Intel shares surge 11% on news of Google contract
The market reacted with a rise in Intel's stock price amid confirmation of a large order from Alphabet. Investors positively assessed the progress of Intel's contract business and plans to expand production capacity in the US amid TSMC's capacity shortage.
Analytical article: Intel soared 11% — triumph or a $50 billion bubble?
The market bought hope for a contract that will start generating money in two years. Until then — billions in losses every quarter.
On June 9, 2026, Intel shares crashed 13.5% at the open. It seemed the week would be a nightmare for the semiconductor sector — $1 trillion in market cap evaporated from the market in one Friday. Then the incredible happened: within hours, shares not only recovered the drop but surged 11%. The trading day closed at $110.27.
What happened? The Information published an article where anonymous sources revealed a Google order for 3 million TPUs for 2028. The market, which had just experienced its worst session in five years, instantly switched to "buy everything" mode.
But let me explain why this 11% jump is pure self-deception. Why behind the facade of a "loud victory" lie $2.4 billion in quarterly losses, technological problems Intel has yet to solve, and a complete lack of confirmed contracts with Nvidia.
[The Gist]: What's Really Happening
Officially: Intel shares rose on news that Google allegedly ordered 3 million TPUs for 2028. Unofficially: the market bought rumors that neither side confirmed.
Let's look at the facts. The Information is a respected publication, but its sources remain anonymous. Google and Nvidia declined to comment. Intel "sidestepped the question." No signed documents are publicly available. Yet the market reacted as if the contract were already on the table.
What is actually known: Google might indeed order TPUs from Intel. But 2028 is an "eternity" in the chip world. By then, TSMC will be on 1.4nm processes, while Intel only promises to finalize 18A and 14A. Plus, Morgan Stanley estimates total TPU production in 2027-2028 at 6 million units. So Intel's share, even if the deal goes through, would be 50% — no more.
And now the most important thing that news feeds don't cover. Intel Foundry is a loss-making business. In Q1 2026, the division posted an operating loss of $2.4 billion on revenue of $5.4 billion. Of that revenue, external orders (i.e., not from Intel itself) were a paltry $174 million. 96% of fab capacity is used for Intel's own chips. And this is while the company spends $5 billion per quarter on capital expenditures alone, resulting in a free cash flow of negative $2 billion.
The market bought the news of a contract that could save Intel Foundry... in two years. Until then, another $20-25 billion in losses need to be financed.
Timeline and Context
To understand how we got here, we need to trace Intel's path over the past year and a half. This is not a one-day story but a series of events that led investors to buy any positive rumor.
January 2026: Intel reports Q4 2025 results. Revenue $13.7 billion (down 4% YoY), net loss $0.6 billion. Guidance for Q1 2026: $11.7 to $12.7 billion. Shares fall 17%. New CEO Lip-Bu Tan admits: "18A process yield is lower than I want." KeyBanc estimates yield at 60% — catastrophically low for commercial production.
April 2026: Tesla announces plans to use Intel's 14A process for Terafab in Austin. This is the first major external customer. But an important detail: 14A is the next process, which should be better than 18A. So Tesla doesn't trust Intel's current technology and is waiting for the next one.
May-early June 2026: A series of positive signals. Microsoft is already using Intel for Maia 2 chips. Morgan Stanley notes sustained demand for server CPUs. At Computex, Lip-Bu Tan states that "several CEOs asked for more CPUs" — demand exceeds supply even with poor yields. Nvidia acquires a $5 billion stake in Intel, SoftBank invests $2 billion.
June 4-8, 2026: The Information publishes material about Google and Nvidia tests. The market soars. Intel CFO David Zinsner at the Bank of America conference tries to reassure investors: "14A is on track, we're ahead of where 18A was at the same maturity stage." But admits 18A was a mistake: "We tried to do too much at once."
Now — the key point the market ignored. In the same Information report, it says Nvidia is testing Intel on "multiproject wafer test runs." This is not a contract. It's "let's see what you can do." A contract is years of testing away.
Who Wins and Who Loses
Winner #1: Intel shareholders (short-term). Those who bought at the 13.5% drop and sold at the 11% peak made 24% in one day. These are speculators. They will sell on the next correction. Long-term investors who entered at $110 will likely end up in the red.
Winner #2: Nvidia (paradoxically). Nvidia did not confirm a contract, but rumors of its "tests" lifted Nvidia shares by 1.6-2%. Moreover, if Intel can relieve TSMC, TSMC will allocate more capacity to Nvidia. Nvidia also owns a $5 billion stake in Intel — Intel's rise benefits Nvidia as a shareholder.
Winner #3: Google (strategically). Even if the deal isn't signed, the leak itself signals to TSMC: "We have an alternative." Google gains leverage over the Taiwanese monopolist.
Loser #1: Investors who bought at the $110 hype. Bank of America has a $40 price target. The consensus of 51 analysts is Hold with an average target of $98.15. The current price ($110) is almost 12% above the target. Even optimists see no upside.
Loser #2: TSMC (momentarily). TSMC shares fell 6% on the news. Investors feared Google was leaving. But that's emotion. TSMC remains a monopolist in advanced processes at least until 2028. One Google contract doesn't change that.
Loser #3: Market rationality. A situation where a company with a loss-making division and 60% yield on 18A rises 11% on a rumor of a contract two years away is an irrational bubble.
What the Media Leaves Out
News feeds write about "Intel's breakthrough" and "victory over TSMC." But they omit three things that change the whole picture.
Insight #1: Intel still can't properly manufacture chips on 18A.
CFO David Zinsner admitted at the Bank of America conference in early June: 18A faced problems because Intel "tried to do too much at once." Yields remain secret — if they were good, Intel would shout it from the rooftops. KeyBanc estimates yield at 60%, which for commercial AI chip production (where every square millimeter costs tens of dollars) is a disaster. For comparison, TSMC's mature processes have yields of 90-95%.
Intel is trying to "fly the plane and fix the wing at the same time." And the market buys shares as if the wing is already fixed.
Insight #2: 2028 is not tomorrow. It's an abyss.
The contract in question is for 2028. But what about 2026 and 2027? Losses. Intel Foundry will burn another $20-25 billion before earning its first dollar from Google. And that's assuming 14A launches on time (which 18A didn't). In 2025, foundry operating loss was $10.3 billion. In 2026, it won't be better.
Intel's market cap is about $500 billion. Yet the company generates negative free cash flow and loses billions each quarter. This is not a "growth company." It's a "company on life support."
Insight #3: Nvidia hasn't signed any contract.
The media mixed two stories: (1) Google may order TPUs, (2) Nvidia is testing Intel 18A. The second is not a contract. It's "trial batches" (multiproject wafer test runs). Nvidia is seeing what Intel can do. But even if tests succeed, mass production is years away. Nvidia, unlike Google, is in no hurry. They have TSMC, and TSMC works.
Moreover, Nvidia already invested $5 billion in Intel. That's not trust in the technology. It's risk hedging: "if TSMC falls, we have a backup." But Nvidia won't put all eggs in one basket.
Forecast: Next 30 Days and 90 Days
Next 30 Days (to mid-July 2026)
1. Intel stock correction of 10-15%. The 11% rise on unconfirmed rumors is an overheat. When the initial emotions fade, analysts will start asking about yields, foundry losses, and the lack of signed contracts with Nvidia. Shares will correct to $95-100.
2. TSMC will announce accelerated capacity expansion. To counter the negative from "Google leaving," TSMC will announce new investments in Arizona or acceleration of the second phase. This will put new pressure on Intel — TSMC will remind the market who's king.
3. Nvidia will officially deny contract rumors. Nvidia doesn't like decisions made for it. After Nvidia shares rose slightly on the rumors, the company will issue a statement: "We are testing various technologies but have no production commitments with Intel." Intel shares will fall 5-7% the same day.
Medium-term Forecast (90 days, to September 2026)
1. Intel will report weak Q2 results. Forecast: revenue around $12-12.5 billion, foundry loss again $2.5-3 billion. Free cash flow negative $1.5-2 billion. Analysts will revise price targets downward. Bank of America will reaffirm $40. Morgan Stanley may downgrade.
2. Meta or Amazon will place a "quiet" order with Intel, but not publicly. To avoid inflating a bubble while securing diversification. The contract will be small ($100-200 million) and for 2027-2028. It won't become public news — only insider sources.
3. China will start poaching Intel engineers. Seeing Intel's weakness on 18A, Chinese companies (SMIC, Huawei) will intensify headhunting. Engineers who worked on 18A and 14A will receive offers with 50-100% salary increases. Intel will lose key specialists at the most critical moment.
4. Rumors of market consolidation: will someone buy Intel? At the current $500 billion market cap, Intel is an expensive asset but not unreachable for a consortium (e.g., SoftBank + Nvidia + Apple). Acquisition rumors will circulate, but a deal is unlikely due to antitrust restrictions in the US and China. Nevertheless, this will support the stock price above fundamental values.
Summary for those who read to the end: Intel's 11% rise is not a victory. It's hysteria. The market is so desperate to find the "next Nvidia" that it's ready to buy any positive rumor. But Intel's fundamental problems haven't gone away: a loss-making foundry, poor yields, dependence on internal orders, and a lack of confirmed large external contracts.
If you bought Intel at $110 — congratulations, you bought hope for 2028 at 2026 prices. If you held Intel and didn't sell at the peak — you missed the moment. And if you're just watching from the sidelines — just remember: in semiconductors, the winner is not the one who shouts loudest about future contracts, but the one who can manufacture chips with decent yields today. Intel can't yet. TSMC can. That's the whole analysis.
— Editorial Team
No comments yet.