The AI labs just became consulting firms

Hours apart yesterday, Anthropic and OpenAI both announced something the consulting industry has been quietly worried about for two years: enterprise services firms of their own, with $5.5 billion of Wall Street capital behind them. Today's issue is the read on what actually changed and who should be paying attention.

In today’s issue:

• Main story: The AI labs just became consulting firms

• AI Headlines since Friday

• Free webinar: Build Notion Agents to Automate Complex Tasks, a 30-minute Lightning Lesson this Thursday at 2 pm ET. Save your seat →

$5.5 billion of Wall Street capital just rewrote the implementation layer for enterprise AI. The consulting industry has not had a competitor with this alignment before.

Anthropic and OpenAI both announced enterprise services firms yesterday, hours apart. Anthropic teamed with Blackstone, Hellman & Friedman, and Goldman Sachs on a $1.5 billion venture. OpenAI lined up TPG, Brookfield, Advent, and Bain Capital, raised $4 billion against a $10 billion valuation, and named the new entity The Deployment Company.

The headline is "AI labs move into consulting." The thing that actually changed is who is going to install AI inside your company two years from now, and the answer is now a different person than it would have been a week ago.

AI READY IN 21 DAYS · FREE

A 21-day AI program. In your inbox. Then it ends.

One email every morning at 7 am ET. Each one is a short read and one real thing to try before lunch. By Day 21, you have nine concrete capabilities, including prompting that hits your bar on the first draft, AI workflows that take five hours a week off your plate, and the language to lead the AI conversation at your company.

Built for ops leaders, COOs, chiefs of staff, founders, and team leads at mid-market companies. The four weeks move you from foundations (effective prompting, catching hallucinations, learning AI with AI) to connect (a personal context layer, daily automation, five hours a week back), to build (reusable skills, agent design and debugging), to lead (tool and model evaluation, leading the conversation at work). Twenty-one mornings, no upsells, no community Slack.

For three years, the implementation layer between a frontier model and a Fortune 5000 customer has been the systems integrator: Accenture, Deloitte, PwC, and the named partners inside the Claude Partner Network. The labs sold the model, and the SIs sold the deployment. Yesterday's announcements broke that arrangement. Anthropic Applied AI engineers are now embedded directly inside the new firm, on engagements the announcement names explicitly: community banks, regional health systems, mid-sized manufacturers. OpenAI's Deployment Company is structured similarly, with 19 PE and consulting backers reaching thousands of portfolio companies. A $5.5 billion combined raise is not the size of a co-marketing deal.

Forward-deployed engineering is now a category, not a job title

Anthropic's announcement names the engagement model directly: "Applied AI engineers from Anthropic will work alongside the firm's engineering team to identify where Claude can have the most impact, build custom solutions, and support customers over the long term." Fortune was more explicit: the structure mirrors Palantir's forward-deployment model.

This matters because forward-deployed engineering is the talent profile that has been impossible to hire at scale. There are not enough people who can sit inside a healthcare network for six weeks, understand the prior-authorization workflow, and ship a working agent against it. Alex Konrad reported last week on a startup hiring future founders specifically to work as FDEs. The $5.5 billion announced yesterday is a vote that this scarcity is the binding constraint on enterprise AI revenue, not model quality.

Private equity is the channel, not the customer

The investor list on each deal deserves attention. Blackstone alone has more than 250 portfolio companies, and the Anthropic press release names PE-backed CFOs as a primary buyer. On the OpenAI side, TPG, Bain, Advent, and Brookfield bring their own portfolios.

The customer profile inside the Anthropic announcement is also specific: the engagement scenarios cite a "multi-site healthcare services group" doing prior authorizations and a "community bank" that "lacks the in-house resources to build and run frontier deployments." That is not a Fortune 50 description. The Big Three SIs do not staff at the price point a community bank can pay. AI agent vendors selling per-seat licenses cannot, on their own, deliver an integration with a regional system. The new firms are pointed at that gap, and Wall Street has now underwritten them to operate inside it at scale.

The sales motion changes, too. A PE-backed CFO under sponsor pressure to embed AI in finance and operations is not going to issue an RFP. The sponsor will introduce a partnership with the new firm, the engagement will start, and the consulting decision will look more like a portfolio mandate than a vendor selection.

The consulting industry just lost its monopoly on AI implementation talent

Anthropic kept the Claude Partner Network alive in the announcement. Accenture, Deloitte, and PwC are still inside it. That is the polite framing.

The actual signal is that the partners with the deepest engineering capability inside the AI labs are now sitting on the other side of the table from the SIs, with Wall Street co-owners and explicit pricing pressure on what an SI engagement should cost. Aaron Levie posted yesterday that the work to implement agents at scale will create demand for "existing consulting firms, new ones that emerge, FDEs from agent vendors, or new internal agent engineering roles." The five-billion-dollar version of his thesis dropped a few hours later.

Where the bear case lives

The strongest version of "this is channel strategy, not a structural shift" is fair. The labs need distribution into the mid-market, the SIs are slow and expensive, and PE is the buyer with the most aggressive AI mandate from its sponsors. A jointly-owned vehicle with PE capital and lab engineering talent is the obvious answer, and obvious answers do not always disrupt categories. The new firms also still depend on Accenture and Deloitte for the global delivery footprint; a $1.5 billion vehicle cannot stand up; they look more like specialty boutiques aimed at PE portfolios than like McKinsey replacements. What is genuinely new is the ownership structure: for the first time, the model vendor and the implementation team share a P&L on a specific category of customer. That alignment was not available a week ago.

What this changes

For most readers, this is a watching brief, not a to-do list. The implementation layer has new owners, and the talent market behind it just got more competitive, but the rollout will play out over twelve to eighteen months.

A few things worth tracking as it does. If your company is PE-backed, your sponsor's relationship with these vehicles is now a budget input; the cost basis on AI services is probably coming down for portfolios with access. If you have consulting RFPs in motion, the question for bidders is whether they have access to forward-deployed engineers from the model vendor and at what cost. And for everyone else, the labor-market read is the one that matters: forward-deployed engineering talent was thin a week ago and is now being absorbed by capital.

The AI lab no longer just sells you the model. It sells you the team that installs it, the firm that finances the installation, and the long-term contract underneath both. The consulting industry was not built to compete with that.

SPONSORED BY CLUTCH

Hire secure AI teammates that work 24/7.

Hire pre-built AI teammates. Give your engineers and operators a platform to ship their own AI apps. Stop losing sleep about what is running where.

Clutch is the platform behind both: pre-built agents for the workflows your ops team should automate first, plus the integration plane your team's vibe-coded apps and Claude Code projects plug into. One platform. Real production. Visible and safe by default.

Built for ops, engineering, and security teams that are tired of the shadow-AI surface area inside their own company.

SINCE FRIDAY

• OpenAI and Anthropic both launched enterprise AI services joint ventures yesterday: Anthropic's $1.5 billion firm with Blackstone, Hellman & Friedman, and Goldman Sachs, and OpenAI's $10 billion-valuation Deployment Company backed by TPG, Brookfield, Advent, and Bain Capital. The implementation layer for enterprise AI just got two new owners, both of them tied to Wall Street capital.

• Cerebras updated its IPO prospectus yesterday to sell 28 million shares at $115 to $125, targeting a $26.6 billion valuation, with Amazon Web Services named as the first hyperscaler customer and a $20 billion OpenAI contract attached. The strongest public-market read so far on AI chip diversification away from Nvidia.

• Cursor open-sourced its internal Team Kit, the slash-command set the Cursor team uses on itself, including a verify-this skill that forces an agent to gather baseline-and-treatment evidence before claiming a change worked. Useful reference for any operator running an internal "make AI prove its work" workflow.

• a16z published Workday's Last Workday?, arguing that the dominant enterprise software incumbents are exposed to the services-as-software shift, with implications for HR, finance, and ITSM stacks that mid-market operators are sitting on right now.

• Ben Thompson's Stratechery interview with Sam Altman and Matt Garman ran last Monday and is the closest read available on how OpenAI now thinks about agents-as-managed-service. Required context for the Tuesday announcement.

REPLY PROMPT

Hit reply and tell me one thing: if your company is on the buy-side of this shift, who is the implementation partner you are actually planning to use for AI in the next 12 months? An SI, an internal team, one of the new lab-PE vehicles, or something else? I read every reply.

FORWARD THIS

If a colleague is sitting in a conversation about "who do we hire to actually deploy this AI we bought," forward this issue to them. That is the conversation this Memo is for.

IF YOU GOT VALUE FROM THIS ISSUE

Keep going. The next 21 mornings, in your inbox.

The analysis you just read came out of the same operator practice that the 21 Days program teaches. Foundations, then connect, then build, then lead. By Day 21 you own nine concrete capabilities: prompting that hits your bar on the first draft, AI workflows that take five hours a week off your plate, agent skills your team can run, and the language to lead the AI conversation at your company.

Free. Twenty-one mornings, then it ends. No upsells, no community Slack. Inspired by Hilary Gridley's Couch to 5K for AI.

Not sure if it is for you? Take the 2-minute diagnostic.