chips

$30 billion worth of Broadcom chips

by

Apple has announced an extension of your agreement with Broadcom which, according to the company itself, will exceed $30 billion. For the company, it is the largest single commitment within its American Manufacturing Program (AMP), but it is also another handful of bills that once again demonstrates how the semiconductor industry is being the central axis of the strategy of big technology. The context. Apple has been trying to demonstrate to Donald Trump’s government for more than a year that is moving part of its production chain to American soil. And it has no other choice, since this commitment is framed within of the 600 billion dollar investment plan in four years that the company announced in 2025, after Trump threatened to impose 25% tariffs on iPhones if Apple did not manufacture more components in the country. According to collect According to the Financial Times, that promise is similar to another similar one that Apple already made during Trump’s first term, when it committed to investing $350 billion in the United States. In detail. The new agreement with Broadcom provides for the production of more than 15 billion chips manufactured in the United States and will create hundreds of jobs, according to Apple. its official statement. Broadcom, which was already working with Apple to supply connectivity components, will expand and modernize its plant in Fort Collins, Colorado, with an additional investment of $1.5 billion. That plant will manufacture advanced radio frequency components, including FBAR filters, which allow Apple devices to connect to mobile, Wi-Fi and Bluetooth networks. Neither Apple nor Broadcom have yet confirmed when they will have everything ready to work with this new capacity. Between the lines. The announcement comes two days after Broadcom will inform to the United States Securities and Exchange Commission (SEC) that it had signed new long-term agreements with Apple to develop custom ASIC-type chips (application-specific integrated circuits, increasingly used in artificial intelligence workloads) that will extend until 2031, according to collects Reuters. So the $30 billion figure now confirmed by Apple puts concrete numbers on those long-term deals. In the last year, Apple has been replacing some of the Wi-Fi and Bluetooth connectivity chips that it previously bought from Broadcom with its own designs. This had raised doubts among investors about whether Broadcom would lose weight as a supplier, something similar to what happened with Qualcommwhose 5G modems Apple has been phasing out in favor of its own C-series chips. However, it appears that Broadcom will continue to supply other RF components and, according to Bloombergis also working on technology that will support the first artificial intelligence server that Apple plans to deploy next year. The voices. Apple CEO Tim Cook has qualified the components that will be manufactured in Fort Collins as “essential to deliver the performance and connectivity our customers expect,” and thanked President Trump and his Administration for “supporting important projects like this.” For his part, Broadcom’s CEO, Hock Tan, said he was “proud to continue working with Apple after decades of joint success” and highlighted that the agreement will allow it to expand its manufacturing footprint in Colorado. Why it is important. Apple is not massively restructuring its global supply chain, which remains highly concentrated in Asia. Your strategy, as explains the Financial Times, has been to focus on chips, which are the highest-value components in its devices, and announce purchase commitments with manufacturers that already operate in the United States, such as TSMC in Arizona, Texas Instruments in Texas or GlobalFoundries in New York. The Broadcom deal is, so far, the biggest piece of that puzzle. And now what. The announcement comes at a time of transition for Apple. Tim Cook will step down as director of the company on September 1, position that John Ternus will assumecurrent hardware manager. Cook will remain with the company as CEO and, according to Bloombergis expected to maintain his role as the main interlocutor with the White House, a role that has been key to sustaining Apple’s relationship with the Trump Administration in recent months. In Xataka | The US taught that access to advanced AI can be cut off. China is studying the same thing, according to Reuters, and Europe is watching from the outside

is going to spend 500 million on the chips of the future

by

If you have traveled to Asia to countries like Japan, China or Singapore, it is possible that along your journey you have put your butt on a smart toilet. And it is also likely that that smart toilet was from the TOTO brand. The Japanese firm is famous mainly for its toilets, but the fact that its core is china allows it to be on the front page in the midst of the AI ​​boom and that is not so well known: it has been in the chip business since 1984, manufacturing high precision ceramic components employed in semiconductors. It is not NVIDIA or AMD, but to give us an idea of ​​how serious Toto is and its level of competition, it has just invested a fortune with one goal: to scale its production to the most advanced one-nanometer chips. It should be remembered that IBM just achieved the first chip of that very small size. Towards 1nm chips. Already in February the Pallister Capital fund (one of the largest shareholders of Toto) qualified to a company as “the most undervalued and overlooked beneficiary of AI memory”, highlighting that its chip component manufacturing segment already accounts for more than 50% of its profit. What he also blamed on the Japanese company is the lack of transparency. A few months later, Toto has picked up the gauntlet: just announced an investment of $495 million over five years to scale its business to the most advanced technological horizon in the chip industry: supporting next-generation manufacturing technologies in the nanometer range. Why is it important. The rise of data centers and AI is fueling demand for advanced semiconductors in search of increasingly smaller and more efficient chips. Without advanced materials like those made by TOTO, the miniaturization needed for one-nanometer technology would not be possible. And this diversification is very profitable for Toto. Nikkei Asia collects its astronomical projected figures for this segment: operating profit of 27,000 million yen (146 million euros) for the fiscal year ended in March 2026, an absolute historical record and 32% more than the previous year. It is already what makes the most money for the toilet company. At the state level, this investment is part of Japan’s effort to strengthen its domestic semiconductor supply chain in a context where several countries wish to reduce their dependence on Taiwan and South Korea. Context. Toto began research in the field of advanced ceramics in the 1970s, as Japan’s period of rapid post-war growth was winding down. As relates Toto ceramics business planning department manager Junji Kameshima said, “We wanted to use our ceramics experience to create high-value products.” In 1984 that area was officially established and in 2020 it went from artisanal and low-performance production to playing in the first division. The jump was thanks to Nakatsu’s highly automated plant, with AI systems trained to detect minimal defects: it went from a performance of 50-60% to more than 90% and delivery times were reduced from 180 days to just over 40. Its product portfolio within the semiconductor area was consolidated around three main products: the most important are e-chucks, ceramic discs that hold the silicon wafer during the etching of NAND memory chips. The second are aerosol deposition components, which protect internal walls of the etching chambers. The third are highly durable structural parts used in large LCD panel manufacturing equipment. The three take advantage of a skill acquired making ceramics for the bathroom: ceramic firing of high precision and purity. In detail. This investment of 495 million dollars over five years has three specific lines of action: Expand the machinery at its Oita and Fukuoka plants, already operating at full capacity. Reorient R&D at its Kanagawa plant toward logic semiconductors. Build a new cooking building in Fukuoka, scheduled for January 2027. Part of the investment has already been decided, but the rest will be available depending on market conditions. Thus, if the demand still cannot be met, Nikkei Asia leaks that Toto will consider the construction of a new plant from scratch. Yesyes, but. It seems that days of wine and roses are coming for Toto, based on its solid figures, but there are aspects to take into account. The first is that NAND memory already collapsed a few years ago and could do it again before the Japanese company recovers the investment. On the other hand, this optimistic speech comes from Pallister, one of those interested in Toto doing well (he is an investor). Also, Toto has very few large clients, so a slowdown from any of them can be a severe blow. In Xataka | The fascinating world of Japanese electronic toilets: sensors, microchips and what’s to come In Xataka | Welcome to the AI ​​duopoly: the sector already has a turnover of 80 billion a year, but OpenAI and Anthropic take 89% of the revenue Cover | Toto and Igor Omilaev

South Korea has a plan to dominate in memory chips and robotics. One of a billion dollars

by

South Korea has put on the table a megaproject for the AI ​​era. This is an initiative made up of three public-private projects spread across semiconductors, data centers and the promising industry of “physical AI”, that is, robots and autonomous systems. The advertisement seeks something very specific: that the country does not depend only on selling memory, but on manufacturing other physical products that it anticipates will be consumed massively. Memory chips rule. The largest item of this ambitious plan is, as we all expected, the one destined for the country’s semiconductor industry. Samsung and SK Hynix have committed to investing $585 billion in new manufacturing plants in the southeast of the country, in addition to strengthening factory construction in the Seoul region. They want to double the production of DRAM memories in five years. The future belongs to data centers. The second large part of this plan corresponds to the data centers that are precisely going to take advantage of a large part of those memory chips. SK Group, CS Group and Naver will invest $357 billion to build large-scale AI data centers in areas until now far from the country’s technology centers. According to the Ministry of Science, the final ambition is to achieve a combined capacity of 18.4 GW by 2035, which would make South Korea one of the world’s great AI “nodes.” Robotics, of course. More surprising is the third leg of the plan: the South Korean government has declared that “physical AI” is a “national strategic industry.” These systems, which allow robots or autonomous vehicles to interact with the real world, also want to be part of the future of the country’s technology industry. In three years they want to create a foundational AI model with the philosophy of “models of the world“—the same in which Yan LeCun or Fei-Fei Li work—. Hyundai has its own plan. The South Korean auto industry giant has set aside $5.8 billion to create a robot manufacturing plant and data center. It’s no coincidence: Hyundai is in fact the parent company of Boston Dynamics since 20221, and is already using its local supply chain to help the American company increase production of its Atlas humanoid robot. The goal: build 30,000 units per year in 2028. Humanoid robots in factories. South Korea wants to commercialize humanoid robots in several key industries before 2028, and to achieve this it wants to train the people who will work with them. The joint plan includes a section that talks about training 10,000 new “AI robotics specialists” in the next five years. But. The announcement coincides with an important internal debate: there are political proposals that seek distribute part of the extraordinary profits of the chip manufacturers, and the unions already see the imminent threat of robotics that will replace positions in all types of assembly chains. The opposition to the South Korean government has also criticized the location of these new production centers, which according to them respond more to a political calculation than to a good industrial strategy. In Xataka | Samsung had been the absolute king of technology in South Korea for decades: SK Hynix has just surpassed it

Samsung and SK Hynix have pledged $880 billion in chips and AI. It’s a South Korean survival story.

by

President Lee Jae Myung has summoned the leaders of the world’s two largest memory companies, calling them “national heroes” and describing the plan as a matter of “survival.” This is not usual in political rhetoric. South Korea has presented its largest coordinated technological investment plan in history: At least 1.35 trillion won (about $880 billion) in semiconductors and AI data centers. Samsung and SK Group will build two chip factories each in the southwest of the country. And other companies, led by Naver, will build 8.4 gigawatts of data center capacity before 2029. Why is it important. South Korea produces most of the RAM and HBM on the planet. SK Hynix is ​​the main supplier of HBM chips to NVIDIA, and Samsung is the second. With Google, Amazon, Meta and Microsoft announcing more than 700 billion in capex by 2026, the memory supply chain is the bottleneck that can stop all that expansion. Apple and Microsoft have already announced price increases for their devices due to the increase in the cost of these components. In figures: 880 billion dollars in total investment, equivalent to about 5% of South Korea’s GDP in 2024. 4 new chip factories in the southwest of the country. 8.4 gigawatts of AI data center capacity by 2029. $295 billion: China’s five-year semiconductor investment plan, the benchmark South Korea has in mind. Between the lines. The plan is industrial but also has a lot of politics: Lee’s approval rating has fallen to its lowest level since he took office a year ago, pressured by the economy, the weak won and housing. Locating the factories in the southwest (far from the Seoul metropolitan area, where all advanced production is concentrated) responds to an agenda of territorial redistribution that Lee has made an electoral banner. Yes, but. The announcement did not sit well with the markets. Samsung fell about 5% on the day of the event and SK Hynix fell 1.7%. Investors have read it as a warning of possible oversupply if demand for AI chips slows before factories are operational. Added to this are the known material challenges in this sector: building state-of-the-art factories requires quantities of water, electricity and talent that the government has promised to support… but without yet detailing how. The context. This is not the first time that the Asian country has announced this type of commitment. The difference now is the urgency: the boom of AI has accelerated deadlines and the competitive framework has changed. Japan subsidizes TSMC to build on its soil, China has been closing the gap for years with state financing and the United States has committed tens of billions through the CHIPS Act. In Xataka | Who are Openchip, the Catalan company that designs RISC-V chips… and has just received 115 million from the Government Featured image | Daniel Bernard

Who are Openchip, the Catalan company that designs RISC-V chips… and has just received 115 million from the Government

by

This Monday, June 29, the Council of Ministers authorized an investment of 115.77 million euros in Openchip & Software Technologiesa microelectronics company based in Barcelona and five years old. The operation is channeled by the Spanish Society for Technological Transformation (SETTthe digital SEPI, dependent on the Ministry for Digital Transformation) through the Next Tech facility of the Recovery Plan. It is the largest one-time injection of public capital received by a Spanish technology company in the sector to date. It comes just a week after another move. On June 23, The Generalitat converted part of a 35 million bridge loan into sharesan operation that gave him 5% of the capital and set a implicit valuation for Openchip in around 700 million. With that reference, the 115.77 million from the SETT would be equivalent to a participation of up to 16.54%, which would place direct public control (State and Generalitat) above 20%. Both administrations will have a seat on the council. The Government also included a veto right over any transfer of the headquarters outside of Catalonia. Added to the 111 million already received via PERTE Chip, public support accumulated is close to 262 million. The public supports a good part of the structure. A company that designs, not manufactures Openchip was born in 2021 as a joint initiative of the Catalan engineering group GTDaround 54% of the capital, and the Barcelona Supercomputing Center (BSC-CNS), the center that operates the MareNostrumaround 46%. Today it employs about 300 people, almost all of them engineers, and operates under a fabless– Designs intellectual property and outsources manufacturing to external foundries. The CEO himself has admitted that this production will leave Europe, which in practice points to where these types of projects usually aim: TSMC. The industrial plan estimates investments close to 500 million to deploy the entire infrastructure. What it designs are processors and accelerators based on RISC-V, the open source architecture that has become the European bet (and, ahem, China) to avoid the dependency on x86 (Intel, AMD) and ARM. Its specific product is a vector accelerator for AI and high-performance computing, integrated into DARE SGA1a 240 million European initiative led by BSC itself that distributes the design between Openchip (vector accelerator), the Dutch Axelera (AI processing unit) and the Czech Codasip (general processor). The goal: a European hardware and software proposal operational by 2028. The schedule, the equipment and the exam This is where it is good for everyone to temper expectations. In November 2025, Cesc Guim (pictured above), CEO and former Intel, said that the company had just sent its first prototype to the factory and that commercial production was planned for 2028. The commercial argument is energy efficiency: its designs promise to reduce electricity consumption by 20% to 30% compared to current alternatives. The real comparison can only be made when there is working silicon, not plans. A few weeks ago, in May, Openchip signed Tobías Martínez as presidentformer CEO of Cellnex for almost a decade. Replaced Carlos Kinder in a change that the company did not officially confirm. His profile provides what a startup of 300 engineers with a round of hundreds of millions was missing: plenty of experience in the capital markets. The operation is sold under the convenient modern mantra of ‘European technological sovereignty’, and the truth is that the framework is real: Europe today designs a minuscule part of the world’s chips, and certainly none of the leading ones. But There remain questions that public investment does not solve on its own: Whether Openchip will achieve a competitive product against rivals with a twenty-year advantage (Guim himself has admitted it). Yes, manufacturing will continue to depend on TSMC, which keeps Europe away from the critical link in a long-term dependence. And whether the intensive financing model, with two administrations on the board and a regional veto over the headquarters, will allow the flexibility and agility that a semiconductor business requires to compete. The State has bought shares, a seat on the council and qualified employment in Catalonia. What remains to be seen is whether the chips arrive. And they work as promised. In Xataka | The Valencia family that made a fortune with guano and Coca-Cola now has another project: photonic semiconductors Featured image | Openchip, Xataka

Honor’s new 16-inch laptop weighs only 1.64 kg and debuts the latest generation of Intel chips

by

The market of laptops focused on productivity just received an important announcement. If you are looking for a generous screen to work comfortably but you are horrified by the idea of ​​carrying a heavy device, the Honor company has a new proposal. The launch of the Honor MagicBook 16 2026a device specifically designed to balance mobility, autonomy and cutting-edge performance. Its price is 1,199 euros (with 16 GB of RAM and 512 GB of storage) The price could vary. We earn commission from these links A very light laptop with an 80 Wh battery The main attraction of this model lies in having advanced technical specifications in an aluminum chassis with a sandblasted finish that measures just 15.9 mm thick and weighs only 1.64 kg. Integrate the new processor Intel Core Ultra 5 325manufactured using the Intel 18A process. This chip distributes its tasks across four high-performance cores and four efficiency cores. Furthermore, for avoid performance drops in multitasking or 4K video editingincorporates a cooling system with a large fan, a heat pipe D12 and enlarged dissipation fins. His 16 inch panel It has WQXGA resolution, a screen-to-body ratio of 92.5% and a fluid refresh rate of 120 Hz. In addition, it adds AI Defocus Display technology to reduce visual fatigue through optical simulation. It has an 80 Wh battery optimized with Honor Turbo autonomy greater than 14 hours in office tasks. The included charger is type GaN with 65W of power and offers support for PPS reverse charging up to 60W. This laptop has the SGS resistance certification Exceptional Reliability. Also, your keyboard has IPX2 certification against splashes thanks to an internal drainage system that diverts liquids away from critical components. Lastly, it can be mentioned that uses the Honor WorkStation platformwhich facilitates file synchronization and multi-screen collaboration with smartphones and tablets of the brand itself. ⚡ IN SUMMARY: honor magicbook 16 2016 laptop ✅ THE BEST Lightness in large format– Achieving a weight of 1.64 kg in a 16-inch size greatly facilitates daily transport in any backpack or briefcase. Peace of mind in case of accidents: The IPX2 keyboard drainage system reduces the risk of breakage due to accidental drops of coffee or water. ❌ THE WORST Unique color range… Initially it has been put on sale only in a single color (gray Starry Gray). Exclusive focus on productivity… By not equipping dedicated gaming graphics cards, its performance is aimed at professional work, content creation and office automation. 💡 BUY IT IF… You work remotely, create content, or require ample screen display space but require a lightweight, durable chassis to move around frequently. ⛔ DON’T BUY IT IF… Are you looking for a computer for purely entertainment purposes to run cutting-edge video games with high graphical load? Other 16-inch laptops that might interest you acer Aspire 16 AI Copilot+ PC A16-52M The price could vary. We earn commission from these links Lenovo Thinkbook 16 – 16” WUXGA Laptop The price could vary. We earn commission from these links Some of the links in this article are affiliated and may provide a benefit to Xataka. In case of non-availability, offers may vary. Images | Honor In Xataka | Best ultrabooks. Which one to buy and 11 recommended light laptops from 500 to 2,000 euros In Xataka | Best gaming laptops. Which one to buy and eight recommended computers from 770 to 3,000 euros

An unexpected salvation for the end user emerges from the memory market debacle: Chinese chips

by

The DRAM memory industry is facing a profound structural transformation. Until October 2025 the price of memory chips evolved in a relatively stable way, but from that moment on began a dizzying climb which still continues. In fact, the consultant TrendForce expects the price of conventional DRAM to rise between 58% and 63% quarter-on-quarter before the expiration of the second quarter of 2026. And the artificial intelligence (AI) is behind all this. The three largest chip manufacturers of memory on the planet, the South Korean companies SK Hynix and Samsung Electronics, and the American Micron Technology, They have reallocated about 70% of its production lines to high-bandwidth memories (HBM) to satisfy the currently insatiable demand of data centers specialized in AI. The consequences of this movement did not take long to appear: standard DDR4 and DDR5 memories and their derivatives, which are the most used in the consumer segment, immediately began to become scarce. And its price skyrocketed. In fact, according to the consulting firm GartnerRAM has gone from representing 16% of the total cost of a laptop to 23%. And it is possible that this escalation will continue to develop in the coming months. However, users can cling to the greatest stabilizing agent in the memory market today: Chinese manufacturers. This is the great opportunity for YMTC and CXMT Yangtze Memory Technologies Co. (YMTC) is one of the largest NAND chip manufacturers in China. Its global market share is approximately 13%making it one of the main competitors of Samsung, SK Hynix, Micron, Kioxia or SanDisk. Its weight in the Chinese market is very great, especially because US sanctions They prevent American and South Korean memory manufacturers from selling their most sophisticated integrated circuits to their Chinese customers. On the other hand, Changxin Memory Technologies (CXMT) is one of the Chinese companies specialized in the production of memory chips, and, like other companies in the country led by Xi Jinping, it has chosen to compete in this very attractive market by deploying a very aggressive pricing policy. CXMT in particular has increased its DRAM chip production capacity almost five times during the last four years, which has allowed it to increase its global market share until reaching a very worthy 7.6%. CXMT has chosen to compete in this very attractive market by deploying a very aggressive pricing policy. While large foreign manufacturers maximize their margins thanks to data centers and the rise of AI, Chinese manufacturers prioritize sourcing from local companies. This scenario allows the supply and prices of memory and NAND chips in China to remain relatively stable, remaining outside the strong premiums charged by the big three (Samsung, Micron and SK Hynix). This is the context in which the Chinese memory module brands Gloway and KingBank have recently announced new DDR5 modules that stand out for using SDRAM memory chips made in China. With a standard configuration of eight chips per module, these companies can produce 24 GB modules and group them into kits of two or four modules to achieve capacities of 48 GB or 96 GB, respectively. Chinese memory chips, particularly those from CXMT, have already begun to spread beyond China’s borders. Corsair has already integrated them into some kits of its Vengeance line, while HP and Dell have begun the process of homologating modules with CXMT chips for their products. This is good news for users, there is no doubt. Even so, we still don’t know if the use of CXMT DRAM will become widespread in response to AI-induced shortages. The market demands new players, wherever they come from, and if YMTC and CXMT are able to fill the gaps left by Samsung, Micron and SK Hynix, they are welcome. Image | Intel More information | Tom’s Hardware In Xataka | China needs to develop a new type of chips immune to US sanctions. And your scientists have just achieved it

China will build the world’s largest AI network with its own chips, although they are not enough

by

The Chinese Government is launching a project that seeks to invest $295 billion in five years in the deployment of a national network of data centers of artificial intelligence (AI). In the current scenario of confrontation in the technological field between the US and China, this plan is not surprising at all. However, we cannot ignore the fine print: at least 80% of the underlying technology, including AI chips, must come from domestic suppliers, such as Huawei or Cambricon. Early October 2024 the Administration sent Chinese AI companies were given a recommendation asking them to use chips produced in China as much as possible. Ten months later, this recommendation became a requirement. The Chinese government forced state-owned data centers across the country to use at least 50% Chinese integrated circuits in their servers. Now, as we have just seen, this figure rises to 80%. Be that as it may, this scenario clearly favors three companies: Huawei, Cambricon and Moore Threads. Nvidia and AMD no longer count for China Huawei invests more than 25 billion dollars annually in developing your hardware for AIso presumably it will not take long to match the performance of the GPUs produced by Nvidia or AMD. However, this company faces an enormous challenge that will probably prevent it from meeting the demand for AI chips from the Chinese market in the short term. In mid-June 2025 Jeffrey Kessler, the Under Secretary of Commerce for Industry and Security at the US Department of Commerce, made this statement in Congress: “Our assessment indicates that Huawei’s Ascend chip production capacity by 2025 will be 200,000 units or less, and we anticipate that most or all of that production will be delivered to companies within China.” This limitation has a specific technical basis: yield per wafer The integration technologies used by the Chinese semiconductor manufacturer SMIC to produce Huawei or Cambricon chips have a very wide room for improvement. The per-wafer performance of SMIC integration technologies has a very large room for improvement SMIC already has the capacity to manufacture 6nm integrated circuits, and will soon also be able to produce 5nm semiconductors, but is limited by the performance of the deep ultraviolet lithography equipment (UVP) that you have in your possession. It is commendable that SMIC and Huawei engineers have managed to refine their integrated circuit manufacturing processes enough to be able to produce 5, 6 and 7 nm chips with ASML’s UVP equipment, but a priori it is very unlikely that with these machines they will be able to go beyond 3 nm. And it is because the technique of multiple patterningwhich is what they are using, imposes important limitations. A note: this strategy broadly consists of transferring the pattern to the wafer in several passes with the purpose of increase resolution of the lithographic process. Its problem is that it usually has an upward impact on the cost of chips and a downward impact on production capacity. For Huawei, it is a big problem not to have the necessary technology to produce cutting-edge semiconductors comparable to those manufactured by Intel, TSMC or Samsung, which is why it is working on the development of its own equipment. extreme ultraviolet photolithography (EUV). This scenario has caused the Chinese semiconductor industry to question whether national hardware can maintain the pace demanded by the Government. SMIC co-CEO Zhao Haijun has warned that this hasty addition of capacity can leave data centers idlecomparing this situation to building highways before there is traffic. In addition, several Chinese executives in the chip sector have recognized separately that his country is five to ten years behind in the cutting edge in silicon for AI data centers. In the current situation, the decision has its logic. Another thing is whether China can comply with it. Image | Cambricon Technologies More information | Tom’s Hardware In Xataka | NVIDIA has to deal with the absolute distrust of several US legislators. Your plan in China is in danger In Xataka | The US wants to end Chinese AI chips sold abroad. And China knows how to defend itself

The world was tired of depending on TSMC to manufacture all its chips. This is what is causing Intel’s great resurrection

by

Who has seen you and who sees you, Intel. The legendary semiconductor firm seems to be leaving behind its painful journey through the desert, and the latest news points to a true resurrection. The signature has achieved a spectacular contract to manufacture three million Google TPUs, and Nvidia is also studying the possibility of use Intel 18A node for future multi-die GPU designs. This is spectacular news for the company. Promising future, at last. The agreement with Google’s cloud division is a huge boost for the chip manufacturing business (foundry) from Intel. This deal will see Intel produce millions of AI chips at its advanced 3-nanometer node. With it, the firm achieves a decisive step to compete with TSMC, which until now was the absolute reference for those who wanted to access advanced semiconductor manufacturing processes. There is another crucial geopolitical factor here: part of these chips will be produced in the US, which helps in the objective of not depending so much on Asian countries for this process. Flirting with Nvidia. But Nvidia also seems to be interested in Intel’s 18A photolithographic process. The company led by Jensen Huang is considering the use of this node for its future multi-die architectures for its GPUs. Nvidia has managed to become TSMC’s main customer, but this manufacturer cannot satisfy Nvidia’s demand, so this company is looking for plans B, and Intel is serving it one on a plate. The signature by the way, already bought 4% of Intel in September 2025, so it is the first interested in Intel doing well. The PowerVia revolution. There are two big technical arguments that are apparently convincing Google and Nvidia. The first, the transistors RibbonFET. The second, even more important, PowerVia technology. This system is a qualitative leap because it physically separates the power and signal lines from the transistors, which avoids bottlenecks and improves both performance and efficiency of the CPUs that use this technology. Chip sovereignty. This decision by Google and Nvidia’s plan respond in part to the pressures that the US government is doing—and boosting with its CHIPS Act— to recover technological sovereignty and avoid dependence on foreign countries. Both companies know that 90% of the planet’s advanced chips depend on that island called Taiwan, and taking advantage of Intel’s renewed capacity is a great opportunity for kill two birds with one stone. They reduce their dependence on TSMC, and comply with the demands of the US government. War makes strange allies. The current situation is unique, because it is causing companies that competed fiercely in the field of hardware (Intel and Nvidia) to now be forced to collaborate out of pure necessity. Intel needs clients of this type to demonstrate to investors that its division foundry can operate independently of its consumer processor or server division. And Google and Nvidia in turn need Intel to break manufacturing monopoly of semiconductors that TSMC had. Intel finally resurrects. The big winner of these agreements is Intel, which has gone through a really compromising stage but has for a year has not stopped growing. We can see it in its valuation on the stock market. A year ago its shares were trading at $20.68, and now they are trading at $107.04 and with these agreements that value may continue to improve. Good for Intel. Image | Intel In Xataka | Bill Gates has X-rayed Intel. And his diagnosis is overwhelmingly accurate.

“The demand for AI chips exceeds us, and will continue to do so for years”

by

CC Wei, the current president and CEO of TSMC, knows exactly what he has on his hands. Recently has communicated to shareholders that this Taiwanese company will not be able to fully meet the global demand for chip production. artificial intelligence (AI). And this situation will last for several years. The trigger for this scenario is the insatiable demand for AI chips from the data centers that are being built around the planet. Intel is setting up several cutting-edge integrated circuit production plants. Samsung, too. And TSMC is working on building new state-of-the-art factories in Taiwan, the US, Germany and Japan. Despite this effort, the future looks full of dark clouds. TSMC maintains its forecast It expects a 30% increase in sales this year, but it could sell more if its production infrastructure were able to absorb the current demand for AI semiconductors. Despite this scenario, the head of TSMC has anticipated that his company will not take advantage of this bottleneck to suddenly increase the price of wafers. And he will not do it for a compelling reason: he prefers to guarantee the stability of his business. What will most likely happen is that the cost will rise little by little, so that it can be absorbed by the market without triggering a fracture. If the AI ​​bubble does not burst, demand will continue to grow The factories that TSMC has in operation in Taiwan, and, above all, the plants that it is building in Hsinchu, Taichung and Kaohsiung, play a leading role in this company’s medium-term strategy. However, its most media project is its new facility in Arizona (USA). The plant that is already in operation has been producing 4 nm chips since 2024 in the N4 lithographic nodewhich belongs to the 5nm FinFET family. TSMC needs to reinforce its production infrastructure in the US in an attempt to meet demand This factory, known as Fab 21, made $514 million in profit last year according to Yeh Chun-Hsienthe minister of the National Development Council of Taiwan. This is not bad at all if we keep in mind that during the first year of operation the semiconductor plants They do not usually provide benefits. Even so, TSMC needs to reinforce its production infrastructure in the US in an attempt to meet the demand of its American customers, among which Nvidia, Apple, AMD and Qualcomm stand out. Their plan involves investing an additional 20 billion dollars in the expansion of Fab 21. In fact, this project is part of the expansion plan of 165 billion dollars that TSMC presented last year. If everything goes as planned, mass production of 3nm integrated circuits will begin in Arizona in 2027. But this is not all. And the purpose of this company is for this site to finally bring together no less than 12 factories, 4 advanced packaging centers and an R&D facility. This is the long-term expansion plan for the Chandler (Arizona) campus. What Wei hasn’t said, but the market has been reading between the lines for months, is that TSMC is in an extraordinarily comfortable position. You can choose who to manufacture, at what price and in what timeframe. Their clients have no real alternative in the short term. Intel tries to sneak into that gap with its 18A integration technologyand some reports argue that Apple has already reached a preliminary agreement with those of Lip-Bu Tan. But this, at best, is a story for 2027. Meanwhile, TSMC’s waiting list continues to grow. And CC Wei knows it perfectly. Image | TSMC More information | Tom’s Hardware In Xataka | Intel’s plan against an unattainable TSMC: beat Samsung and consolidate itself as the second largest chip manufacturer

Log In

Forgot password?

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections

Here you'll find all collections you've created before.