What $10 billion of “Forward Deployed Engineering” launches means for Asia’s enterprises and why the biggest opportunity just moved closer to home
By Sandeep Joshi, Founder & Managing Director, MASSIVUE
Something extraordinary happened in enterprise AI over the past eight weeks and most business leaders in Asia haven’t yet registered what it means for them.
On May 4, Anthropic announced a US$1.5 billion enterprise AI services venture with Blackstone, Hellman & Friedman, and Goldman Sachs backed, notably for us in this region, by Singapore’s GIC. Days later, OpenAI unveiled the OpenAI Deployment Company, raising US$4 billion from 19 investors including TPG, Bain Capital, and McKinsey, and acquiring the applied-AI firm Tomoro to staff it from day one. On June 30, AWS committed US$1 billion to a dedicated Forward Deployed Engineering organisation, sending pods of five to six engineers into customer sites for 45-day sprints. And on July 2, barely 48 hours later, Microsoft launched Frontier Company: US$2.5 billion, 6,000 embedded specialists, led by the executive who until now ran Microsoft’s Asia business.
Four launches. Nearly US$10 billion of committed capital. One shared conviction: the model is not the product anymore. The deployment is.
The quiet admission behind the noise
Strip away the branding and these announcements are a collective confession from the world’s most powerful technology companies: selling access to frontier AI has not been enough.
MIT’s NANDA research found that 95% of enterprise generative AI pilots produce no measurable business impact. The models aren’t failing – the implementation is. Enterprises bought the licences, ran the pilots, and watched impressive demos stall against the realities of legacy systems, data governance, compliance, and human workflows.
The industry’s answer is “Forward Deployed Engineering” – a term borrowed from Palantir, meaning engineers who embed inside a client’s operations and build working systems rather than delivering recommendations. It is, frankly, what serious transformation practitioners have always done. The giants have now validated it with ten billion dollars.
For those of us who have spent years arguing that AI transformation is an operating-model challenge, not a procurement decision, this is vindication on a global scale. The world’s largest technology companies just agreed that the hard part of AI is the part that happens inside your business, with your people, on your data.
But read the fine print: this wave is not built for Asia’s real economy
Look carefully at who these ventures are designed to serve.
Anthropic’s venture targets the portfolio companies of its private-equity backers. OpenAI’s DeployCo routes through McKinsey, Bain, and Capgemini into global enterprise accounts. AWS deploys pods of engineers whose compensation – median FDE base salaries now sit around US$180,000, rising past US$500,000 at the frontier labs – prices them for Fortune-scale engagements. Microsoft’s 6,000 specialists will flow first to its largest global customers.
Now consider the structure of Asia’s economy. SMEs and mid-sized companies account for the overwhelming majority of enterprises across ASEAN and the bulk of employment. These are the manufacturers in Penang, the logistics operators in Jakarta, the family-owned financial services groups in Bangkok, the healthcare networks in Manila and the professional services firms in Singapore. They face the same 95% pilot-failure problem – arguably more acutely but the unit economics of a Silicon Valley FDE pod simply do not reach them.
The giants have validated the category and priced themselves out of most of this region’s economy in the same breath.
And even where they land, there is a day-46 problem
AWS’s engagements run roughly 45 days. The FDE model, by design, embeds, builds, and leaves. AWS itself markets customer self-sufficiency as the goal – a goal I genuinely respect.
But anyone who has led real transformation in this region knows what happens on day 46. The system is live; the capability is not. Model behaviour drifts. Regulations evolve, MAS guidance in Singapore, PDPA obligations, Indonesia’s data-protection regime, sector rules in every market. Staff rotate. The workflow the pod redesigned meets the organisation the pod never had time to change.
Deployment is an event. Transformation is a capability. The ten-billion-dollar wave solves the first. The second still has to be built locally, in the language of the business, over quarters rather than sprints.
Collaboration, not competition – at different altitudes
Here is the narrative I believe Asian business leaders should adopt, and the one we operate by at MASSIVUE.
These new deployment companies are aircraft carriers: enormously capable, globally scaled, and built for open water. But carriers do not dock at every port. They need pilot boats, local harbours, and crews who know the waters. The frontier labs and hyperscalers bring the models, the tooling, and increasingly the embedded engineering pattern. Local and regional partners bring what cannot be shipped from San Francisco or Seattle: regulatory fluency, cultural and linguistic context, mid-market economics, and the patient work of building capability that stays after the engagement ends.
That is not a consolation prize. It is the layer where value is actually realised. Industry analysis has long observed that for every dollar spent on software, roughly six are spent on the services around it. The giants are chasing that services layer globally; the regional share of it will be won by those who can land these technologies inside real Asian businesses and keep them working.
This is why we welcome these launches. AWS has already opened a Partner-Led FDE motion, explicitly building credentialed delivery capability inside consulting partners. Microsoft named its global SI ecosystem as central to scaling Frontier Company. The doors for regional collaboration are open for those ready to walk through them with production-grade delivery standards.
What Asian enterprises should do now
If you lead a mid-sized enterprise in this region, three moves matter this quarter.
First, stop piloting and start deploying but deploy with a capability plan. Ask any implementation partner, global or local, one question: what does my team own on day 46?
Second, insist on regional realities from day one. Data residency, local regulation, multilingual users, and change management across your actual workforce are not “phase two” items. They are the difference between the 95% and the 5%.
Third, choose partners who transfer capability, not dependency. The most honest test of an AI partner is whether they are structured to make themselves progressively less necessary.
At MASSIVUE, we have built our model around exactly this embedded implementation paired with structured capability building through our Academy, so the transformation belongs to the client, not the consultant. The world’s largest technology companies just spent ten billion dollars proving the first half of that thesis. The second half — the last mile, the day after, the local reality — is where Asia’s advantage lies.
The frontier has come to us. Landing it is our job.
Sandeep Joshi is the Founder and Managing Director of MASSIVUE, a Singapore-headquartered AI transformation, consulting and capability-building firm serving enterprises across APAC. MASSIVUE’s Consulting + Academy model has trained over 5,000 professionals and supports AI deployment from strategy through production and beyond.
For the full research and data behind this article, download the MASSIVUE white paper: “The Deployment Decade: Forward-Deployed AI and the APAC Last Mile.“