Intel's $8.9 Billion Moment

The Trump administration's equity stake in Intel marks a turning point in U.S. industrial policy. Unlike past subsidies, this is ownership: 433.3 million shares purchased at $20.47 each, giving Washington a 9.9% stake in America's only leading-edge chipmaker. Importantly, this wasn't fresh cash so much as a conversion of $5.7 billion in CHIPS Act grants and $3.2 billion in Secure Enclave funding layered on top of $2.2 billion already delivered. The symbolism is powerful: permanent capital, not clawbacks and conditional grants.¹

The deal is deliberately hands-off. Washington gets no board seat and no day-to-day influence, and its warrant for another 5% stake only activates if Intel loses majority control of its foundry.² It is an ownership signal, not a management takeover. Taxpayers get equity at a discount, Intel gets a balance-sheet backstop without new debt, and the entire industry gets a message: the U.S. government is not just subsidizing semiconductors, it is betting on them.

From a shareholder perspective, the structure is designed to balance political symbolism with financial prudence. For Intel, the benefit is equally clear: a reinforced balance sheet, the elimination of clawback provisions that could have constrained capital planning and a level of confidence that simply writing another grant check could never deliver. Symbolically, this marks a break from subsidy-style industrial policy toward explicit equity participation, with the state standing shoulder to shoulder with private shareholders in assuming risk and return.

Intel's Position: Assets and Struggles

Intel still controls some of the most valuable industrial assets in America: fabs³ in Arizona, Ohio, New Mexico and Oregon, many of which are being tooled for the 18A and planned 14A nodes. These facilities anchor more than $100 billion of recent investment.⁴ Yet the same fabs that prove Intel's strategic importance also drag on its financials. Capital expenditures are running at $18 billion annually, while the foundry division posted a $5.5 billion loss in the first half of 2025.⁵ They are crown jewels—but crown jewels that bleed cash.

Financially, Intel has lost $45 billion in free cash flow in just more than three years, eroding a market capitalization now down 64% from its 2020 peak. Gross margins, once reliably 55%–60%, hover in the mid-30s. Execution has faltered too: of the vaunted "five nodes⁶ in four years" pledge, only Intel 4 counts as a clear success. External foundry revenue in 2025 was just $53 million—symbolic in the face of multibillion-dollar losses. The exit of Pat Gelsinger, replaced by Lip-Bu Tan, underlined how fragile confidence had become. As Barron's observed, "money alone won't solve the company's problems, which are rooted in manufacturing."⁷

Analysts highlight the central bind: Intel needs both scale and credibility to fund its roadmap, but its repeated missteps have eroded confidence. Without a stronger central processing unit (CPU) roadmap and market-share recovery, Intel lacks the scale to justify the very investments it is making.⁸ That, more than any headline, explains why the government stepped in. It wasn't saving a company that was "just fine"; it was preventing a national champion from failing through sheer lack of time and customers.

Ways to Win: Why the Stake Matters⁹

The Intel stake works because it multiplies paths to success. The first is simple downside protection: permanent capital that reduces the chance of outright failure. With Washington on the cap table,¹⁰ the probability of insolvency collapses.

The second path is confidence-driven. If the government refuses to let Intel fail, then hyperscalers,¹¹ defense primes and fabless¹² U.S. chipmakers can rationalize new wafer orders. Notably, the government explicitly hates failures. One anchor contract at 14A¹³ could spark a hedging cascade; others would follow to preserve optionality. The dynamic here is less about present competitiveness than about reducing counterparty risk. When Washington backs a company, customers feel safer making long-dated bets.

The third path is policy leverage. Intel becomes a bargaining chip in broader negotiations: access to U.S. markets, favorable tariffs or accelerated expensing might tilt toward customers who support onshore Intel production. These fiscal levers matter more than subsidies because they change the total cost of ownership. Washington does not have to tell customers, "Buy Intel." It simply has to make it financially smarter to do so.

Finally, there's the breakup option. Intel has already carved out its Foundry unit. The government's warrant is explicitly tied to its control. A clean separation with Foundry as a national champion, and CPUs and datacenter assets free to find their own partners, could unlock value that the integrated model cannot. History offers precedent: Standard Oil's breakup revealed more value in the parts than the whole. Each path, be it downside protection, customer orders, fiscal favoritism or breakup, creates more ways to win than to lose.

What Comes after Intel?

The bigger lesson is that the Intel stake is not an anomaly. It signals a new playbook: whenever a company sits at a chokepoint, lacks the financial scale to compete and matters for both security and economics, do nothing is not the base case. Washington now has equity, procurement and tariff levers it is willing to use. That changes the distribution of likely outcomes across the industry.

Continuing this thread, in an August 26, 2025, interview on CNBC, Commerce Secretary Howard Lutnick revealed that the administration is "actively discussing" stakes in major defense and tech contractors, including Lockheed Martin, Boeing and Palantir.¹⁴

Other potential candidates stand out. Each has reasons for and against intervention. Our aim here is not to predict or call out the next action of the U.S. government, but rather to present an illustrative mental model that presents a way of trying to think of the sorts of things the U.S. government may think of in the future:¹⁵

• GlobalFoundries (GF)

Why it might happen: GF is critical for autos, defense and radio frequency chips, ensuring resilience in supply chains most exposed to shocks. It also represents a U.S.-domiciled alternative to Taiwan Semiconductor Manufacturing Co. (TSMC) for less-advanced but strategically vital nodes.

Why it might not: The company already has stable niches and profitability at the lagging edge;¹⁶ the government may view it as less urgent since it does not compete directly in the bleeding edge where national security concerns are most acute.

• Micron

Why it might happen: Micron anchors U.S. memory and high-bandwidth memory (HBM) for AI accelerators.¹⁷ HBM, in particular, is a national security interest, given its role in AI compute.

Why it might not: Micron is relatively healthy financially compared to Intel, and while capital hungry, it has demonstrated execution strength. Interventions would likely be more targeted (loan guarantees or procurement) rather than equity.

• Wolfspeed

Why it might happen: Wolfspeed is central to silicon carbide (SiC) semiconductors, which underpin electric vehicle (EV) drivetrains, power electronics and defense systems. Its balance sheet has been stretched by fab delays, making it a candidate for risk-sharing capital.

Why it might not: The SiC market, while strategic, is smaller in scale than logic or memory.¹⁸ Washington might conclude that demand pull-through, rather than ownership, is enough to support Wolfspeed.

• Amkor/Advanced Packaging

Why it might happen: Advanced packaging is the hidden chokepoint of the AI era, and U.S. capacity is limited. A national packaging utility could be a natural candidate for public–private capital.

Why it might not: Politically, "packaging" lacks the resonance of "chips," and subsidies may flow more naturally than equity. A shared facility funded through procurement may be viewed as sufficient.

• Coherent (Lasers/Optics)

Why it might happen: Coherent plays a role in lasers, optics and photonics critical for interconnect and defense. Fragile financial performance could create pressure for support.

Why it might not: Compared to memory or foundry, photonics is a layer removed from the AI/logic core. Washington may prioritize more direct chokepoints before turning to optics.

Each is imperfect, but that is the point: Intel, too, was imperfect when Washington acted. The calculus is not about the U.S. government backing winners, but about reducing the chance of strategically vital failures. Investors should see this as an emerging policy pattern: chokepoint companies with weak balance sheets now have more ways to win than the market currently prices.

Closing Thought

Washington's stake in Intel is not about "picking winners" in the caricatured sense. It is about reducing failure risk in industries too strategic to leave exposed. Equity, not subsidy, may be the new currency of industrial policy. The question is no longer whether government will step in, but where and how—and which chokepoints it chooses to reinforce. Intel shows that the era of subsidies-only industrial policy is over. What comes next may be smaller, subtler and industry-specific, but the direction is clear: if the U.S. government sees a company as indispensable, "just fine on its own" may never be the starting assumption again.

For those considering how WisdomTree places Intel into either the geopolitical or artificial intelligence contexts:

• WisdomTree GeoAlpha Opportunities Fund (GEOA): GEOA is designed to track the total return performance of, before fees and expenses, the WisdomTree GeoAlpha Opportunities Index. This strategy generates exposure to companies primarily benefiting from geopolitical and global policy shifts. It focuses on the following categories of geopolitical and global policy shifts: geopolitical events, fiscal and monetary policy shifts, innovations in technology and shifting consumer preferences. As of August 22, 2025, Intel was a 2.78% weight in GEOA. Holdings are subject to change.
• WisdomTree Artificial Intelligence & Innovation Fund (WTAI): WTAI is designed to track the total return performance of, before fees and expenses, the WisdomTree Artificial Intelligence & Innovation Index. As of August 22, 2025, Intel was a 1.55% weight in WTAI. Holdings are subject to change.

¹ Source: Intel Corporation, "Intel and Trump Administration reach historic agreement to accelerate American technology and manufacturing leadership" [press release], Business Wire, 8/22/25.

² Source: Intel Corporation, 2025.

³ Fabs refers to facilities that directly make physical semiconductors.

⁴ Source: Intel Corporation, 2025.

⁵ Source: Intel Corporation. "Form 10-Q for the quarterly period ended June 29, 2025," U.S. Securities and Exchange Commission, 2025.

⁶ Nodes refer to "process nodes" that are usually defined in relation to the brand label for a generation of process technology—capturing improvements in performance, power efficiency and density, but not tied to a literal nanometer metric.

⁷ Source: Barron's, "Intel's $10 billion lifeline: Why Washington's stake may not be enough," Dow Jones & Company, 8/21/25.

⁸ Source: J. Moore, M. Wayne, S. Brett, E. Tulchinsky and N. Kozhukhov, "Intel Corporation (INTC): What's next for Intel?", Morgan Stanley Research, 8/21/25.

⁹ Sources for the "Ways to Win" section: Intel Corporation, "Intel and Trump Administration reach historic agreement to accelerate American technology and manufacturing leadership" [press release], Business Wire, 8/22/25; Intel Corporation, "Form 10-Q for the quarterly period ended June 29, 2025," U.S. Securities and Exchange Commission, 2025; J. Moore, M. Wayne, S. Brett, E. Tulchinsky and N. Kozhukhov, "Intel Corporation (INTC): What's next for Intel?", Morgan Stanley Research, 8/21/25.

¹⁰ The cap table references the list of owners, usually equity stakes, of a company.

¹¹ Hyperscalers refers to the largest providers of public cloud infrastructure, typically Amazon Web Services, Microsoft Azure, Google Cloud and, more recently, Oracle.

¹² A fabless chipmaker would design chips and send those designs to a fabrication facility.

¹³ 14A refers to a process node that Intel is currently working on and would represent a significant technological leap forward.

¹⁴ Source: "U.S. could seek piece of defense companies, Commerce chief says" [news article], Reuters, 8/26/25.

¹⁵ Sources for following bullet points: each specified firm's 10-K and 10-Q filings.

¹⁶ Lagging edge refers to less advanced chips that may be important and in demand but that do not require the most advanced technologies to make.

¹⁷ HBM chips are essential to be used alongside AI accelerator graphics processing units, such as Nvidia's Blackwell GPU.

¹⁸ Logic refers to chips that process information, like CPUs and GPUs. Memory refers to chips that store the results of computations done by logic chips.