Hyperscaler FDE Units: How AWS, Microsoft and Google Cloud Embed Engineers With Customers

Three hyperscalers, three models

All three units put engineers inside customer organizations to take AI from prototype to production, and all three are paid for by the cloud provider itself. They differ in who does the work: AWS and Microsoft deliver directly with their own engineers, while Google Cloud embeds its engineers alongside systems integrators and funds partners to do most of the delivery.

AWS Forward Deployed Engineering

AWS announced its Forward Deployed Engineering unit on 06-30-2026 with a $1 billion commitment that covers scaling the engineering team, market development funds, cash and credit investments, and partner allocations¹. It is led by Francesca Vasquez, VP of Frontier AI Engineering and Services, and is funded entirely from AWS's own balance sheet¹². The Next Web reported it as the first hyperscaler commitment to an FDE unit at this scale².

The engagement model is the most detailed of the three. Engineers work in pods of five or six, each embedded with one customer at a time, alongside the customer's business, engineering and security teams². AWS packages the pace as "AI 45": 45 minutes to an idea, 45 hours to a prototype and about 45 days per engagement, varying by use case¹. Pricing is fixed and outcome-based rather than billed by the hour, and every engagement is designed to end with an exit strategy so the customer can run the system itself¹. AWS says it will use enterprise-wide semantic context and ontology in "probably almost all" engagements to keep agents consistent¹.

Early customers named in coverage include the Allen Institute, Cox Automotive, the National Football League, the National Basketball Association and Ricoh, with regulated industries such as financial services, healthcare and the public sector next¹². AWS describes the unit as complementary to its partners and says it expects to work with the FDE arms of both Anthropic and OpenAI¹².

The launch also drew the sharpest public skepticism of the year. Constellation Research CEO R "Ray" Wang said at the launch: "I'm tired of looking at fake FDEs," and Constellation noted it was not yet clear whether AWS's engineers would combine technical and business depth the way Palantir's do¹. See Forward Deployed Engineers for that debate.

Microsoft Frontier Company

Microsoft announced the Microsoft Frontier Company on 07-02-2026, two days after AWS, as a $2.5 billion operating business of about 6,000 industry and engineering experts who work alongside customers to build, launch and keep improving AI systems at scale³⁴⁵. Judson Althoff, CEO of Microsoft's commercial business, announced it and positioned it as more than an FDE team: "This goes beyond what has been labeled as Forward-Deployed Engineering," he wrote, saying it "will be the largest, most capable, outcome-driven engineering organization in the industry"⁵.

HCAMag reports that Rodrigo Kede Lima, president of Microsoft Asia, leads the company, and that its focus is what Microsoft calls "Frontier Transformation"⁴. Launch partners named in coverage are the London Stock Exchange Group, Unilever, Land O'Lakes and Accenture⁵. In the LSEG example, Microsoft's engineers and industry experts embedded AI into LSEG Workspace so finance professionals can get answers from structured and unstructured financial content⁴. Althoff also said customer data is not used "to train models in ways that commoditise what differentiates them"⁴.

Microsoft has not published pricing, a delivery method or a headcount breakdown. The 6,000 figure is the largest stated FDE-style workforce of any 2026 entrant, but Microsoft has not said how many of those people are new hires rather than existing staff assigned to the company.

Google Cloud's partner-led approach

Google Cloud announced a $750 million fund on 04-22-2026 at its Next conference in Las Vegas to help partners in its 120,000-member ecosystem move joint customers' agentic AI projects into production⁶. The fund covers AI value identification, prototyping, agent building and deployment, upskilling, and teams of embedded Google forward deployed engineers⁶. Google will embed those engineers alongside Accenture, Capgemini, Cognizant, Deloitte, Devoteam, HCLTech and TCS to support deployments and resolve difficult technical problems⁶.

Nine AI-native services firms are launching Gemini Enterprise practices under the program: Altimetrik, Artefact, Covasant, Deepsense, Distyl, Northslope, Quantium, Tribe.ai and Tryolabs⁶. Two of them sit elsewhere on this site: Distyl is an independent DeployCo, and Northslope agreed to be acquired by the OpenAI Deployment Company in July 2026 (see DeployCo Companies).

Google Cloud CEO Thomas Kurian followed in May 2026 with a hiring call for more forward deployed engineers "to help us scale customer AI transformation," noting that "having FDEs is not new for Google Cloud"; CIO Dive counted 59 open FDE roles in the US, London, Paris and Hong Kong, with applied AI FDE base salaries of $127,000 to $183,000 in New York and Atlanta⁷. Google has not announced a standalone unit or a headcount target.

How hyperscaler units differ from DeployCos

A hyperscaler unit is a third shape alongside the lab-backed and independent DeployCos described in Types of DeployCos. The differences that matter to a buyer:

• Funding and ownership. The OpenAI Deployment Company and Ode were set up with private equity partners; the three hyperscaler units are funded internally with no outside investors¹²⁵.
• What is being sold underneath. A lab-backed DeployCo pulls through its parent's models. A hyperscaler unit pulls through cloud consumption, and AWS and Google Cloud both offer models from several labs, so the unit can be more model-neutral while still tied to one cloud.
• Channel conflict. Every hyperscaler depends on systems integrators. AWS calls its unit complementary to partners, Microsoft launched with Accenture, and Google routes most of its money through partners¹⁵⁶. How work is split when a hyperscaler pod and an integrator want the same project is not public.
• Pricing. AWS is the only one of the three, and one of the few 2026 entrants, to state fixed, outcome-based pricing¹. See DeployCo vs Consulting vs SaaS.

For questions to put to any of these units before signing, see How to Evaluate a DeployCo; the exit-plan and lock-in questions matter most when the partner also runs the cloud.

Frequently asked questions

References

1. Constellation Research, "AWS launches forward deployed engineering unit with $1 billion investment," 06-30-2026. constellationr.com
2. The Next Web, "AWS commits $1bn to forward deployed engineers for enterprise AI," 06-30-2026. thenextweb.com
3. TechRadar, "Microsoft is spending $2.5bn on deploying AI engineers to its customers," 07-2026. techradar.com
4. HCAMag, "Microsoft invests $2.5 billion to help companies embed AI at work," 07-06-2026, citing Microsoft's 07-02-2026 blog post. hcamag.com
5. TechBriefly, "Microsoft backs Frontier Company with 2.5 billion dollars," 07-03-2026. techbriefly.com
6. HPCwire (Google Cloud press release), "Google Cloud Commits $750M to Accelerate Partners' Agentic AI Development," 04-22-2026. hpcwire.com
7. CIO Dive, "Google Cloud is hiring an army of AI deployment engineers," 05-14-2026. ciodive.com