AI news5
Anthropic launches enterprise AI services joint venture with Blackstone, Hellman & Friedman and Goldman Sachs at a $1.5 billion valuation
Anthropic announced on 4 May a joint venture with Blackstone, Hellman & Friedman and Goldman Sachs to build an AI-native enterprise services company that will compete head-on with the major consulting firms for corporate AI transformation work. The Wall Street Journal first reported the venture is valued at $1.5 billion with a $300 million commitment each from Anthropic, Blackstone and Hellman & Friedman, alongside backing from Apollo Global Management, General Atlantic, GIC, Leonard Green and Sequoia Capital. The venture will use a forward-deployed engineer model popularised by Palantir, with engineering teams sitting alongside clinicians, IT and other staff to build into existing workflows.
fortune.com ↗Why it matters
Frontier-AI labs are now directly entering the consulting layer that nonprofits typically buy from McKinsey, BCG and Deloitte, and any sector AI engagement quoted in 2026 needs to be benchmarked against this new vendor class before signing.
OpenAI parallels with The Deployment Company at a $10 billion valuation, $4 billion raised from 19 investors including TPG, Brookfield and SoftBank
Bloomberg reported hours before the Anthropic move that OpenAI has raised more than $4 billion for The Deployment Company, a new majority-owned joint venture valued at $10 billion that will help businesses adopt OpenAI software. Backers include TPG, Brookfield Asset Management, Advent, Bain Capital, Dragoneer Investment Group and SoftBank Group, who collectively bring access to more than 2,000 portfolio companies and clients. Brad Lightcap, OpenAI's COO, was promoted last month into a new special-projects role overseeing the enterprise software push that the venture is now built around.
techcrunch.com ↗Why it matters
The two largest US AI labs have launched competing enterprise distribution arms inside the same 24 hours, which means any nonprofit running an RFP for AI implementation now has two distinct vendor options whose incentives are tied to enterprise deal volume, not just API consumption.
NVIDIA ships Nemotron 3 Nano Omni as a 30B-A3B open-weight multimodal model with 9x agent throughput
NVIDIA released Nemotron 3 Nano Omni on 28 April as an open multimodal model unifying vision, audio, image and text in a single hybrid mixture-of-experts architecture with 30 billion parameters and roughly 3 billion active. The model claims 9x higher throughput than other open omni models at the same interactivity level, eliminating the need for separate perception models. Open weights, datasets and training techniques shipped alongside the model on Hugging Face, OpenRouter and build.nvidia.com, with deployment paths from NVIDIA Jet on local devices through DGX Spark and into data centres.
blogs.nvidia.com ↗Why it matters
A single small open model that handles voice, vision and text agentically lowers the technical floor for any nonprofit considering on-prem or sovereign AI deployment, and the open-weight licence means no per-token vendor lock-in to a frontier lab for multimodal agents.
OpenAI open-sources its monitorability evaluation suite for chain-of-thought safety auditing of frontier models
OpenAI's alignment research team open-sourced a subset of its chain-of-thought monitorability evaluation suite on 23 April, releasing intervention, process and outcome-property datasets and reference code via github.com/openai/monitorability-evals. The release includes the g-mean2 metric implementation and a new cross-fit filtering strategy that improves intervention-style estimates against noise-dominated test instances. OpenAI also back-filled monitorability results for GPT-5.4 thinking, GPT-5.2 thinking, GPT-5 thinking and o3 in their published system cards.
alignment.openai.com ↗Why it matters
Any nonprofit deploying frontier reasoning models in high-stakes workflows like grant decisions, clinical triage or legal aid now has an open evaluation suite to argue procurement against, and the released code is the first vendor-built open-source toolkit for chain-of-thought safety auditing.
WashU Olin Business launches first-of-its-kind $10K AI Workforce Transformation Scholarship and new MS in AI for Business
Washington University Olin Business School announced on 4 May a $10,000 AI Workforce Transformation Scholarship paired with a new one-year Master of Science in AI for Business launching in fall, with three industry tracks across finance, marketing and technical disciplines. In parallel Olin launched AI Transformation for Business, an executive-education series delivered inside organisations to translate AI experimentation into enterprise-wide impact. The programme reflects Olin's nearly decade of consulting collaborations across retail, sport and pharmaceuticals.
source.washu.edu ↗Why it matters
A dedicated reskilling scholarship for displaced workers and recent graduates is the cleanest tertiary signal yet that AI workforce displacement is being treated as a degree-level intervention, and any AIH member running training programs can use Olin's curriculum framing as evidence that AI adoption needs paired workforce strategy not just tooling.
AI in the nonprofit sector4
Australian Centre for AI in Marketing launches Unlocking the Power of AI for Good with IBM ANZ and 12+ leading AU charities
ACAM launched a four-week online programme called Unlocking the Power of AI for Good on 4 May for AU social-impact-sector leaders, with IBM ANZ as backing partner. The programme was shaped through ACAM's AI for Good Think Tank with Wayside Chapel, Gidget Foundation, Food Ladder, Pankind, Canteen, Variety - the Children's Charity, Black Dog Institute, Barnardos, Autism Awareness, Lifeline, plus support from UNICEF Australia, World Vision and The Smith Family. Founder Louise Cummins, former World Vision Australia CMO, framed the gap as not access to technology but knowing what to prioritise, what to refuse and how to embed AI into real workflows.
ecommercenews.com.au ↗Why it matters
This is the first IBM-backed AU NP AI capability programme co-designed by the largest charities in the country, and it sits as a direct neighbour to AIH on the practical-AI-for-impact map, worth tracking for both partnership leverage and competitive positioning.
Uganda Re-Imagining TB Care project deploys AI screening chatbot in four local languages for Village Health Teams
Uganda's Re-Imagining TB Care project launched an AI-powered screening chatbot on 28 April for Village Health Teams, the country's frontline community health workforce that connects communities to formal TB services. The platform shipped in Lumasaba, Lusoga, Luganda and Runyankore to match local health worker languages, supporting early TB detection and on-the-spot health education. Funding sits with the Korea International Cooperation Agency, with implementation backed by the US Centers for Disease Control and Prevention and the Stop TB Partnership.
africa-press.net ↗Why it matters
This is a working template for any health-focused nonprofit asking how to deploy AI in low-bandwidth multilingual community settings without losing trust, and the four-language rollout shows the model can be localised faster than most legacy translation pipelines.
International Rescue Committee scales Signpost AI to 4,700 teachers in northern Nigeria and launches Alma multilingual newcomer assistant in the US
The International Rescue Committee published an update on 28 April reporting that AprendIA, its Signpost-platform AI tutor for teachers in Borno, Adamawa and Yobe states in Nigeria, has scaled from roughly 500 to more than 4,700 teachers and is on track to reach 22,000 by end-2026. The same Signpost platform now powers Alma, a multilingual virtual assistant that helps newcomers in the US navigate work permits, driver licences, school enrolment and housing systems by drawing on verified IRC sources with escalation to human caseworkers when needed. The IRC frames the work as purpose-designed AI agents with defined boundaries, evaluated extensively before launch.
restofworld.org ↗Why it matters
This is the cleanest end-to-end frontline-deployment story of 2026 for a nonprofit that scaled an AI tool from pilot to country-program scale, and the Alma case is the first concrete refugee-resettlement AI agent worth showing to AIH funder partners exploring humanitarian deployments.
VNS Health and Nestmed scale ambient AI clinical documentation across New York's largest home-based-care nonprofit
VNS Health, a 133-year-old New York nonprofit and the largest home and community-based care organisation in the state, partnered with Nestmed on 1 April to roll out ambient AI clinical documentation across home-care teams, becoming the first NY organisation to scale the technology at this size. Nestmed is the leading AI-powered documentation platform for home health and hospice, with more than 15,000 clinicians and millions of annual visits already on the platform. The deployment targets the time-consuming after-visit documentation that home-care clinicians currently extend beyond patient-contact hours.
reporter-times.com ↗Why it matters
Aged-care and disability-services nonprofits in Australia run almost identical documentation-burden patterns, and the VNS Health rollout is now the largest production reference for ambient AI in home-based community care, useful as a procurement comparable for AIH members in those subsectors.