Intel, a symbol for decades of the U.S. industrial muscle, remains at the heart of the debate. After years of misguided decisions and loss of leadership, several former executives who, it is claimed, participated in the era that led to the decline, have resurfaced with a proposal as ambitious as it is controversial: take Intel private, split it into five parts, and relaunch it from scratch. The storyline, previewed in statements reported by Fortune and amplified by the tech ecosystem, comes at a unique moment: Intel is starting to regain ground with public support (less than 10% of U.S. government stake) and the entry of NVIDIA with around 5%. Is this a rebound in optimism… or a move to dismantle the company’s historic vertical integration?
The proposal lays out what many have debated privately: separating manufacturing from chip design, selling non-core assets such as Mobileye (estimated at around $15 billion), and divesting from venture capital arm. The case is adorned with parallels: the “conglomerate discount” allegedly penalizes Intel for its diversified businesses; the example of General Electric, which gained traction after spinning off; and, as a historical mirror, the AT&T breakup in the 1980s. On paper, the plan aims to be completed by 2028, with units ready to be sold or go public, and, in passing, taxpayers and private shareholders celebrating capital gains.
The issue — and the reason for controversy — is that self-criticism is absent from the proposal: the advocates are part of the generation that pushed for resurrecting an obsolete and rigid vertical integration model, which fell behind behind TSMC and the rise of the fabless trend. Practically, Intel missed the boat, saw internal morale dwindle, and suffered talent drain toward more attractive competitors during the AI boom. Now, these ex-executives offer themselves as surgeons for a patient that, paradoxically, has just received infusion of public and private capital — including $5 billion from NVIDIA, viewed by markets as a vote of confidence in Western manufacturing and the foundry’s roadmap.
The five-move plan
1) Going private (temporary privatization).
The thesis: quarterly results and market pressure prevent deep reforms. In private hands, they say, corporate “surgery” becomes feasible.
2) Split the “heart” into two: manufacturing on one side, design on the other.
The goal is for foundry to compete head-to-head with TSMC, while the design unit battles in CPU/SoC markets without fab constraints. Essentially, this means giving up vertical integration that made Intel iconic.
3) Selling non-essential assets.
Mobileye, valued at around $15 billion, would be spun off; the same applies to the venture capital arm holding stakes in private companies.
4) Revaluation through spin-off.
The “conglomerate discount” is invoked to argue that each unit would be worth more separately. By estimation, proponents mention around $100 billion for PC and another $100 billion for servers and data centers. Simple calculation, uncertain impact.
5) A consortium of Big Tech and the government to buy out public capital.
The most sensitive part: a consortium led by the U.S. government with heavyweights like Microsoft, Apple, Amazon, Google, Qualcomm, and Broadcom would acquire all listed shares and commit up to $100 billion over ten years to rebuild the foundry. NVIDIA already paved the way with its ~5% stake, signaling that “the market would applaud.”
Why now? Conflicting narratives
The proponents’ narrative.
- Intel cannot implement essential reforms under the quarterly performance pressure.
- Vertical integration is untenable as currently structured; separation is necessary.
- The spin-off removes the “conglomerate discount”, unlocks value, and directs capital toward where it’s needed most — the factory.
- With public-private support, the U.S. foundry can stand up to TSMC within less than 10 years.
The critics’ narrative.
- The proposal arrives just as Intel is rebounding with fresh support; it feels more like an opportunity window than firm conviction.
- “Breaking up” could fragment R&D, double costs, and erode synergies when design and process need to be coordinated, especially during a critical node transition.
- Relying on the same management that failed, the plan lacks credibility without self-criticism or an independent team.
- The AT&T parallel feels forced: that was a regulatory breakup of telecom monopolies; semiconductors operate in a different league, with global supply chains and colossal capex.
Valuations and figures: quick math versus manufacturing reality
The circulating estimates — around $100 billion for PCs and another $100 billion for servers — are approximations that overlook product cycles, margins, and the historical load of transitioning nodes. The foundry remains the capital sink: building and maintaining competitive fabs requires decades of know-how and successive investments. This explains the proponents’ push for injecting up to $100 billion over ten years. But money isn’t everything: talent, supplier ecosystem, yield per wafer, and customer pipeline are real bottlenecks.
Supporters of privatization cite GE as a success story and the “conglomerate discount”. Critics point to the complexity of separating design and process without losing speed during a market — AI-driven — that’s moving fast. And regulation adds an extra layer: a Big Tech consortium buying all of Intel’s shares would face a high-level antitrust review in the U.S. and elsewhere, not a trivial hurdle.
Undisputed fact: the context has changed
The proposal emerges as Intel already has capital and allies lined up. The U.S. government backs it with resources (less than 10% stake) as part of a political push to relocalize critical manufacturing; NVIDIA has committed ~$5 billion; and, parallelly, the industrial discourse in Washington is clear: the U.S. aims to reduce reliance on Taiwan for advanced manufacturing. In this climate, Intel’s foundry — whether it’s doing well or not — is not just a private gamble but a piece of industrial policy. Any attempt to split it would encounter state interests.
Clear risks and potential benefits
Risks
- Fragmentation of capabilities: separating process and design during a node transition could slow down the roadmap.
- Duplicated costs and complex governance: two “parent” companies where there was one demand precise coordination.
- Impact on talent: another massive reorganization could accelerate departures.
- Antitrust regulation: a Big Tech consortium buying all stock would face extraordinary scrutiny.
Potential benefits
- Focus: each unit optimizes its P&L and strategy without cross-commitments.
- Client attraction for the foundry: with design outside the perimeter, fabless competitors might trust the factory more.
- Removal of the “discount”: if the market penalizes conglomerates, spin-offs could revalue assets.
What does industry think? Mixed signals
The NVIDIA entry is seen as a boost to Western manufacturing capacity — and a way to diversify risks, not necessarily an endorsement to break up Intel. The public mobilization (less than 10%) aims to absorb the capital expenditure of the foundry and speed up critical nodes. Meanwhile, large data center and AI clients want second sourcing beyond TSMC. The timing of the proposal, therefore, smells of opportunism: some analysts see it as an attempt to take advantage of a tailwind to push through a reconfiguration that was unthinkable just two years ago.
Key unanswered questions
- Governance: who would lead the separate foundry? What anchor contracts and purchase commitments would ensure the ramp-up?
- Design: how do we guarantee technical coordination during a transition requiring co-evolution of design and fab?
- Employment: what happens to workforces in a multiple spin-off?
- Regulatory risks: can a Big Tech consortium buying all Intel shares pass antitrust checks in the U.S. and other jurisdictions?
The underlying irony: who proposes the “fix”?
As the original text emphasizes: the same profiles that resurrected an obsolete vertical integration, eroded morale, and enabled talent drain are now asking to break the company apart. The argument could be twofold: they know the “crack” — “we know where the fault line is” — or they lack the credibility required for such surgical intervention. In either case, the debate is on. Particularly when new capital (from NVIDIA, the government) appears and AI’s cycle enhances the strategic value of manufacturing.
What can the industry expect by 2028?
The 2028 deadline is seen as a horizon for executing the breakup and listing the resulting units. Yet, that timeline depends on several conditions: secured funding (up to $100 billion over the decade), committed customers, operational node maturity at the foundry, and favorable regulation. Not trivial. Meanwhile, Intel will continue operating, producing, and competing in PCs, data centers, and AI, with Mobileye as a monetizable asset should the plan succeed, and a foundry that the market will demand to show growing yield per wafer.
Conclusion: between urgency and strategic interest
The idea of dividing Intel into five parts is ambitious, controversial, and profoundly impactful. It advocates for focus, value unlocking, and industrial rescue; it omits transition costs, fragmentation risks, and credibility concerns of its supporters. Moreover, it arrives at a turning point: Intel receives backing and re-enters the conversation with AI and industrial policy as catalysts. If surgical intervention is the path, the patient needs more than slogans: independent governance, verifiable investment commitments, and transparency about each actor’s incentives. Otherwise, it risks repeating history: big promises, few results.
Frequently Asked Questions
What does “dividing Intel into five parts” entail according to the proposal?
It involves taking Intel private, separating manufacturing (foundry) from chip design, selling Mobileye (estimated value ~$15 billion), and divesting the venture capital arm. The fifth component would be a consortium led by the U.S. government and major tech firms to buy all listed shares and fund the foundry with up to $100 billion over ten years.
Why propose this now when Intel has received public and private support?
Proponents argue that quarterly pressures hinder reforms and that the conglomerate suffers from a valuation discount. Critics counter that the timing coincides with capital inflows (less than 10% from the government, ~5% from NVIDIA, and $5 billion), and the plan could serve the interests of those involved in the company’s decline.
What are the potential benefits and risks of splitting design and manufacturing?
Potential benefits: focus, removal of the “conglomerate discount”, attracting fabless clients to the foundry.
Risks: RD fragmentation, cost duplication, loss of synergies, regulatory challenges with a Big Tech consortium, and internal fatigue during critical node transitions.
Is the comparison to AT&T or General Electric realistic?
They’re useful references but not directly comparable: AT&T was a regulatory divestment of telecom monopolies; GE operated diversified businesses. In semiconductors, advanced foundry demands massive capex, specialized talent, and accumulated know-how. Success isn’t guaranteed and hinges on clients, regulation, and flawless execution.