Weekly: Parallel profiles of Intel, Samsung
18.5 min read.
Highlights
US-China and chips. On politics and geopolitics, Chris McGuire of CFR writes about the Trump-Xi summit and necessity of AI dialogue. He also comments on the chip export controls and argues that loopholes should be closed. Separately, CSIS experts write on the impact of US tariffs on the AI data centre build out in the United States, demonstrating that they impose costs across the tech stack and risk their competitiveness relative to foreign actors. A third piece, a chart-driven article by the FT, shows that tech stocks, and especially semiconductor stocks, are pushing the entire US stock market up, obscuring the impact of the Iran War.
Intel and Samsung. Bloomberg’s Ian King profiles Intel CEO Lip-bu Tan, who has achieved many quick, high-profile wins since his appointment as CEO last year. Despite investor optimism and White House backing, Intel’s developments over the past year seem mostly external-facing and light on details (such as a deal with Tesla and SpaceX to build Musk’s Terafab). King says that a lot of change still needs to happen inside the firm and that some inside Intel could still see the firm’s manufacturing and design arms split.
There are analogues between Intel and Korea’s giant chipmaker, Samsung. Both are stalwart powerhouses of the semiconductor industry. Both have struggled in recent times but have seen rapid and promising turnarounds since. The Economist profiles Samsung, the world’s largest chipmaker.
SemiAnalysis on Cerebras, EDA. SemiAnalysis has two of their trademark deep dives out this week. One is an investigation into Cerebras in light of their IPO and another is a primer on the EDA industry, which is emerging as yet another AI bottleneck as semiconductor designs get increasingly complex.
Thanks for reading.
Table of Contents
Chris McGuire, “How Trump Should Approach AI Talks With China: Targeted Dialogue, Maximum Pressure,” CFR, 05/13/2026.
Evan Brown, Michael Gary, Kate Koren, and Philip Luck, “The Impact of Tariffs on the AI Data Center Buildout: Balancing Supply Chain Security and AI Infrastructure Leadership,” CSIS, 05/14/2026.
Clara Murray, “How AI mania is disguising big companies’ hit from Iran war — in charts,” FT, 05/11/2026.
Ian King, “Intel CEO Who Won Over Trump and Musk Now Needs a Breakthrough,” Bloomberg, 05/08/2026.
Asa Fitch, “AI’s Next Phase Plays Into TSMC’s Hands,” WSJ, 05/11/2026.
The Economist, “The strange Japanese companies minting money from AI,” The Economist, 05/14/2026.
The Economist, “Samsung has staged a stunning comeback,” The Economist, 05/14/2026.
Myron Xie, Jordan Nanos, Max Kan, et al., “Cerebras — Faster Tokens Please,” SemiAnalysis, 05/14/2026.
Gerald Wong, Dylan Patel, and Sravan Kundojjala, “The EDA Primer: From RTL to Silicon,” SemiAnalysis, 05/12/2026.
Shruti Mittal, “The Unresolved Challenges in U.S.–India Semiconductor Cooperation,” Carnegie Endowment for International Peace, 05/14/2026.
1.
Chris McGuire, “How Trump Should Approach AI Talks With China: Targeted Dialogue, Maximum Pressure,” CFR, 05/13/2026.
U.S. President Donald Trump and Chinese President Xi Jinping plan to discuss issues related to artificial intelligence (AI) when they meet in Beijing—as they should. AI is increasingly underpinning global economic growth, driving technological innovation, and reshaping world battlefields. Modern AI models are the most powerful cybersecurity and hacking weapons that have ever been created, and they are doubling in capability every four months. AI is simply too important to ignore.
If the United States significantly tightened export controls on China, it could expand the U.S. lead from eight months, to eighteen or twenty-four—an eternity in AI development. Chinese firms remain extremely dependent on U.S. computing power, which is the most critical input into AI development. China will only produce about 2 percent of the AI computing power of U.S. firms this year, and the computing power needs to develop and serve a leading AI model are increasing exponentially. U.S. export controls have materially slowed China’s AI development, but they contain significant loopholes that allow China to purchase U.S. AI chips, remotely access them via the cloud, smuggle them via third-countries, or use U.S. chipmaking technology to manufacture them. The presence of these loopholes is not an inevitability; it is a policy choice that can be changed.
Trump’s goal in Beijing should not be to reach an agreement with China on AI safety, but to create the conditions for such an agreement down the road. If the Trump administration does establish a dialogue with China on AI, it must set clear expectations with the Chinese that the dialogue will be narrowly focused on AI safety issues and not cover export controls. And simultaneously, any such dialogue must be coupled with a “maximum pressure” campaign that imposes robust export controls that close all existing loopholes to maximize the U.S. lead over China. Just as the United States and the Soviet Union never assisted each other’s nuclear weapons development programs, the United States and China should not assist the other’s efforts to develop advanced AI models.
2.
Evan Brown, Michael Gary, Kate Koren, and Philip Luck, “The Impact of Tariffs on the AI Data Center Buildout: Balancing Supply Chain Security and AI Infrastructure Leadership,” CSIS, 05/14/2026.
The Issue
The United States is on track to invest more than $2.7 trillion in data center infrastructure by 2030, with semiconductors representing approximately 54 cents of every dollar spent. The Trump administration has embraced two goals that are fundamentally in tension: an aggressive push to build out U.S. AI infrastructure, and broad use of Section 232 tariff authority to restrict the semiconductor and metal imports the buildout requires. A January 2025 proclamation that applied a 25 percent tariff on a specific subset of advanced semiconductors carved out imports for U.S. data center construction but left open the possibility of wider levies on all semiconductors and derivative products. A 100 percent tariff on all semiconductors and products containing them would likely impose an additional $1.4 trillion burden on the buildout. While such a maximalist approach to semiconductor tariffs is not the expected path for the administration, even more moderate tariff scenarios would function as a tax on the United States AI ambitions. This brief examines how semiconductor and metal tariffs interact with data center economics, assesses cost implications across multiple tariff scenarios, and offers policy recommendations to resolve the tension between supply chain security and AI infrastructure leadership.
Conclusion
The United States has a genuine national security interest in reducing its dependence on semiconductor imports from geopolitical adversaries. However, the administration’s ambition to lead global AI infrastructure development depends on the very global supply chains that its tariff agenda is designed to reduce. Resolving this tension does not require abandoning either goal, but it requires precision in how tariffs are designed and applied and complementary policies to support domestic growth. The administration’s restraint on tariffs affecting AI data center inputs to date suggests a recognition that indiscriminate measures are at odds with buildout goals. However, Section 232 tariffs that fall outside these carveouts will still impose real costs on the broader technology ecosystem. Sectors such as domestic research and development and electronics manufacturing, if exposed to higher input costs, risk being disadvantaged relative to foreign competitors. A national security–driven tariff strategy should therefore take a holistic view of the technology stack and avoid shifting vulnerabilities to less visible segments of U.S. competitiveness.
3.
Clara Murray, “How AI mania is disguising big companies’ hit from Iran war — in charts,” FT, 05/11/2026.
Airlines are cutting flights, consumers are tightening their belts and companies are warning of price rises as the Iran war upends global business. Yet, a little over two months after US-Israeli attacks on Iran sparked the conflict, the world’s biggest companies have added more than $5.4tn in value, an increase of 4.2 per cent.
Tech groups have held global equity markets aloft as investors hunt for the sectors least exposed to the Middle East conflict. Huge share price gains for chipmakers such as Intel and Taiwan Semiconductor Manufacturing Company and a blistering recovery in Big Tech valuations, helped by bumper first-quarter profits, have outweighed the falls in other sectors.
Investors were turning back to tech at a time of “extreme macro uncertainty”, said Luca Paolini, chief strategist at Pictet Asset Management, lured back to the “certainty of earnings delivery” in US tech. “Post-ceasefire, it was all about AI again.”
The combined value of semiconductor companies with a market valuation of more than $10bn has risen 26 per cent, or $3.7tn, since the war began in late February.
Companies with a smaller footprint in the Middle East — such as Norway’s Equinor, which is up 24 per cent — have performed the best, followed by companies with large trading desks that have been able to take advantage of the volatility in the market. BP and TotalEnergies are up 14 per cent and 16 per cent respectively.
The worst performers have been companies whose oil and gas production in the Gulf has been shut down or hit by missiles. ExxonMobil and Shell are both facing multibillion-dollar bills to repair the damage at Qatar’s Ras Laffan industrial city. Exxon’s value is down 4 per cent, or $28bn, since the conflict began.
4.
Ian King, “Intel CEO Who Won Over Trump and Musk Now Needs a Breakthrough,” Bloomberg, 05/08/2026.
After Lip-Bu Tan became chief executive officer of Intel Corp. in March of last year, the struggling company’s shares went nowhere for seven months while the chipmaker was getting trounced in the market for artificial intelligence.
But after forging ties to the world’s biggest tech titans — and winning over US President Donald Trump — Tan is kicking off year two on a decidedly higher note. Apple Inc. and Tesla Inc. are showing interest in the company’s manufacturing. The processors it makes are back in demand, and budding optimism that Intel will finally start to benefit from the AI boom has sent its stock to a record.
Before Tan can deliver on shareholders’ rising expectations, he has to make changes within the 57-year-old company that was formerly a leader in semiconductor manufacturing. Since becoming CEO, Tan has spent far more time outside the company than inside, and has not widely explained his specific plan to fix products and manufacturing to employees, according to more than a dozen current and former staff, who were not authorized to speak publicly.
The fundamental issues remain, they said: Intel needs products that can win back lost share, and manufacturing that’s so good, even rivals will have to give it billions of dollars in orders. Neither of those is a given.
Some at Intel still believe the company should be broken up, to split manufacturing and product design and accelerate progress. Tan said that can’t happen anytime soon and that there are advantages to keeping the two units tied together. Over time, he could see an arrangement along the lines of EMC Corp.’s former operation of VMware as a majority-owned subsidiary.
5.
Asa Fitch, “AI’s Next Phase Plays Into TSMC’s Hands,” WSJ, 05/11/2026.
Big tech companies are tripping over themselves to get as many chips as possible to secure AI supremacy. So much so that the world’s ability to make what they want is coming under growing strain.
Microsoft, Meta Platforms, Alphabet and Amazon.com collectively plan capital spending of $725 billion this year, much of it on artificial-intelligence chips. That bodes well for the makers of those chips, particularly the largest contract manufacturer, Taiwan Semiconductor Manufacturing Co.
The spotlight for chip investors has largely been elsewhere lately. Markets have recently blessed memory-chip manufacturers whose sales and prices have skyrocketed. Intel and Advanced Micro Devices are also celebrating a shift within AI toward using autonomous agents where their central processors are increasingly important.
There is an argument, though, that nobody in the chip world is better positioned to reap the rewards of this new phase than TSMC. And despite all that, its stock doesn’t look expensive.
While the company doesn’t make memory chips, it is the go-to manufacturer for just about everything else—including Nvidia’s market-leading AI chips and Apple’s smartphone chips. It has already seen sales and profits soar over the past few years.
The strength of TSMC’s position is evident in the recent expansion of its gross margins. Those margins get fatter when sales grow faster than expenses, something that happens for TSMC when its factories are running closer to full blast.
High capacity utilization offsets the large fixed costs of maintaining chip factories. With demand soaring, the company’s gross margins climbed to around 66% in the first quarter, from roughly 59% a year before.
TSMC and the chip industry aren’t immune to cycles, even if the current uptrend has been unusually strong. But TSMC’s trajectory suggests a lot of reward ahead with relatively little risk for a company at the heart of the AI boom.
6.
The Economist, “The strange Japanese companies minting money from AI,” The Economist, 05/14/2026.
Ajinomoto has spent well over a century supplying monosodium glutamate (MSG), a chemical that gives food an umami kick. Now another of the Japanese seasoning giant’s products is whetting investors’ appetites. Ajinomoto Build-up Film (ABF) is a material used to insulate artificial-intelligence processors from circuit boards. It was originally made from by-products of MSG manufacturing. Ajinomoto controls more than 95% of the market. Booming demand for AI chips has made the film scarce, pushing Ajinomoto’s share price up by 65% since the start of the year, around three times the gain in Japan’s benchmark Nikkei index.
Toto, another century-old Japanese firm, has lately enjoyed an equally improbable flush of prosperity. Best known as the world’s largest toilet-maker, it has found a profitable seat in the semiconductor supply chain. The firm is a leading producer of electrostatic chucks: ceramic plates that hold silicon wafers firmly in place while memory chips are etched. Toto’s operating profit from advanced ceramics now accounts for more than half its total.
The AI frenzy has produced obvious winners in semiconductors: American chip designers, South Korean memory-makers, Taiwanese foundries. Japan has its equivalents, with giants such as Tokyo Electron and Advantest that make the sophisticated equipment used to fabricate and test chips.
But like Ajinomoto and Toto, many of the country’s AI winners are in less flashy trades. Hoya, a health-care company that makes spectacles and contact lenses, is a leading supplier of photomask blanks: transparent plates coated with light-sensitive material that lithography tools use to etch chip designs onto silicon wafers. Sakura, a stationery brand, has adapted technology once used for coloured pencils to spot defects in chip-manufacturing processes. Nitto Boseki (or Nittobo), which began life as a textile company in 1923, is today the sole supplier of “T-glass,” an ultra-thin glass fibre essential for packaging AI chips.
7.
The Economist, “Samsung has staged a stunning comeback,” The Economist, 05/14/2026.
Not long ago Samsung Electronics was in the doldrums. In 2024 the South Korean giant apologised for failing to maintain “technological competitiveness” and “falling short of the market’s expectations”. It has no need for contrition these days. This month its market value, which has soared by 400% in the past year, hit $1trn for the first time, propelled by furious spending on artificial-intelligence infrastructure. In the first quarter of 2026 its operating profit rose to 57trn won ($38bn), more than eight times as much as a year before. Analysts expect profits to keep rising at a blistering pace, thanks in particular to the seemingly insatiable demand for its advanced memory chips.
Samsung Electronics manages a wide portfolio of products, making everything from fridges to phones. Increasingly, however, its business centres on chipmaking. Semiconductors accounted for 61% of sales and 94% of operating profits in the first quarter. It is one of just three firms capable of making at scale the memory chips needed for AI, alongside SK Hynix, a South Korean rival, and Micron, an American one. The number of memory chips Samsung sold in the first quarter was up by about 20% on the preceding three months, but the average selling price rose by 90%. The firm boasts that memory-starved buyers are approaching them to demand long-term purchase agreements. It foresees the shortage lasting well into next year.
Samsung is expanding capacity, but relatively slowly. A new facility will start mass-producing chips for sale later this year, and the firm will begin building another, which will cost it some $55bn, in July. Yet that factory, known as P5 Fab 2, will not be ready until 2030. And although capital expenditure is set to rise by 55% this year, according to Daniel Kim of Macquarie, a bank, it is falling as a share of revenue.
Samsung’s management, perhaps unsurprisingly, continues to display caution. Chipmaking factories are enormously expensive and take years to complete, resulting in a historical cycle of booms and busts. The company will not want a repeat of the last flash-memory boom-bust cycle, when it overbuilt capacity as demand surged, then saw its operating profit fall by half in 2019, notes Jukan Choe of Citrini Research, a firm of analysts.
A coalition of Samsung unions is also demanding that 15% of the memory division’s profits be distributed to workers, similar to an arrangement already in place at SK Hynix. They are threatening a multi-week strike beginning on May 21st, which would cost Samsung some 30trn won. (Ironically, union members also lambasted Samsung for failing to capitalise on the AI boom during a strike in 2024.) The problems of success may feel like a welcome change for Samsung’s bosses. But they are problems nonetheless.
8.
Myron Xie, Jordan Nanos, Max Kan, et al., “Cerebras — Faster Tokens Please,” SemiAnalysis, 05/14/2026.
With the arrival of fast tokens on the mainstage and a 750MW compute deal with OpenAI notched, Cerebras is feeling ready for the scrutiny of public markets. Up until just 6 months ago, we felt that the Wafer Scale Engine, despite its bold innovations, had some technical weaknesses that were too hard to cover up. Thus, the continued popularity of HBM-based accelerators such as GPU and TPU. The strengths of Cerebras (namely: speed), have been overlooked for years in favor of total throughput. But now, with frontier labs releasing fast, priority, standard and batch tiers of the same model weights, the world has revealed their preference for fast tokens with their wallets. This brings Cerebras’s strengths to the fore and is the key reason why OpenAI is willing to fork over tens of billions of dollars for Cerebras compute.
Demand is so strong it’s making everyone look good.
Today, on the verge of Cerebras’s IPO, and because we love the wafer, we are shipping an article that is as long as 4 normal articles. Inside, we will dive deep on:
1. Fast inference
2. WSE-3, Cerebras’ unique wafer-scale chip
3. CS-3, Cerebras’ system, with its unique architecture
4. Provide a BOM cost analysis
5. Explain when and how the wafer wins for fast inference
6. Describe some of the wafer’s limitations, showing tradeoffs
9.
Gerald Wong, Dylan Patel, and Sravan Kundojjala, “The EDA Primer: From RTL to Silicon,” SemiAnalysis, 05/12/2026.
With this trifecta of increasing chip complexity, compressed design timelines and a shortage of engineers, a massive bottleneck has formed at the design stage. The latest AMD MI455X packs 320 billion transistors across 12 logic dies on 2nm and 3nm processes with advanced Hybrid Bonding 3D die stacking, HBM4 memory integration and high speed 224G SerDes. Designing something at this scale is not a matter of hiring more engineers or buying more verification servers. It tests a company’s tooling, methodology, and human capital organization as to whether the design succeeds or fails.
After spending hundreds of millions of dollars on a new SoC design, there is no guarantee the chip will work. Multiple steppings are usually required that need new mask sets, with A0 rarely going into production. When a single advanced mask set costs tens of millions of dollars, every respin is a gut punch to the balance sheet. Furthermore, it adds months to the schedule for high volume production start.
As designs get more complex, testing is becoming more important to ensure all modules within a chip are interoperable and locally sound. Verification, the process of proving a design does exactly what it should before committing it to silicon, now consumes up to 70% of total project effort, depending on the design. Verification engineers are the fastest-growing job category in chip development, and the industry still cannot hire them fast enough.
While chip complexity grows at roughly 50% per year, driven by new nodes and larger SoCs, design productivity improves only about 20% each year. This design productivity gap means every new generation of silicon demands exponentially more engineering effort, more compute, and more sophisticated automation.
The semiconductor industry’s ability to keep building more powerful chips depends not on physics or lithography alone, but on EDA (Electronic Design Automation) software. These tools effectively translate human intent into manufacturable silicon. Without EDA, no chip designed after the mid-1980s would exist.
This primer is your guide to EDA in the semiconductor industry. In this first part, we will walk the entire journey from RTL (Register Transfer Level) code, the high-level hardware description language that engineers actually write, all the way to manufactured, packaged silicon. We will name the tools, explain the tradeoffs, and show why EDA is one of the most consequential and underappreciated sectors in technology.
In part 2, our EDA Market Primer dives deep into the business of EDA, profiling the major companies (Synopsys, Cadence, Siemens) and their revenue and business models. We provide comprehensive market analysis and monitoring the Chinese EDA effort, as well as IP licensing and outsourcing to design partners and the transition to Customer Owned Tooling (COT) with hyperscaler ASIC designs.
Part 3 then assesses how AI is disrupting the EDA industry, covering the full gamut from startups and engineer dashboards to agentic chip design flows from NVIDIA and the big three. The concept of using AI accelerators to create superhuman designs that go into future AI accelerators is the most exciting development that our industry has seen in decades. Stay tuned as we cover the incoming revolution in chip design.
10.
Shruti Mittal, “The Unresolved Challenges in U.S.–India Semiconductor Cooperation,” Carnegie Endowment for International Peace, 05/14/2026.
In February 2025, Indian prime minister Narendra Modi and U.S. president Donald Trump launched the Transforming the Relationship Utilizing Strategic Technology (TRUST) initiative, naming semiconductors as a core area for building trusted and resilient supply chains. Recently, India was invited to join the Pax Silica initiative launched by the U.S. government to coordinate artificial intelligence (AI) and semiconductor supply chain security amongst allies and trusted partners.
The U.S.–India semiconductor cooperation story is well-stocked with top-level strategic intent. What remains unresolved, however, are some underlying challenges that will determine whether the cooperation actually functions. Three such friction points stand out. First, the state of export control regimes on both sides. Second, the absence of a clearly articulated economic case for U.S.–India semiconductor cooperation. Third, a newly articulated anxiety on the U.S. side about India becoming a China-like strategic risk.
A more productive framing perhaps lies less in asking whether India can become another China and more in asking what kind of governance architecture could make the U.S.–India partnership robust enough to put such questions to rest. This architecture could include a shared framework for jointly generated intellectual property at both the pre-competitive and competitive stages, as well as mechanisms to ensure bilateral transparency on end-use.
While such friction points in U.S.–India semiconductor cooperation are real, none are intractable. Resolving them, however, requires quieter, technical work of alignment and political will on both sides.
–