Why should Intel seize the opportunity of strong stock prices to conduct a seasoned equity offering?

Original title: Intel Should Raise Capital

Original author: Semianalysis

Translation: Peggy, BlockBeats

Editor's Note: Since breaking out in early April, Intel's stock price has been steadily recovering, with two key catalysts emerging in June: first, the market reported that Google placed an AI chip order with Intel, driving its stock price up sharply in a single day; second, Bank of America rarely upgraded Intel directly from "underperform" to "buy," raising the target price from $96 to $135. Behind this rebound, the market is repricing not just Intel's short-term performance, but its strategic position in AI CPUs, advanced process manufacturing, and the US domestic chip supply chain.

INTC Stock Price Trend

Today, Intel's transformation narrative is shifting from "self-rescue" to "re-expansion." With Lip-Bu Tan taking over as CEO, a new board overhaul, and strategic capital from the US government, SoftBank, and Nvidia entering the fray, market expectations for Intel have clearly been repaired. However, this article warns that what truly determines whether Intel can return to the core table of advanced manufacturing is not just customer commitments and stock price rebounds, but whether it has sufficient capital to actually build out its foundry capacity.

The author argues that much of Intel's problems over the past decade stemmed from financial engineering: selling assets, bringing in joint venture partners, using Smart Capital (reducing capital expenditure pressure through joint ventures and asset disposals) to ease cash flow pressure, while also ceding the long-term returns of core assets like wafer fabs.

Now, the most important thing Intel should do is not buy back stock, but conduct equity financing while its stock price is strong. The reason is straightforward: on one hand, the current valuation is already high, and a 4% to 5% equity dilution could raise approximately $250 billion, enough to significantly enhance Intel's ability to build advanced process capacity. On the other hand, the US government, SoftBank, Nvidia, and others entered at prices below the current stock price, so a new issuance does not necessarily "punish" new shareholders; instead, it could increase book value per share and provide these strategic investors with paper gains.

More importantly, alternative financing methods Intel has tried in the past have proven to be costly. Whether it was selling its NAND business, reducing its stake in Mobileye, relinquishing control of Altera, or introducing partners like Apollo and Brookfield through SCIP (Semiconductor Co-Investment Program, exchanging long-term rights to wafer fab earnings for external capital), each essentially traded assets and future earnings for cash. Now, Intel spending $14.2 billion to buy back Apollo's stake in Fab 34 precisely proves that ceding wafer fab economic benefits was not cheap in the first place. Continuing to add debt would strain the balance sheet, further asset sales have limited room, and equity has become the cheapest and cleanest source of funding currently available.

Therefore, the core thesis of this article is: Intel no longer lacks a "comeback story"; what it truly lacks is the capital needed to execute that story. Demand for Agentic CPUs (a new type of CPU for the AI agent era), potential major clients like SpaceX and Tesla, and orders from Nvidia and Google have given Intel a demand foundation it can present to the capital market. For Intel, issuing new shares is not simply dilution but, when the market window opens, using cheap capital to secure the execution rights for advanced process capacity, its foundry business, and the silicon sovereignty narrative. Missing this window could be more costly than the financing itself.

The following is the original text (edited for readability):

We have written a lot about Intel. This company holds a special significance for us; it can almost be considered the starting point of the semiconductor industry. Merely saying we love Intel and recognize its role in the world still falls short. In the past, when Intel's early products faltered, we pointed out the issues very directly; regarding its transformation, we have also been supportive and hopeful. One of our firmest judgments is that the Intel board was one of the biggest factors responsible for Intel's decline, and recently, we have finally seen the changes we have been waiting for.

Franky Yeary stepped down after 17 years on the board. The new board is now composed of people who truly understand the industry, not just financial engineering. The new chairman previously served at Qualcomm, Lip-Bu Tan is the CEO, and the board also includes Steve Sanghi and Stacey Smith from Microchip, and Eric Meurice from ASML. In other words, this board finally understands technology.

However, although Intel's transformation has partially begun, the road to fully revitalizing the company remains long. We believe today that Intel, under the leadership of this new board, should make another major strategic bet: not stock buybacks, but issuing enough new shares to fundamentally repair Intel's financial condition in one fell swoop.

Lip-Bu Tan has pulled Intel back from the brink and raised approximately $20 billion through investments from the US government, SoftBank, Altera, and Nvidia. Intel should not stop halfway; it should continue to capitalize on the current strong stock price. In the past bad years, the company was a net buyer of its own stock; now is the time to issue equity while the stock price is strong. If done correctly, this will make Intel's transformation much easier to succeed.

Note: Lip-Bu Tan is Intel's CEO, appointed in March 2025, and also sits on Intel's board.

Equity dilution now actually rewards investors who have already bet

Look at the prices at which these funds came in. The US government subscribed for up to 433 million shares at $20.47 per share, representing a 9.9% stake at signing; as of the end of Q1, 149 million shares were still in escrow. SoftBank's entry price was $23.00, and Nvidia's was $23.28. All these holders are now in profit.

Therefore, the intuition that financing would punish recent investors is actually misguided. Issuing stock at today's price, which is much higher than these entry prices, would increase book value per share and generate paper gains for the US government, SoftBank, and Nvidia. The nearly 10% sovereign capital anchor is also a significant reason Intel can execute large-scale issuances at a relatively low cost. Intel is one of the very few companies globally that can sell large amounts of stock when market sentiment is hot, while also having the US government as a backstop. As long as this leverage exists, it is worth using.

Intel needs capital to execute its transformation

Based on its performance over the past twelve months, Intel has almost never been as expensive as it is now since the dot-com bubble in 2000. We believe in the company's bright prospects, but to realize them, one of the most critical elements is capital; the current stock price does not fully reflect the real execution risk.

More importantly, even in the best-case scenario where demand for Agentic CPUs recovers, Intel cannot solely afford all the investments required for an upside scenario. We believe it is time for Intel to execute a "reverse buyback": raising equity capital while there is still demand for its stock issuance in the market.

Equity is now the cheapest capital available to Intel

Detractors might say Intel has other ways to finance its fabs. But it has tried them all, and has just signaled to the market that these methods are not effective.

Apollo invested $11.2 billion for a 49% stake in the Fab 34 joint venture; Brookfield structured a financing deal for the Arizona fab project; Silver Lake acquired 51% of Altera at an $8.75 billion enterprise value, bringing Intel approximately $4.3 billion in net cash. Intel also phased out its NAND business to SK hynix and continued selling Mobileye shares. "Smart Capital" (a capital strategy using joint ventures and asset sales to reduce capital expenditure pressure) was once Intel's core narrative.

Then, on March 31, 2026, Intel agreed to buy back Apollo's 49% stake in Fab 34, completing the transaction on April 8 for a total of $14.2 billion, consisting of approximately $7.7 billion in cash and $6.5 billion in bridge loans. Management stated this buyback would be accretive to earnings, and they were right, which is precisely the key point. If buying back the wafer fab equity is accretive, then selling the wafer fab's economic interest to partners in the first place was essentially an expensive form of financing all along. SCIP essentially cedes a portion of the long-term earnings rights from the company's best assets to external partners in exchange for capital that appears cheaper on the surface but is actually more costly. Intel has now proven with its own checkbook that it prefers to own its fabs and bear the associated debt rather than continue ceding their earnings.

So, let's cross out the other options. Doing more SCIP deals is exactly the type of choice management just spent $14.2 billion to reverse. Adding more debt would pile onto the existing $45 billion in debt on the balance sheet; factoring in the Apollo bridge loan, the total debt reaches about $51.5 billion. Major asset sales are also mostly complete; Mobileye and Altera have either been sold or control has been ceded. What remains is equity financing. And at current valuation levels, equity is the cheapest capital Intel has.

With the announcement of the massive Terafab project and the spillover demand from severe N3 shortages, Intel's foundry business is just getting started. To truly seize this unique window, Intel must become a critical supplier for the entire industry during this period of tight supply for advanced process wafers. The funding required for this enormous bet far exceeds what Intel can cover with its current operating cash flow.

A mere 4% to 5% equity dilution could raise approximately $25 billion, enough to turn even the most optimistic supply capacity narrative into reality at this critical juncture.

Agentic CPU demand isn't enough to foot the Terafab bill

Customer commitments from SpaceX, Tesla, and those represented by Terafab are key to solving the 14A capacity issue. The initial goal is to reach 100,000 wafer starts per month (WSPM), with a further expansion to 1 million WSPM. This will be very difficult and carry immense capital pressure. But it must happen because Lip-Bu Tan has publicly told the market: if there are no customers, he will shut down the foundry business. Now the customers are here, so it's time to build.

Beyond the Terafab partners, Intel's order book is also filling up. Nvidia's DGX Rubin NVL8 configuration lists dual Intel Xeon 6 host CPUs; Google has signed a multi-year agreement covering Xeon and custom IPUs; SambaNova has also joined in for inferencing. The wafer volumes behind these orders are not fully disclosed, but the capital market funding a visible order book comes at a much lower cost than funding a transformation story. And Intel finally has orders it can point to for the market to see. Equity financing backed by signed demand has a completely different pricing logic than financing based on a promise.

Due to lower-than-expected CPU demand, Intel has been struggling to delay capital expenditures. But now, it's time to go all-in again, much like in the Gelsinger era. This is a critical moment for silicon sovereignty, and Intel must continue to increase its bets.

This complete multi-phase project at Intel could cost up to $119 billion. While SpaceX will provide initial capital, Intel must also make a meaningful contribution. Even just marginal capital matching implies hundreds of billions of dollars in new funding needs, which were not even on Intel's capital expenditure decision matrix a month ago.

Now is the time to end the decade of financial engineering and issue stock immediately. Because while the capacity ramp is exciting, it will be very expensive. The current window for equity issuance is the widest it has been in some time; if Cerebras can raise $5.55 billion, Intel can raise $25 billion. This argument is only stronger because Intel's approximately $498 billion market cap can easily support a much larger follow-on offering. In our view, this window appears to be fully open. Below is some data from recent comparable issuances.

The trading window is open

In other words, Intel's real problem now is no longer "does it have a story," but "does it have enough capital to turn the story into capacity." With strategic capital from the US government, SoftBank, and Nvidia already in place, and the advanced process supply in a tight window, equity financing is no longer just a defensive move that dilutes shareholders, but could become an offensive choice to restart Intel's foundry ambitions and bet on silicon sovereignty. For Intel, missing this financing window might be more expensive than the issuance itself.

Original Link