Google orders over 3 million AI chips from Intel as TSMC capacity crunch forces industry shift

Google reportedly books Intel for packaging more than 3 million TPUs in 2028 -- SK hynix is testing Intel's EMIB packaging for HBM integration

TSMC's CoWoS lines are sold out through 2027, and Intel's EMIB is the only credible second source. Google has placed an order for Intel to build more than 3 million of its TPUs in 2028 after months of testing Intel's advanced packaging, according to The Information, citing four people familiar with the matter. They claim that Nvidia is evaluating Intel to build a future processor that fuses four GPU dies into one unit, tied to its Feynman architecture due in 2028, and that SK hynix is testing whether its high-bandwidth memory works reliably with Intel's packaging. Specifically, SK hynix needs to know whether Intel can run packaging to the standard that AI accelerators demand. TSMC's CoWoS is the industry-standard process for it and has been oversubscribed for more than two years. Intel's embedded multi-die interconnect bridge, or EMIB, is the only alternative AI chip makers can realistically qualify at volume before the end of the decade. This isn't a first for Intel: Google and Amazon were reported to be in active discussions for their custom AI processors back in April, but the remarks from these sources move those "discussions" to a solid unit figure and production timeline, adding in SK hynix qualification that would ultimately determine whether any of it reaches Nvidia accelerators. CoWoS bottlenecked TSMC's leading-edge wafer lines and its CoWoS packaging are both at capacity. At the company's annual shareholders' meeting in Hsinchu on June 4th, CEO C.C. Wei said, "It will be a long time before we can meet customer demand," telling shareholders that the company simply can't satisfy American customer demand for years, even as it builds out U.S. capacity. He had already told the Semiconductor Industry Association last November that TSMC's advanced-node capacity falls "about three times short" of demand. The queue for CoWoS is concentrated across a handful of buyers. Nvidia is naturally expected to account for the majority of global CoWoS demand -- about 60% this year -- with Broadcom and AMD absorbing another 26% between them, leaving custom-ASIC designers and smaller AI-chip makers waiting behind the largest GPU order book in the industry. But the industry can't wait, and both these smaller players and hyperscalers alike with multimillion-unit roadmaps need to qualify a second packaging solution rather than wait for capacity that TSMC says will be short for years. As for EMIB vs. CoWoS, they solve the same problem in opposite ways. CoWoS mounts every die on a large silicon interposer that all signals and power must cross, and the interposer scales with package size, so reticle-class designs waste silicon at the edges. EMIB, meanwhile, embeds small silicon bridges in the organic substrate only where two dies need to connect, with no interposer at all. Intel cites package utilization near 90% EMIB against roughly 60% for interposer-class packaging, because small bridges tile efficiently while large interposers don't. Bernstein analysts estimate EMIB packaging costs a few hundred dollars per chip against $900 to $1,000 for CoWoS on a Rubin-class processor, though the firm flags the fact that there's a "lack of an external production track record" in that estimate. As always, there's a trade-off: standard EMIB routes power around the bridge through the substrate in long, resistive paths. That might have been acceptable for Sapphire Rapids and Ponte Vecchio, but not for HBM4-class accelerators that draw more current. EMIB-T closes that gap by adding through-silicon vias to the bridge die for vertical power delivery, and it's set to enter production fab rollout this year. Intel has said EMIB-T supports HBM3, HBM3E, HBM4, and future HBM5 stacks and scales to a 120mm x 180mm package carrying more than 38 bridges and over 12 reticle-sized dies. Jaguar Shores, the successor to the canceled Falcon Shores accelerator, is the likely first product to use it. Gated by SK? Working with SK hynix could be a huge boon for Intel, with the qualification of its packaging by the South Korean memory giant potentially deciding whether it reaches flagship AI silicon or not. SK held a 57% share of HBM revenue in Q4 2025 per Counterpoint Research, and UBS expects it to take roughly 70% of the HBM4 supplied for Nvidia's Rubin platform this year. HBM stacks are themselves a packaging problem: multiple memory dies bonded vertically through TSVs, then mounted next to a host processor with tight tolerances on power and thermal behavior. Validating those stacks on EMIB rather than a CoWoS interposer is the test of whether Intel can package memory to the standard Nvidia and Google require. An official thumbs-up from SK, or an HBM-4-on-EMIB-T production result, would convert Intel's packaging from "tested" to "trusted." But, until (or if) that happens, the split between accelerator types will remain: ASIC designers running lower memory bandwidth, including Google and Meta, can adopt EMIB sooner, while bandwidth-bound GPUs stay on CoWoS longer. Intel still needs to prove EMIB No named external AI customer is in EMIB or Foveros volume production today. Intel runs EMIB in its own server CPUs, including the 18A Clearwater Forest part whose 17-tile package uses 12 bridges, but every specifically named outside engagement so far, including Google's order, points at 2027 or 2028 products or remains an evaluation. Intel Foundry lost $10.3 billion on $17.8 billion of revenue in 2025, and in Q1 2026, the division posted $5.4 billion in revenue against a $2.4 billion operating loss, with external customers accounting for just $174 million of the total. CFO David Zinsner told the Morgan Stanley TMT conference in March that the foundry is close to closing deals worth "billions per year in terms of revenue" on advanced packaging alone, against a pipeline he had earlier measured in the hundreds of millions. Another unknown is process yields: Intel uses 18A, its first node with gate-all-around transistors and backside power, for Panther Lake and Clearwater Forest, an internal proving ground before courting outside logic customers. However, Intel's most recent guidance is that yields are improving 7 to 8 percent each month, accelerated by enhanced cooperation with external partners.

Intel stock jumps after report claims Google ordered 3 million AI chips

Serving tech enthusiasts for over 25 years. TechSpot means tech analysis and advice you can trust. Rumor mill: Intel's effort to regain its footing in advanced chip manufacturing is drawing fresh interest, as some of the biggest players in AI look beyond their traditional suppliers. Alphabet's Google has reportedly ordered more than three million tensor processing units from Intel for production in 2028. If the order materializes, it would give Intel's contract manufacturing arm a significant boost as the company works to regain its standing in leading-edge chip production. The reported deal comes amid mounting strain on the global chip supply chain, reports The Information, which cited people with direct knowledge of the talks. TSMC, long the dominant player in producing cutting-edge chips, has struggled to keep up with demand fueled by the rapid expansion of artificial intelligence. That imbalance has pushed major AI chip designers to look at other manufacturers, including Intel, rather than relying solely on TSMC. Google's Tensor Processing Units are central to that shift. The company has been steadily developing its in-house chips to reduce reliance on Nvidia's GPUs, while also using them to strengthen its cloud offerings. Securing enough manufacturing capacity for future generations has become increasingly important as AI workloads grow more complex and compute-intensive. Nvidia itself appears to be testing the waters with Intel. The company is evaluating whether Intel's technology could support a design that combines four graphics chips into a single processor, according to the report, though it has not placed an order. Even so, Nvidia's consideration of Intel shows how tight high-end chip capacity has become. Intel's strategy under CEO Lip-Bu Tan has focused heavily on turning its manufacturing arm into a competitive alternative to TSMC. The company has rolled out an ambitious roadmap that includes next-generation process technologies like 14A, while actively courting large external customers. There are signs that the approach is starting to resonate. Intel has drawn investment from the US government and SoftBank and secured Tesla as a customer for its 14A process, which will be used for chips tied to Elon Musk's planned Terafab AI facility in Austin. A separate report last month said Intel had also reached a preliminary agreement to manufacture chips for Apple after more than a year of talks. Government support has also played a role. "Beyond the standard need to diversify, Google and Nvidia are even more motivated than usual to work with Intel. Supporting Intel supports US-based manufacturing, which is important for the relationship with the US administration," D.A. Davidson analyst Gil Luria told Reuters. Investors reacted quickly to the latest developments. Intel's shares rose more than 9% in early trading following the report, adding to a surge that has already seen the stock climb nearly 169% this year. Intel's year-to-date share price Intel declined to comment on the reported Google order, while Alphabet and Nvidia did not respond to requests for comment. Reuters said it could not independently verify the report. As demand for AI continues to strain existing supply chains, companies that once relied heavily on a single manufacturing partner are starting to hedge their bets. For Intel, that shift could open the door to a long-awaited comeback - if it can deliver on the technology.

Google and Nvidia eye Intel as a TSMC backup for AI chips

Google has reportedly placed an order for more than 3 million of its TPUs in 2028, and Nvidia is trialling Intel's 18A process, a rare vote of confidence in America's struggling foundry. For years, the AI boom has run through a handful of factory floors in Taiwan. That dependence is starting to look risky enough that even Nvidia and Google are shopping for a backup, and the unlikely name on the list is Intel. Google has placed an order with Intel to manufacture more than three million of its in-house tensor processing units in 2028, while Nvidia is evaluating Intel's advanced packaging and its most cutting-edge 18A process for future chips, The Information reported on Monday, citing four people with direct knowledge of the talks. Intel's shares jumped about 12 per cent on the news. The driver is scarcity. TSMC, the Taiwanese giant that manufactures virtually every leading-edge AI chip, is straining to keep up with demand, with the squeeze worst in the advanced-packaging lines that stitch chips and memory together. For the companies designing the world's most sought-after silicon, relying on a single supplier in a single country has become a strategic liability. The two approaches differ in seriousness. Google's is a firm order: more than three million TPUs in 2028, after months of testing Intel's packaging, part of a build-out Morgan Stanley estimates at over six million TPUs across 2027 and 2028. Nvidia has not committed. It is running early trials, including multiproject wafer runs on 18A, and testing whether Intel can build a processor that fuses four graphics chips into one, a design tied to its Feynman GPU architecture due in 2028. For Intel, even cautious interest is a milestone. The one-time leader of American chipmaking has spent years trying to turn its contract-manufacturing arm into a credible rival to TSMC, with limited success and heavy losses. It has been courting Apple as a foundry customer, and both the US government and Nvidia have taken equity stakes in the company. Actual orders, rather than goodwill, are what the turnaround has lacked. The near-term opening is in packaging rather than the most advanced chipmaking, where Intel still trails. SK Hynix is reportedly testing whether its high-bandwidth memory works reliably with Intel's packaging too. Whether 18A can match TSMC on yield, the share of chips that come out usable, remains the open question that has tripped Intel up before. It also underscores how far Google has gone to control its own silicon. Its tensor processing units now ship in the millions, and it has been spreading orders across a roster of partners to reduce its reliance on both Nvidia and TSMC. Adding Intel is one more hedge. None of this dethrones TSMC, whose lead in leading-edge manufacturing remains vast. But it is a sign that the AI industry's most powerful buyers no longer want all their eggs in one Taiwanese basket, and that Intel, written off by many, may yet have a role to play in supplying them. For a company that has needed a customer to believe in its comeback, Google just became one.

Google reportedly orders at least three million chips from Intel to arrive in 2028, as TSMC struggles to keep up with the AI boom

Despite a rough couple of years for the company, Intel could be set to make a turnaround with both its handheld chips performing well, and its Foundry business taking on substantial orders, at least if recent reports prove to be true. It seems like the AI boom has really squeezed semiconductor leader TSMC's supplies dry, and Intel is now being tapped to provide some chips in its place. As reported by The Information, Google has reportedly ordered at least three million chip orders from Intel, in the form of TPUs (Tensor Processing Units), according to two people familiar with the company. This Google-designed unit is specifically created for use in neural network machines. In case that sounds like gibberish, it's basically Google, yet again, further committing to AI. These TPUs not only go towards developing Google's own AI services but will also be sold to companies like Apple and Meta. Google is reportedly expected to make more than six million TPUs between 2027 and 2028, so it will presumably be relying on both TSMC and Intel to provide the facilities to help make these chips. TSMC is the world's leading semiconductor manufacturer, and Intel's biggest competition. So, in a sense, the ongoing memory crisis sucking up all the resources is somewhat beneficial to Intel, which may not have been able to make such a deal if TSMC were able to meet demand. That's before mentioning that the TPU demand wouldn't be so high without the AI boom. As well as this, Nvidia is reportedly testing Intel to see if its chips can be used in its next major project, seemingly its next GPU architecture. Currently codenamed Feynman, it supposedly combines four graphics chips into one unit, so if Intel can win the contract, that could be just one more step in establishing the Foundry's foothold in the gaming market. Nvidia is reportedly TSMC's biggest buyer, beating out Apple, which used to be at the top of the list, with the company providing the chips in all of its phones, laptops, and tablets. However, TSMC has been struggling to meet demand for a while now. It is now reportedly sold out until at least 2028, with even its Arizona fab being fully booked before it is even built. Just last year, TSMC announced a $100 billion investment plan into three US-based fabs, but it's not the only one with expansion plans. Just last week, Intel seemingly started targeting a $3.3 billion factory in India. It seems both companies have big plans to ride the AI wave as far as it will take them. One can only hope that this extra investment will contribute to stabilising supply sooner, rather than later. I can't hold out on upgrading my PC forever.

Intel Foundry comeback! Is a bigger Intel stock surge ahead? INTC shares surge 9.16% today as Google's 3 million TPU order and Nvidia testing fuel Intel's AI revival

Intel stock is back in focus. INTC shares surged 9.16% after reports that Google ordered more than 3 million TPU AI chips from Intel for 2028. Nvidia is also testing Intel's advanced 18A process. As TSMC faces capacity pressure, Intel Foundry is emerging as a serious AI manufacturing contender, reshaping investor expectations. Intel stock surged over 10% on Monday in a session where most of the market was bleeding out. The S&P 500 fell 2.6%, the Nasdaq dropped 4.2%, and semiconductor peers struggled to hold ground. Yet Intel opened at $111 and held firm near $109, defying gravity. The reason was a single, credible report from The Information: Google had placed a real order with Intel to manufacture more than 3 million TPU chips, and Nvidia was actively testing Intel's most advanced chip process. For a company that spent years fighting for credibility in the foundry space, this was more than a headline. It was a signal that Intel's long and painful turnaround may finally be finding solid ground. Intel stock has lived in a state of perpetual comeback narrative for years. But narratives don't move stocks by 10% on a down day. Hard orders do. And the Google TPU deal -- confirmed by four people with direct knowledge, cited in the report -- gave Intel's foundry story exactly the commercial validation it had been missing. Investors who had watched Intel lose ground to TSMC and Samsung got something tangible to price in. Why Google's TPU order to Intel changes everything for Intel stock Google's Tensor Processing Units are not generic chips. They are the proprietary AI processors that power Google's model training and inference, and Google now sells TPU computing access to major customers including Apple and Meta. When Google decides to manufacture over 3 million of these chips through Intel by 2028, it is not dabbling. It is placing a significant production bet after months of testing Intel's advanced packaging technology. That testing period matters -- it means Intel passed real technical scrutiny before this deal was struck. Morgan Stanley's read on this deal is especially telling. The bank estimates Google will produce more than 6 million TPUs across 2027 and 2028 combined, which means the Intel Foundry relationship could scale substantially beyond the initial order. With Morgan Stanley also projecting roughly 30% year-over-year revenue growth in Intel's data center segment for 2026, the foundry win and the CPU tailwind are reinforcing each other at exactly the right moment for Intel stock. Is Nvidia testing Intel's 18A process a bigger long-term catalyst? The Nvidia angle is less certain but potentially larger in scale. According to two sources cited in the report, Nvidia is running early trials using Intel's 18A process -- the company's most advanced 1.8-nanometer manufacturing node -- through multiproject wafer runs. The specific goal is to evaluate whether Intel can produce a processor that integrates four GPU dies into a single package, tied to Nvidia's Feynman series GPU architecture planned for 2028. No order has been placed. But the fact that Nvidia -- the world's most valuable chipmaker -- is seriously testing Intel's manufacturing process rather than defaulting entirely to TSMC is a credibility inflection point for Intel's foundry ambitions. "In the last four weeks, I have had all CEOs calling me, saying 'I need more CPU.'" -- Lip-Bu Tan, Intel CEO Intel CEO Lip-Bu Tan had already signaled that AI inference workloads were driving a sharp surge in demand for Intel's data center CPUs. That demand, layered on top of the foundry developments, is building a coherent growth thesis for Intel stock that didn't exist cleanly even six months ago. TSMC capacity crunch is quietly handing Intel a historic opportunity Behind all of this is a structural shift that is easy to underestimate. TSMC's manufacturing capacity is strained. The AI chip boom has overwhelmed its leading-edge wafer lines and advanced packaging production. Companies like Google and Nvidia are not turning to Intel because Intel is suddenly perfect. They are turning to Intel because TSMC cannot do everything, and the risk of single-source dependency in AI hardware is becoming too large to ignore. Intel is benefiting from being the most credible alternative at a moment when the industry desperately needs one. For Intel stock, the significance of Monday's move isn't just the 10% gain. It's what that gain represents in a deeply red tape -- a company-specific catalyst powerful enough to defy a broad market selloff. That kind of counter-trend strength reflects genuine re-rating, not just sentiment. Intel's foundry strategy has struggled for years to earn market confidence. Monday, in a session where almost nothing else worked, it earned some. What investors in Intel stock should watch next The immediate catalyst is priced in. What matters now is whether the Google TPU relationship scales toward Morgan Stanley's 6 million unit estimate, whether Nvidia moves from testing to an actual 18A order, and whether Intel's 18A process yields at commercial scale -- something it has not yet demonstrated publicly. Intel stock has a 52-week range of $18.96 to $132.75. At $108, it is far from those highs but moving with real momentum for the first time in a long while. The foundry narrative just got a facts-based foundation. Whether Intel can execute on it remains the central question.

US market today: Alphabet taps Intel to make three million in-house chips: Report

Nvidia has not placed an order with Intel yet, but has been evaluating whether the company's technology can be used to make a processor that combines four graphics ⁠chips into a single unit, the report said. Alphabet's Google has placed an order with Intel to manufacture more than three million tensor processing units in 2028, The Information reported on Monday, citing people with ⁠direct ⁠knowledge of the discussions. Nvidia has not placed an order with Intel yet, but has been evaluating whether the company's technology can be used to make a processor that combines four graphics ⁠chips into a single unit, the report said. US MarketsPowered By As on 06 Jun 2026, 01:30 AM IST S&P 500 Top Gainers Cooper Companies67.34(8.58%) Kimberly-Clark99.04(6.19%) Clorox94.14(5.03%) Kenvue17.71(4.92%) Gainers" S&P 500 Top Losers Micron Technology864.01(-13.25%) Teradyne357.93(-12.03%) First Solar279.01(-11.41%) Intel99.17(-11.28%) Losers" Intel's shares surged more than 13% ⁠in premarket trading. Alphabet, Intel and Nvidia did not immediately respond to Reuters requests for comment. Reuters could not independently verify the report. The news is the latest sign that a turnaround at Intel is gathering steam after years of management blunders allowed TSMC to take the lead in chip manufacturing. Tesla ⁠CEO Elon Musk said in April that the EV maker plans to use Intel's next-generation 14A manufacturing process to make chips at its Terafab project, an advanced AI chip complex Musk has envisioned in Austin.

Intel gains ground on Google and Nvidia in AI chip manufacturing

Intel could receive a major boost from Google for its foundry business. According to The Information, the group has reportedly secured an order for over three million Tensor Processing Units (TPUs) scheduled for production in 2028. The report also indicates that Nvidia is exploring the use of Intel's manufacturing technologies for a future processor integrating four graphics chips, although no formal agreement has been reached yet. This news provided strong tailwinds for Intel shares, which surged by over 12% on Monday. If confirmed, the Google order would represent a significant milestone in the turnaround strategy led by CEO Lip-Bu Tan. Intel is seeking to revitalize its semiconductor manufacturing business after losing ground to TSMC. Surging demand linked to artificial intelligence and capacity constraints at the Taiwanese foundry are prompting several industry players to diversify their suppliers, offering Intel an opportunity to regain prominence in the global production chain. The group has already secured several financial and industrial endorsements in recent months, notably from the Trump administration, Nvidia, and SoftBank. Intel has also signed an agreement with Tesla for its future 14A manufacturing process and has reportedly reached a preliminary deal with Apple to produce certain chips. For many analysts, partnering with Intel now serves not only to diversify supply chains but also to bolster semiconductor production on US soil, an objective encouraged by American authorities.