Intel’s Volatile Turnaround: From 270% Surge to Sharp Pullback as AI and Foundry Bets Face Tests

Intel shares have traced a remarkable path in 2026. The stock soared more than 270 percent in the first half of the year. Then it gave back chunks of those gains in early July. One investment fund felt the absence keenly.

Columbia Threadneedle Investments missed out on the rally in its Columbia Seligman Global Technology Fund. The fund did not own Intel. That lack of exposure hurt performance as the chipmaker’s shares climbed sharply during the second quarter. Yahoo Finance reported the details from the fund’s investor letter.

But the story runs deeper than one missed position. Intel posted its strongest revenue growth in more than 15 years for the second quarter. Revenue reached $16.13 billion. That topped expectations. Adjusted earnings came in at 42 cents a share. Analysts had forecast 22 cents. The stock jumped in extended trading after the July 23 release. Investor’s Business Daily covered the beat and the immediate market reaction.

Lip-Bu Tan, Intel’s CEO, pointed to AI as the force behind the numbers. “AI is driving unprecedented demand for compute, and as we continue to execute, Intel is well-positioned to capture sustainable growth across our CPU franchise, ASICs, advanced packaging and vast wafer foundry network,” he said in the earnings release. Dave Zinsner, the CFO, added that the quarter showed “robust demand and improved execution, including volume upside driven by higher factory yields and improved cycle times.” Both quotes come directly from Intel’s official press release.

Data Center and AI revenue surged 59 percent to $6.3 billion. Client Computing and Physical AI Group brought in $8.9 billion, up 13 percent. Intel Foundry posted $5.8 billion, a 31 percent increase. These figures reflect real momentum in areas the company has targeted for years. Yet the market’s response proved anything but steady.

Just weeks earlier the shares had crashed. On July 8 Intel fell 7.67 percent to $110.68. That extended a 21 percent drop over seven days from roughly $140. The decline erased much of the earlier euphoria. Three pressures hit at once. A delay in profitable yields for the 18A process pushed meaningful returns to late 2026 or even 2027. AMD passed Intel in data-center revenue for the first time in the first quarter. And a broader semiconductor selloff followed a Bank of America note warning of an AI chip bubble plus weak results from Samsung. Phemex laid out the sequence and the price levels in detail.

The 18A node sits at the heart of Intel’s ambitions. It entered risk production on schedule. Intel 18A-P followed, delivering 9 percent higher performance at the same power or 18 percent lower power at the same performance. The company moved subsets of its Core Ultra Series 3 processors, code-named Panther Lake, into high-volume manufacturing using ASML’s High NA EUV tools. Progress continues. But yields remain the variable that decides the financial payoff. Investors wanted clearer proof before the July earnings. They didn’t get enough in the preceding weeks. So they sold.

And yet the second-quarter results showed the foundation strengthening. Gross margin expanded more than 12 percentage points on a non-GAAP basis to 41.8 percent. Operating margin swung from a loss to 17.2 percent. The company generated $7 billion in cash from operations. Zinsner noted the team was “meaningfully increasing our investments in equipment, clean room space, and substrates” to meet expected growth in products and foundry services.

Intel has pushed hard into physical AI and robotics. More than 130 customers now adopt or test its Core Ultra Series 3 processors for edge applications. The firm introduced OpenVINO Physical AI, an open-source framework for vision, language, reasoning and motion control. New rack-scale AI infrastructure solutions pair with partners like SambaNova and Foxconn. Xeon 6+ became the first server product on 18A. These moves aim to broaden the addressable market beyond traditional data centers.

Foundry efforts gained traction too. A major reported win involves Google ordering more than three million Tensor Processing Units to be made by Intel in 2028. The deal would mark one of the largest external AI chip manufacturing agreements for the foundry business. Intel also collaborates with Fortinet on a security processor and invests €5 billion to expand capacity for Xeon processors on Intel 3. HotHardware highlighted the Google development and its implications for challenging TSMC’s dominance.

Wall Street had grown optimistic earlier. Price targets rose as the stock extended its AI-fueled run. Some analysts saw the surge as a catch-up after years of underperformance. Intel’s market capitalization briefly topped $550 billion. The shares hit all-time highs not seen in more than two decades. Optimism around potential Apple chip deals and government support under previous policy environments added fuel. Yet the July volatility served as a reminder. Expectations had grown stretched. A price-to-sales ratio above 12 times left little room for disappointment on execution timelines.

Recent updates show 18A yields improving steadily. Intel 18A-P entered risk production in June. The company maintains its roadmap even as it acknowledges the path to industry-standard yields may stretch into 2027. That binary outcome looms over the next several quarters. Third-quarter guidance called for revenue between $15.8 billion and $16.8 billion. Non-GAAP earnings per share of 38 cents. The midpoint implies continued growth but not the explosive acceleration some had priced in.

Leadership changes accompanied the operational shifts. Alex Katouzian now leads the Client Computing and Physical AI Group. Pushkar Ranade serves as chief technology officer. Seok-Hee Lee heads advanced packaging. These appointments signal a sharper focus on execution and technology delivery. Tan has emphasized speed, accountability and customer focus since taking the role.

The competitive picture adds complexity. AMD’s data-center gains reflect strong uptake of its EPYC processors. Nvidia continues to dominate AI accelerators. Intel’s response combines its own Gaudi accelerators, Xeon CPUs and a foundry model that offers an alternative to pure-play manufacturers. Success depends on converting design wins into high-volume, profitable production. The Google TPU order, if realized at scale, would validate the strategy. So would sustained share gains in AI PCs and edge inference.

Investors now watch the balance between near-term financial delivery and longer-term process leadership. Q2 delivered on revenue and margins. The stock still trades well above levels from a year ago. But the July selloff stripped away some of the premium built on hype. Future movement will hinge on tangible proof that 18A yields meet cost targets by the end of this year and that external customers commit at meaningful scale.

Intel’s transformation has always carried high stakes. Massive capital investments in U.S. facilities, pursuit of leading-edge nodes and a shift to foundry services demand patience. The market granted some of that patience in the first half of 2026. It withdrew a portion in July. The coming quarters will test whether the operational progress can outrun the skepticism. One thing appears clear. The CPU renaissance driven by agentic AI has arrived. The question is how large a slice Intel ultimately claims.