Rise of the OREO Company in the AI Epoch
Onshore Resident Executing Offshore
In the frantic global gold rush of artificial intelligence, where compute power is the new oil and regulatory scrutiny is the new border wall, a stealthy new corporate species is quietly dominating the ecosystem.
These aren’t your grandfather’s multinationals chasing cheap labor for call centers.
They’re something far more deliberate and strategic and what I call: OREO Companies—Onshore Resident, Executing Offshore.
Born from the unique pressures of the AI age, they keep their public heart, brand, funding, and regulatory face firmly planted in the United States (or other Western markets), while deliberately planning the most expensive, compute-heavy, and talent-intensive essential work overseas. Not as a component, as a constitutive of this new business model.
It is more than incidental or accidental outsourcing.
It’s a consciously engineered, bi-operational business model—part domestic, part foreign—designed from day one to exploit the brutal economics of frontier AI while dodging the political, legal, and cost landmines of operating entirely onshore or entirely offshore.
Lawyers would see immediately, it can give any operation a regulatory “call” / Toggle to Jurisdiction A or Jurisdiction B as the circumstance would dictate.
Defining OREO: From 2000s BPO to Intentional Bi-Operational Powerhouses
Flash back to the early 2000s.
The big trend was Business Process Outsourcing (BPO).
Western giants like GE, Citibank, and American Express shipped repetitive, low-value back-office functions—call centers, payroll processing, basic data entry—to India and the Philippines.
I know because I was there. I invested in these companies. I “papered” the deals.
The goal was simple cost arbitrage: cut labor expenses by 60-70% while the “real” company, the strategy, the IP, and the customer-facing core stayed firmly onshore.
It was additive outsourcing, not structural levering and corporate reinvention.
OREO Companies represent the next evolutionary leap—and a fundamentally different philosophy.
These are intentionally planned bi-operational entities. From inception (or through deliberate restructuring), they split their corporate DNA and operations across borders [literally and also obliquely]:
Onshore Resident: U.S. headquarters (or equivalent), American CEO and board visibility, primary fundraising from Silicon Valley VCs, U.S. stock listing or regulatory filings, domestic sales and marketing teams, and strict compliance with American data-security and national-security rules. Also present federal tax and other benefit from siting in the USA. This is the “face” the public, investors, and regulators see.
Offshore Execution: But the heavy machinery of AI—massive GPU clusters for training and inference, core model development, data labeling at scale, low-level engineering, and much of R&D—happens in cost-advantaged, high-talent jurisdictions like India, Southeast Asia, or select Middle Eastern hubs. This is where the real work really gets done, 24/7, at a fraction of the cost.
As an intentional builder leaning on India-sited teams, I have also not just seen but done this first-hand with jurisphere.ai voxlex.ai and some others.
The offshore piece isn’t an afterthought or a cost-cutting sideshow.
It is core to the business model, baked into the org chart and cap table from the beginning.
OREO isn’t “let’s move the back office to Bangalore.”
It’s “let’s build the entire AI engine in Hyderabad while keeping the American brand, capital, and regulatory shield intact.”
The result is a hybrid organism optimized for the AI era’s unique constraints: exploding compute demand, GPU shortages, stratospheric U.S. salaries, energy bottlenecks, and tightening export controls and data-sovereignty laws.
Why the AI Era Supercharged This Model
The OREO playbook isn’t a loophole.
It’s the logical, almost inevitable corporate (law) evolution to a world where building the most powerful technology in history has never been more expensive or more politically fraught.
It takes advantages of cost arbitrages, and demographics, many many more number of educated young in jurisdictions not America.
In the 2000s, BPO helped companies save on overhead. In the 2020s, OREO Companies are using deliberate bi-operational architecture to win the AI race outright.
3 Examples : Cerebrus Tik Tok Blaize
Example 1: Cerebras Systems — U.S. Chip Pioneer Goes Deep into India for Compute
Cerebras Systems, the California-based AI supercomputer maker known for its massive wafer-scale engines, exemplifies the OREO model.
Headquartered in Sunnyvale with a U.S.-centric identity, the company has offices in Bangalore and Hyderabad—and it’s executing major compute projects offshore.
In February 2026, Cerebras partnered with UAE’s G42, India’s Centre for Development of Advanced Computing (C-DAC), and Mohamed Bin Zayed University of Artificial Intelligence to deploy an 8-exaflop AI supercomputer in India. The system, powered by Cerebras’ WSE-3 chips, will support sovereign AI applications for Indian academia, government, and SMEs while complying with local data residency rules.
This isn’t just market expansion—it’s strategic “bi-shoring” — ab initio offshoring of the most expensive part of AI (massive-scale training and inference) to leverage India’s infrastructure push, talent, and lower operational costs.
Cerebras remains “resident” onshore for U.S. customers, funding, and IP control while executing where economics favor it.
Example 2: TikTok (TikTok USDS Joint Venture) — Formerly Chinese, Now U.S. Resident with Offshore Roots
TikTok offers a regulatory-driven variant of the OREO playbook.
Originally a ByteDance creation with deep Chinese ties, the app faced existential U.S. threats over data security and foreign influence.
In response, ByteDance restructured U.S. operations in early 2026 into the TikTok USDS Joint Venture LLC—a majority American-owned entity (80%+ controlled by U.S. investors including Oracle, Silver Lake, and MGX; ByteDance retains just 19.9%).
This new onshore-resident structure complies with executive orders on national security, with U.S.-based data handling via Oracle, localized content moderation, and algorithm safeguards.
Yet core execution—algorithm development, much of the engineering, and backend tech—remains tied to ByteDance’s offshore expertise and resources.
The U.S. entity licenses technology while maintaining separation. It’s “formerly Chinese” in the public U.S. facing sense, allowing market dominance (hundreds of millions of American users) without full onshore costs or risks.
This hybrid has become a template for navigating bans while preserving efficiency.
Example 3: Blaize — Nasdaq-Listed AI Hardware Innovator with Half Its Workforce Executing in India
A third compelling case is Blaize, a California-based, Nasdaq-listed startup founded in 2011 that builds programmable AI processors and software for edge devices and data centers.
Focused on energy-efficient, real-world AI applications (think smart traffic, healthcare analytics, defense, and agriculture), Blaize has raised over $330 million from investors like Temasek, Mercedes-Benz, and Samsung.
Crucially, roughly half of its ~300 employees are based in India, with a significant R&D operation in Telangana (Hyderabad) established over a decade ago.
The company partners locally (e.g., with Yotta for data center software) and recently signed an MoU with the Telangana government to deepen AI and semiconductor collaboration.
While headquartered and listed onshore in the U.S. for credibility and capital access, Blaize executes core hardware design, software development, and AI optimization offshore—tapping India’s talent pool and cost advantages to accelerate deployment of its chips in global markets.
This model lets Blaize compete on price and speed in the cutthroat AI silicon space without ballooning U.S. overhead.
If you’re founding, investing in, or scaling an AI company today, the question is no longer “Should we offshore some tasks?” It’s “Are we fully onshore, fully offshore… or strategically, intentionally, unapologetically OREO?” (Bi-Shore)
The winners are already choosing the third path and I predict it is a business model who’s time has arrived.