The Philippines Built a $40 Billion Industry It Doesn't Own

The Philippines captured income without capturing control.

IT-BPM is a $40 billion industry. 1.9 million jobs. A genuine middle class that didn't exist thirty years ago. And roughly $12 billion of that $40 billion exits the country annually — through IP licensing fees, management service charges, and dividend repatriation flowing back to parent companies in Charlotte, Paris, and Zurich. Structures the Philippine government helped design. Structures that are individually legal, collectively significant, and almost never discussed at the industry galas.

This is the structural condition that makes the AI transition different here than everywhere else. Not just a technology shift. An exposure event.

I talked through this with Bernd Rennebeck, who's been building in the Philippines for 16 years and has been publishing some of the most data-dense analysis of this specific problem at blog.brennebeck.com. His framing cuts to it: "The Philippines perfected hospitality as industrial policy. It became very good at welcoming other people's companies and very poor at ensuring that a meaningful share of the strategic upside would remain Filipino when the arrangement changed."

That arrangement is now changing.

Capability is not control

The distinction that matters is between capability and control.

The Philippines has formidable capability. Deep domain expertise in customer operations, compliance workflows, healthcare support, financial reconciliation, multilingual service environments. Millions of skilled, disciplined, English-fluent workers built over two and a half decades. These are real assets.

Capability without ownership means that when value starts moving away from labor — toward software, workflow automation, and retained intellectual property — you're not at the table where decisions get made.

India is the contrast that makes this concrete. TCS, Infosys, Wipro are Indian firms. When they automate a customer-service operation, the productivity gain stays in India. It gets reinvested in R&D, in new service lines, in the next-generation workforce. As of Q3 FY2026, TCS reports more than 217,000 employees with advanced AI skills and annualized AI services revenue at $1.8 billion. They're not defending a headcount model. They're converting one. The gains from that conversion flow to Indian shareholders and Indian institutions.

The Philippines has no firm in this position. The $40 billion sector is almost entirely controlled by Accenture, Teleperformance, Concentrix, and their global peers. When those firms automate Philippine operations, the employment adjustment lands here. The upside accrues elsewhere.

Bernd put it plainly: "For India, it's existential for their business and their country. Here, we are a byproduct of it."

The math

The Philippines' flagship AI reskilling program for IT-BPM — Project UNLAD — carries a four-year budget of roughly $13 million. For a 1.9-million-strong workforce, that's $1.70 per worker per year.

India's IndiaAI Mission: $1.25 billion over five years. 38,000 GPUs of publicly backed compute. 1.85 million workers enrolled in a digital reskilling program with over 337,000 course completions.

Thailand — comparable GDP, smaller population — committed approximately $770 million to government AI programs over two fiscal years. That's roughly $10.80 per capita versus $0.38 in the Philippines. A gap of nearly 28-fold between two countries at roughly the same development stage.

The government isn't spending nothing. The National AI Strategy was approved in 2025. There are roadmaps, pilot programs. The gap is between movement and scale, and the scale of what's coming isn't being matched.

Bernd's scenario modeling puts a probability on the downside: crisis-level outcomes carry roughly a one-in-six chance; the broader recession-and-displacement scenario approaches one-in-three. The ILO has found that 89% of Philippine BPO workers hold roles with high automation exposure. The blast radius extends beyond the sector — IT-BPM accounts for 44% of all Philippine office space. The towers of BGC and Ortigas exist because of that bet. Everything downstream — the malls, the condominiums, the informal service economy — runs on the same wager.

AI doesn't create this vulnerability. It exposes it.

The founding class sitting inside the problem

The people who get 10x more valuable in the AI era are not the most specialized. They're high-agency individuals with wide knowledge breadth, core curiosity, and the ability to learn fast. Generalists who can direct AI across domains. The Philippines already has hundreds of thousands of these people — they're just not being seen as the founders they could be.

They're the senior operations managers, the compliance leads, the training directors, the finance-and-accounting specialists inside BPO delivery centers. They've spent a decade absorbing domain knowledge that is genuinely hard to build. That knowledge is the most valuable input to the next wave of Philippine software — if someone converts it into a product rather than watching it become a cost center.

Bernd frames it this way: "The mid-career workers in the upper layers of BPO operations are perhaps the most important constituency of all. They are not liabilities waiting to be automated away. They are the founding class of the next wave — if they are given the tools, the capital, and the cultural permission to see themselves that way."

Before, the 20% was strategy — the hard part. The 80% was execution. AI is eating the execution. Which means you're always in strategy mode now, which is exhausting, but it also means the barrier between "I have an idea" and "I have a product" has collapsed.

The economics follow the timeline compression. A $15,000–25,000 angel check from a diaspora connection buys several months of runway for a two-person team in Manila. You don't need institutional venture capital to get to something real. You need domain knowledge, curiosity, and the willingness to start.

"Not everything has to be venture capital," Bernd said. "If you're curious and you're hungry, those are really the only prerequisites."

The ownership gap

There's one counterexample worth naming: GCash. Built through Globe's technology arm. Genuinely Filipino-owned. Real scale, 81 million users. It shows exactly what domestic capital can produce when it makes a patient, long-horizon bet on capability rather than tenancy.

One GCash in twenty years, from a $40 billion sector full of Filipino talent, is not a success story. It's the exception that makes the absence of everything else harder to excuse. Filipino conglomerates — Ayala, SM, Aboitiz, JG Summit — had the capital, the talent networks, and the institutional relationships to build domestic technology champions over the past two decades. The returns from real estate, retail, banking, and utilities were more immediate. That's where the money went. The sector got payroll and rent. It got far less compounding control.

The window to correct this isn't closed. But it's not permanent either.

The soft landing isn't defending legacy headcount. It's using the current BPO base as a bridge — converting workers with irreplaceable domain knowledge into builders, before the bridge decays.

What Bernd is building

Bernd is looking to run AI workshops out of Vantis, a coworking space on the 49th floor of PBCom Tower in Makati — show-and-tell sessions for entrepreneurial-minded people, demonstrating what these tools actually do in practice. His argument is that the bottleneck isn't the technology. It's AI fluency spreading fast enough to reach the people who most need it. If you're looking for a place to work, book a seat at vantis.ph.

He's not a pessimist about the Philippines. He's a realist who chose to stay and is betting on the outcome going the right way. His daughter will live in whatever future this country builds. That's not rhetorical framing. It's why he's spending his time on this now, while the window is still open.

The conversations that should be happening aren't happening at the scale they should be. This one is a start.

Bernd Rennebeck writes The Philippine AI Transition at blog.brennebeck.com — five pieces, heavily footnoted, drawing on IMF working papers, AMRO scenario modeling, and primary regulatory documents.

Listen to our full conversation on the Oblique Podcast.