Hi operators and happy $SPCX IPO day for those who celebrate. This week saw a huge amount of backlash against Anthropic for degrading performance based on specific tasks. We heard versions of this from a huge range of AI services founders as the world begins to grapple with the reality on a small number of companies controlling widely used infrastructure. “I think it's going to change which models people decide to adopt, not if”, said Joshua Wohle from AI training startup Mindstone. “But lots of companies were defaulting to Anthropic based on good vibes, and I think they're going to be questioning that path now. I personally think open source is going to make real strides over the next six months". It will be interesting to see how clients respond to this. Let us know what you’re hearing.
On that note, we have a useful longer piece on the other AI partner ecosystems you should be considering for clients too.
-Daria
💰 Market Monitor (AI services M&A + Investment)
Tracking developments across the AI enablement services sector. Have we missed an interesting deal? Or do you want to correct details? Just hit reply to let us know.
Notable Investments
(UK, data & AI consulting) Advancing Analytics has received a strategic growth investment from Lead Edge Capital. The London-based Databricks consultancy, founded in 2018, embeds delivery teams alongside clients in financial services, healthcare and retail to build AI-ready data platforms. Amount undisclosed. [Link]
(UK, AI platform) PhysicsX has raised $300M in a Series C round at a $2B+ valuation. The London-based company develops AI-native simulation and engineering platforms to accelerate hardware design, with Amazon and Nvidia among backers. [Link]
(US, AI consulting) Ramp launched an AI services motion built around forward-deployed engineers specialising in finance operations. The core thesis: the gap between "we have AI tools" and "agents run our workflows" is where the real value is, and most enterprises can't close it alone. Notable signal of a SaaS company with a data moat explicitly moving into professional services. [Link]
(US, AI training) Google committed $50 million to train 300,000+ skilled trade workers across the US - electricians, fiber technicians, welders and HVAC specialists - through 14 labor unions and four trade associations, focused on the workforce needed to build and maintain data centres. The announcement came alongside eight policy proposals on workforce development. Meta ($115M) and Anthropic ($150M Claude Corps fellowship) made similar commitments the same week, reflecting a growing consensus that labor availability, not capital or compute, may be the binding constraint on AI infrastructure expansion. [Link]
Notable M&A
(UK, AI implementation) Scale Factory has been formed by the merger of Ten10 and The Scale Factory. The combined UK consultancy covers quality engineering, cloud, data and AI, targeting regulated sectors including financial services, defence and the public sector. AWS Advanced Consulting Partner, ISO 27001 certified, 3,500+ projects delivered. [Link]
(US/UK, AI services) Whalar has been acquired by Accenture. The New York-based creator and social agency joins Accenture Song, expanding its influencer marketing and creator-economy capabilities. [Link]
(US, AI services) Knownwell has been acquired by 2X to form what the combined company is calling the first human-agentic GTM services business. Knownwell brings agentic AI engineering, commercial intelligence and data science capability; the deal values the combined entity at over $400M. [Link]
(US, professional services) Crowe is selling a majority stake to KKR in a deal valued at nearly $3 billion. The top-10 US accounting and advisory firm becomes one of the largest professional services firms to take on private equity ownership. [Link]
(US, AI implementation) Griffin Global Technologies has been acquired by Doeren Mayhew. The Alpharetta-based software development, AI automation and data analytics firm joins the top-50 accounting firm as its second tech acquisition of the year. [Link]
🙌 Partner Program Updates (models + platforms)
Tracking emerging partner programs for frontier models, cloud and leading AI application vendors. Did we miss a partner program update from any of the major players which you feel is important?
Anthropic: Formalised the Claude Partner Network this week with a cluster of major SI sign-ups ahead of its IPO. TCS joined as a Global Premier Partner, deploying Claude to 50,000 associates with a joint go-to-market across financial services, healthcare and public sector. [Link] DXC Technology signed a multi-year Global Premier partnership to integrate Claude into the systems of banks, airlines and other regulated enterprises. [Link] CI&T joined with 1,000+ certified AI engineers. [Link] And Bounteous joined as a launch Preferred Partner. [Link]
Google Cloud: NTT DATA expanded its partnership to deploy agentic AI solutions built on Gemini Enterprise at scale — the latest in a pattern of large SIs formalising lab-aligned delivery capabilities. [Link]
Microsoft: Atos expanded its partnership with Microsoft to deploy secure agentic AI across its own operations and client environments — part of the broader movement of large European SIs into agentic delivery via lab partnerships. [Link]
IBM Consulting: Launched a joint Google Cloud Practice with Google, combining IBM Consulting Advantage with Gemini Enterprise and staffed by forward-deployed engineers, targeting government and regulated-industry modernisation — described as a multi-billion-dollar services opportunity. [Link]
Deloitte: Launched "Adopt 100" with NVIDIA to accelerate AI adoption in large enterprises and connect AI startups with major corporate clients — a signal of Big-4 firms using chip-maker partnerships to scale AI-adoption consulting. [Link] Also: formed a strategic alliance with Ironclad to build agentic contract lifecycle management capabilities, signalling the pace at which Big-4 are building AI-legal and AI-operations practices. [Link]
Converteo: The Paris-based data and AI consultancy launched an international AI and agentic systems offering dedicated to Pharma & Life Sciences - an example of European AI consultancies building differentiated vertical practices rather than competing on generalist delivery. [Link]
Josh Bersin Company: Launched The Josh Bersin Institute, repositioning as a professional standards and capability-certification body with an AI-driven, research-backed HR and AI learning curriculum — entering the AI enablement credentialing space directly. [Link]
The operator essay: Four partner programs worth your time (or finding opportunities in less crowded places)
There is a lot of noise about OpenAI and Anthropic's partner programmes. For founders building in the field, becoming a certified partner of either company is both a valued credential and a possible fast track to an acquisition. Tomoro, whose acquisition we covered last week, is a case in point.
But OpenAI and Anthropic are not the only AI vendors running partner programmes. We track 10-plus AI vendors actively advertising one. At 10xHumans we are big believers in finding opportunity in less crowded places. Below we highlight four that represent interesting openings, and, critically, who we think would benefit most from partnering with each.
Clay: the next frontier in GTM
What. Clay is a GTM orchestration platform, starting with a data marketplace of 150-plus external sources, and enabling automation of everything from prospecting to ABM, signals-based outreach, and paid ads. It sits between your CRM and your go-to-market motion, replacing the manual work. It crossed $100 million ARR in January 2026, growing from $1 million in just two years. A formal Solutions Partner Programme launched in August 2025, though an organic community of agencies had already formed before that, driven largely by the fact that all but the most sophisticated customers struggle to find experienced talent to run Clay at full capacity in-house.
Why now. Clay has a partner team of twelve, which means partners who show up get attention. The commercial structure pays 20% referral revenue on new licences brought in, but the bigger opportunity is in implementation, consulting, and managed services, where reported monthly retainers run $3,000 to $15,000 per client. Not transformational on a single engagement, but GTM engineers are cheaper and more accessible than full forward-deployed engineers, which means the model scales. Clay is also making a deliberate push to partner with private equity funds: a lead channel we covered when we looked at the five things the fastest-growing AI services businesses are doing differently. For a services firm looking to build relationships with PE as a referral channel, this is a short cut in. Another interesting signal: in our AI GTM deep dive, a finding that divided opinion was that some leaders believe the space is moving too fast for in-house teams to keep up. Making working with a specialist agency a more practical choice.
Best fit. Clay has use cases beyond outbound (recruiting, dealflow, competitive intelligence) but to make the most of this partner opportunity, you need to be a GTM services firm. The natural fit is an established Salesforce or HubSpot implementation practice looking to add AI capability on top of an existing offering, or a B2B marketing agency that wants to move up the value chain from execution to strategy and automation.
Replit: the early mover’s window
What. Replit is an AI-powered software creation platform with 50 million users and enterprise deployments at 85% of the Fortune 500 - though you would not know it from its reputation. It is still widely perceived as a prosumer tool, a place where developers tinker. That gap between perception and reality is part of the opportunity. The company did $240 million in revenue in 2025. Less than two weeks ago it launched a formal Solution Partner Programme, with Accenture, Slalom, and Hexaware as founding enterprise partners.
Why now. The partner directory is almost empty. The founding enterprise partners are global SIs chasing Fortune 500 contracts; they will not compete for mid-market, or sector-specific clients. The commercial model includes a structured referral and co-selling programme, as well the opportunity to sell more services. The range of what partners are charging is wide: anecdotally from $335/ hour for basic website builds up to $7 million for a full application for a government client. Of the four programmes covered here, this is the most open window right now, measured simply by the fact that almost no one has walked through it yet.
Best fit. The sweet spot is serving less technical teams inside larger organisations. The ones for whom Claude Code is overkill but off-the-shelf SaaS tools are not flexible enough. These teams know they need custom internal tooling; they do not have the engineering resource to build it. Replit sits exactly in that gap. The firms best positioned to partner around it are those that can do two things at once: help clients discover where custom tools would actually move the needle, and then build and train on them in the same engagement. Training is half the value, not just an add-on. A client who cannot use what you built without you is not a successful deployment. If your practice is already oriented around internal process automation or operational tooling for mid-market and enterprise clients, and you can lead with consulting before you lead with build, this is a natural fit.
Mistral: European regulated-Industry specialist
What. Mistral is a French AI company and the only credible European alternative to OpenAI and Anthropic for organisations where data sovereignty is a hard requirement. It is targeting €1 billion in revenue in 2026. Its strategic partners include Accenture, IBM, Capgemini, SAP, and NTT. The core proposition is sovereign AI: models that can be deployed on-premise or within European cloud infrastructure, with full data residency and no US dependency.
Why now. Mistral is the only credible full-stack European AI option for regulated industries. Its partnership programme is selective and invitation-driven. There is no published tier structure or affiliate model, which means the commercial relationship is bespoke. That cuts both ways: harder to get into, but far less crowded. Accenture and Capgemini hold the top of the market. Below them, the mid-market specialist tier is wide open.
Best fit. This is not the right programme for a generalist AI services firm. It suits businesses with genuine vertical depth in financial services, healthcare, public sector, or defence-adjacent industries where data sovereignty is a hard requirement. The pitch to Mistral is not "we implement AI" but "we have existing enterprise relationships in [regulated sector] and can bring you into accounts that neither Accenture nor any US-model provider can enter." If your firm has that position, this is structurally the most defensible partnership on this list.
n8n: Technical Automation Builders With a Data Sovereignty Angle
What. n8n is an open-source workflow automation platform that has native LangChain integration, 70-plus AI nodes, persistent agent memory, and vector database support for RAG workflows. It is self-hostable, has no per-execution pricing at scale, and is increasingly the platform of choice for technical teams building genuine agentic systems rather than simple trigger-action automations. Its affiliate programme pays 30% revenue share for 12 months for referred new customers, though much of the money is made independently rather than through a formal rev-share.
What makes the opportunity distinct. The Expert Partner programme, the tier for implementation agencies, which carries different and unpublished commercial terms, was closed to new applicants in mid-2025 while n8n revamped it. It has not yet been relaunched publicly. The existing expert directory is thin. This means early positioning as a founding expert of the relaunched programme is available to a small number of firms that move now. Project-based engagements typically run $2,000 to $15,000 depending on complexity: a law firm CRM sync at the lower end, a real-estate lead-routing build at the higher. One detailed practitioner breakdown of a $20,000 project illustrates the margin profile well: process mapping ($2K), solution design ($6K), twelve production workflows ($5K), data warehouse ($2.8K), dashboard ($1.8K), and GDPR hardening ($2.4K), delivered over 67 days at roughly 35% net margin. Not enormous on a single project, but the work is repeatable and the skill set that commands it is genuinely scarce.
Best fit. This suits technically deep automation and AI engineering firms. The ideal client profile is mid-market or enterprise organisations with data residency requirements, sensitive data environments, or agentic ambitions that Zapier and Make simply cannot handle. Firms that position as "AI automation engineers" rather than "automation consultants" and can demonstrate production-grade system design will find n8n a more defensible home than the more commoditised Zapier and Make ecosystems.