Intel (INTC) Stock Falls: A ~$23 Billion Raise, a SoftBank Headline, and the Turnaround Bet

Why Intel stock dropped — and the bigger bet underneath it

On Tuesday, August 18, 2026, Intel (NASDAQ: INTC) fell about 6% to roughly $97, from a $103.49 prior close [1]. The clearest driver was not an Intel headline but a broad, same-day pullback across chip stocks — Micron dropped about 5.3% and Applied Materials about 5.2% on the day, as investors took profits in a group that has run hard all year [7]. Sitting behind that were two pieces of Intel-specific context that had been building for several days: the company had priced a large stock offering — ultimately about $23 billion — that dilutes existing holders, and a widely shared "SoftBank has bet its portfolio on Intel" statistic turned out to be less than it looked. The bigger picture is the stock itself: Intel has risen roughly 296% over the past year on a turnaround thesis, so a down day is as much about how much optimism is already priced in as about any single new development [3].

Intel is not a pure-play anything: it is an integrated device manufacturer that both designs and builds chips. Its large Products business (PC and data-center processors) is solidly profitable, and it is using those profits — plus outside capital — to fund an expensive push into foundry, contract-manufacturing chips for other companies around its new 18A process, in competition with TSMC and Samsung. That build-out is why the headline financials look strange: Intel reported a trailing GAAP net loss of about $11.3 billion, yet its underlying business is far healthier than that number suggests — it was operating-profitable in the latest quarter and generated roughly $8.1 billion of operating cash flow in the first half of 2026. The net loss is dominated by a non-cash accounting charge, explained below, rather than by operating losses [3][8].

The ~$23 billion raise: dilution, and a vote of demand

The concrete corporate event behind the recent volatility is Intel's equity sale. On August 10 the company priced 210.5 million shares at $95 — below the market price — upsizing the deal to about $20 billion from an initially planned $15 billion. Demand was strong enough that underwriters fully exercised their option for roughly 31.6 million more shares, bringing the total to about 242.1 million shares and ~$23 billion of gross proceeds (net proceeds near $22.6 billion) [2]. Intel's disclosed use of proceeds is deliberately broad: general corporate purposes, including capital expenditures and working capital. Management has separately guided 2026 capital spending above $20 billion, which supports both its products and its foundry capacity — but that capex plan is not the same thing as an earmark for this specific offering [2][5].

The market read the deal two ways, and both are in the tape. Issuing 240-plus million new shares increases shares outstanding by roughly 4.8% against the pre-offering count, diluting existing holders, and pricing below the market weighed on the stock. At the same time, the originally marketed $20 billion deal drew more than $100 billion of institutional demand — roughly five times oversubscribed — and Bank of America's Vivek Arya, reiterating a Buy while trimming his target to $145 from $160, framed the raise as evidence of confidence in Intel's foundry ambitions rather than a defensive scramble [2][6]. Strong demand at $95 shows investors were willing to fund the build-out on those terms; it does not, by itself, prove the turnaround will work. Intel is asking shareholders to finance a multi-year effort before the foundry profits arrive.

The government stake — a backer, and the reason for that "loss"

One of the least-understood parts of Intel's story is that, in its 2026 proxy, the U.S. government is listed as its largest disclosed beneficial owner — 433.3 million shares, about 8.4% (as of March 20, assuming full release of escrowed shares), ahead of Vanguard and BlackRock; after the August issuance, that stake is closer to 8.2% of the enlarged share base. It stems from a 2025 deal that converted roughly $8.9 billion of CHIPS Act grants and Secure Enclave funding into equity at $20.47 a share. The stake is passive: no board seat, and the government agrees to vote with Intel's board on most matters [9]. That same arrangement is why Intel's GAAP results look alarming at first glance. A portion of those shares sits in escrow and is accounted for as a derivative liability that is marked to market each quarter. As Intel's stock soared, the fair value of that derivative liability rose with it, producing a $12.5 billion non-cash mark-to-market charge in the second quarter alone. Strip it out and Intel's Q2 operating income was positive at about $1.8 billion — its Products business earned roughly $4.8 billion of operating income while Intel Foundry lost about $2.1 billion — and although trailing operating income was only roughly breakeven, the company still generated about $8.1 billion of operating cash flow in the first half [8]. In other words, the headline "$11 billion loss" is largely an artifact of a rising stock price inflating a government-share obligation, not a company burning cash.

The SoftBank headline that isn't quite what it looks

Part of this week's Intel narrative was a striking statistic: SoftBank's Masayoshi Son had 66.81% of his firm's disclosed U.S. equity holdings in a single stock — Intel, a roughly $12 billion position within an ~$18 billion 13F [4]. Read quickly, it sounds like one of the world's most famous investors making an all-in bet. Read carefully, it is more mundane: the figure covers only SoftBank's U.S.-listed, 13F-reportable holdings, not its entire portfolio, and SoftBank reported the same 86,956,522 Intel shares at both recent quarter-ends — so the Intel position's jump in value came from price appreciation, not fresh Intel buying [4]. SoftBank does have real skin in the game — a $2 billion strategic investment that closed in 2025, alongside NVIDIA's $5 billion equity stake — but the "67% of the portfolio" figure is a byproduct of the rally, not a new vote of confidence, which is part of why the headline lost its lift.

Why it matters

Intel is the largest listed bet on whether the United States can build a leading-edge chip manufacturer to rival TSMC — and, with the U.S. government now among its largest shareholders, it is unusually entangled with national industrial policy. A ~$23 billion raise, marquee strategic backers, and a near-quadrupling in a year all say the market is taking the turnaround seriously; the size of the Foundry losses and the capital bill say it is far from finished. That is why a day like August 18 matters beyond Intel: when a group that has priced in years of AI and turnaround success gives back 5–6% together, it is a reminder of how much optimism is embedded across the semiconductor complex, from memory to equipment to foundry.

How the sector traded the same day

Intel's decline was part of a broad, single-day pullback in semiconductors — a profit-taking rotation rather than an Intel-specific verdict [7]:

That the biggest AI-era winners fell together points to a group-wide rotation. Intel dropped a little more than its peers, consistent with its added Intel-specific overhangs — the dilution and the SoftBank re-read — layered on the same profit-taking [7].

What the Street did with it

The notable analyst move around the raise came from Bank of America's Vivek Arya, who reiterated a Buy but trimmed his target to $145 from $160, treating the capital raise as evidence of foundry ambition rather than distress [6]. More broadly, the Street is split: the consensus 12-month target sits near $115, above the ~$97 price, but the range runs from about $75 to $200 [6]. Few large-cap stocks carry a target spread that wide, and it captures the core disagreement — how quickly Intel's foundry arm can turn from loss to profit, and how much to pay today for that possibility.

What to watch

• Foundry customers. The turnaround thesis needs external customers committing volume to Intel's 18A process; watch for named design wins and capacity commitments [2].
• Foundry losses narrowing. Intel Products is profitable; Intel Foundry lost about $2.1 billion in the quarter. Watch the Foundry operating loss, its gross margin, and any timeline toward break-even [8].
• Capital needs and further dilution. This ~$23 billion raise, plus its roughly net-cash balance sheet after the deal, funds the near-term plan; watch whether Intel needs to raise again, which would dilute holders further [2].
• The backers — including Washington. The U.S. government (~8.4%), NVIDIA and SoftBank are part of the bull case; watch whether their involvement deepens into commercial relationships, and how the government-share accounting swings future GAAP results. Moves are tracked on the INTC stock page and the earnings calendar [9].

Illustrative valuation sensitivity

The scenarios below are a descriptive, author-weighted exercise — not a forecast, target, recommendation, or intrinsic fair-value calculation. Because Intel's GAAP earnings are distorted by the government-share charge and its Foundry arm is pre-profit, the price points are a Street-target-based scenario exercise, anchored to the analyst target distribution (average ~$115; range ~$75–$200) and to turnaround milestones rather than to an earnings multiple [6]. The weights are the author's subjective assumptions and sum to 100%.

Weighting those (0.30 × $200 + 0.45 × $115 + 0.25 × $75) gives an author-weighted scenario value near $130, above the ~$97 quote [1]. The point is the width of the band, from $75 to $200: analysts disagree sharply on how Intel's foundry bet resolves, and the same capital that funds the bull case is the dilution that deepens the bear case. This is a Street-target-based scenario exercise and descriptive analysis of a corporate event, not investment advice or a valuation substitute.

INTC data snapshot — August 18, 2026 (intraday, session open)

The August 18 figures are intraday snapshots that moved through the session, not closing prices; Intel was down about 4% at the open and roughly 6% later in the morning.

The setup at a glance