, a price, or a revenue figure 3. What they do disclose is the money structure, and it is the part nobody has read closely: DXC licenses LangSmith, LangChain's paid platform for testing, evaluating and observing agents 1, while LangGraph, the open-source runtime orchestrating DXC's OASIS agents, carries no license fee per the release 1. The releases do not say whether Quantiphi licenses LangSmith the same way.
Services giants are selling the governance of their own labor
The pattern is bigger than one vendor. Within roughly 48 hours, DXC and Quantiphi re-cut themselves as the governed-deployment layer for AI agents 3, while HCLTech, a third services giant, signed a Global Premier Partnership with Anthropic on October 12, 2026, with a dedicated Anthropic$350B — Anthropic, private, latest valuation $350B business unit and Claude embedded in HCLSoftware products 6. The firms that once sold the human labor for enterprise IT now sell the infrastructure that governs the automation of that labor. Calling each platform choice a rail that binds customers is this piece's framing, not the releases' language, but the picks are real: LangChain for DXC and Quantiphi, Anthropic for HCLTech 126.
DXC printed numbers, and they are a seller's numbers
DXC printed numbers, which none of the recent deals did. Across 500 measured use cases, its OASIS platform cut incident resolution time by 95% on average, with diagnostic accuracy of 92% or better, validated against DXC's own senior engineers 1. Filesystem incidents fell from 53 to 90 minutes to 3 to 6; problem management that ran 2 to 6 days now runs 8 to 10 minutes 1.
DXC reports minutes instead of days once its agents run the fix
- Before OASIS
- After OASIS
Data
| Before OASIS | After OASIS | |
|---|---|---|
| Filesystem incident | 90 (53–90) | 6 (3–6) |
| Database incident | 180 (120–180) | 12 (5–12) |
| Problem management | 8,640 (2,880–8,640) | 10 (8–10) |
These are DXC-reported figures checked against DXC's own staff, with no independent benchmark and no named customers 3. An industry audit cited in the coverage found only 22 of 150 documented AI deployments disclose cost or financial impact at all 3, so DXC's 95% is the best evidence on the table and still a seller's evidence.
The observability meter sits in the most crowded slot
The read-through for investors is DXC, which is scaling OASIS toward 125 customer environments 1, adding LangSmith licenses as a new cost line on the way. The pressure on LangChain is that its meter sits in the most contested part of the stack, and CIO's own ecosystem map proves it: it lists LangSmith among observability upstarts alongside Arize, Braintrust and Galileo, with Datadog−2.54% — Datadog, down 2.54 percent today and Dynatrace−0.97% — Dynatrace, down 0.97 percent today as leaders and Langfuse as a free open-source alternative 4. When a competitive survey names your paying product as one of many, the slot is crowded. The integrators picked their rails, but the meter they run has competition at every price point, and neither signed deal will say what it pays.
Watch the next DXC earnings cycle for whether the 125-environment target and the LangSmith licenses show up in reported spending, and whether HCLTech's Anthropic unit repeats its Advanced AI revenue disclosure, which reached $219 million, up 78% year over year, in the September quarter 6.
Deepdive
AI-generated from this story and its cited sources. Not investment advice.
