I’ll be in New York from 7 to 14 September to meet the Funda team there and some of you. Please reach out if you’d like to meet.

We have been more cautious about the market in recent weeks, but as we note below, there are significant bright spots:

• Nand pricing is rebounding.

• The Koreans are returning capital.

• Alibaba made numerous helpful comments on ROI and AI investment payback.

Finally, Funda is working on a study of enterprise adoption of AI. This is going to be important as we head into the tail end of this year.

NAND

BofA reporting that NAND wafer prices “rebounded sharply” in early August.

Hearing this being attributed to Vera Rubin mass production ramping up.

Trendforce post on X on the same new has a bearish tilt, saying “overall transaction momentum [in consumer] continues to wither”:

The spot market exhibited a visible vertical consolidation recently still attributed to the overall sluggish consumer demand, and overall transaction momentum continues to wither without sizable improvement to the prospect of the end sector, despite some of the suppliers willing to compromise on prices. Spot prices of 512Gb TLC wafers rose by 0.39% this week (Aug.17), arriving at US$21.208.

An earlier post from 30 July describes Agentic AI as a force that could “significantly increase demand for high-speed SSDs, helping absorb incremental supply” and bring it into “equilibrium”.

Trendforce does not appear to treat AI driven demand as the base case.

Hynix: MS says on track to complete buyback in 37 days

Hynix registered for 650k shrs buyback (12% ADV) today compared to 650k (100% filled) yesterday or 2.7% of the total done i.e. on track to complete the full size 24m shares in 37 days (by Oct 15) at this pace. Separately, US ADR +4.4% (premium 33% or flat DoD).

Samsung: Board meeting to approve shareholder return program expected today (Nikkei)

South Korean semiconductor giant Samsung Electronics is expected to unveil a shareholder return program worth a total of 100 trillion won ($72 billion), Nikkei has learned.

According to sources, Samsung is expected to announce its plan following a board meeting on Friday. The program would mark the largest-ever shareholder return in South Korea.

Alibaba: Comments on ROI and Payback

Some helpful comments from Alibaba for the AI trade:

Read more

Dr He confirming what we’ve flagged for several months now - memory and ABF are the two most severe shortages in semis:

One of the immediate challenges I see right now, and this is true for the whole industry. In the last couple of years, and for the next few years, the whole industry is facing tremendous shortage, not only on memory, but also on something called ABF substrate. Your sourcing team is knowingly taking risk of multiple sourcing because of the capacity shortage.

Ibiden is a key beneficiary. AT&S is a relatively undervalued player who took more space in Kulim last week to support a growing workforce. AT&S is one of our favourite ideas.

Singapore raises GDP growth forecast to 4.5 to 5.5% (PR)

Ministry of Trade and Industry:

[A]cceleration in global AI-related capital expenditure should boost growth in the electronics and precision engineering clusters of the manufacturing sector. Notably, demand for AI-related semiconductors (e.g., networking and memory chips) from the data centre end-market is expected to remain strong amidst the rapid rollout of agentic AI across industries, while capacity expansions by semiconductor firms will raise demand for semiconductor equipment

In the same week:

• AEM, AEHR’s bigger cousin, reported net profit up 876% y/y. AEM has been getting more orders from AMD and Intel in past 2 months, and is likely struggling to fulfil AMD’s demand.

• Metasurface reported its order book grew 2.6x.

YMTC third in shipments but fifth in shipment value (Counterpoint; Nikkei)

As we reposted in early August, only 12% of 73 investors surveyed by MS thought YMTC was the most important factor for Kioxia going forward.

That is not to say YMTC is not important, its just important to have a measured take on their impact on the eSSD market.

Counterpoint:

The global NAND market in Q2 2026 was defined by the migration of AI workloads from training to inference, which pushed enterprise SSDs to 48% of total bits shipped and left consumer supply short. The resulting supply tightness pushed industry revenue to record levels, rising fivefold from Q2 2025.

Nikkei:

YMTC is a semiconductor manufacturer supported by the Chinese government, and is expanding its sales of consumer electronics such as smartphones and personal computers. The company’s shipments increased by 22% compared to the same period last year.

However, in terms of shipment value share for the April-June quarter, YMTC ranked 5th, falling behind Kioxia. This is because Kioxia and others sell advanced products for artificial intelligence (AI) servers, which have high unit prices and high profitability. YMTC has not been able to fully penetrate the server market and is content with 5th place in terms of shipment value.

Counterpoint:

The Chinese vendor plans to shift its mix further towards eSSDs in the second half of the year to cement its third-place position globally, supported by growing avenues for capital support. The same dynamic now frames the industry: with servers set to take most NAND bits, profitability through 2027 will be decided less by who ships the most bits, but who ships the right mix.

JPM 9 August memory note:

On the NAND side, our research indicates YMTC’s bit/wafer is largely on par with the leading memory makers indicating limited technology gap. We expect YMTC’s capacity/bit share to account for ~16% of the global supply by 2028E. This puts the risk of oversupply more in NAND vs. DRAM, on a relative basis. We expect the leading NAND makers to focus on manfuacturing high-end NAND (SLD based NAND or HBF) to widen the gap. Given the existence of US regulations, we believe YMTC’s NAND will likely be for local uses. Recent news stating that YMTC is considering to manufacture DRAM (link) is a positive for the overall NAND S-D, in our view.

What local uses? Apart from consumer electronics, China has its own DC buildout. Given that this newsletter started with coverage of the beneficiaries of the DC boom in Johor, it is striking to now see Mongolia dwarf that market.

Inner Mongolia: 3x the DC pipeline of Johor

UBS out with a note this week reporting that Ulanqab, Mongolia has received 12.5GW capacity commitment from data center operators and internet/AI companies as of Jun 2026, implying 10x+ capacity expansion from its c.1.2GW live capacity as of 2025.

This is three times the total upcoming data center capacity in Johor, Malaysia.

UBS sees GW-scale projects in Ulanqab underway from 3 of the top 4 data center operators in China, namely GDS, VNET and Chindata.

UBS also cites projects from Envision, ZDATA, Centrin, and self-built campuses Kuaishou and UCloud, ByteDance, Z.AI and other leading internet/AI companies

Apple: Chinese suppliers might help with supply but not necessarily pricing

On its latest earnings call, throwing cold water on the idea that CXMT and YMTC will price too aggressively.

In terms of the sources of supply, primarily the DRAM market has 3 suppliers. And obviously, if there were more suppliers, that would be good, and it would help us on the supply side and perhaps the pricing side. It’s unclear on the pricing side, but it could help on the supply side. And so we’re evaluating all options.

Temasek denies new interest in Samsung and Hynix, already has position (BT)

Business Times:

Temasek said in response to queries from Bloomberg News that it did not seek advice from the Korean government on the timing of investments in either SK Hynix or Samsung, and that it first invested in both companies more than two years ago.

Nikkei:

In the near term, Temasek noted “compelling opportunities” in AI, identifying five focus areas that it plans to dive deeper into: energy and data centers, semiconductors, cloud services, foundation models, and related software and applications.

Temasek has already taken bets in U.S. powerhouses that are at the forefront of the tech trend, such as ChatGPT creator OpenAI and Anthropic, which pioneered the Claude platform, in addition to semiconductor player Nvidia.

In South Korea, Temasek has positions in memory chip giants SK Hynix and Samsung Electronics, both of which have caused the country’s stock market to surge this year but have also been drivers of volatility as investor sentiments made short-term swings.

Intel: LBT podcast

'I used to [not] invest in memory because it’s kind of commodity business, right? But now [it has] become different. There’s a lot of new technology come out. So we are kind of looking… one of my pet projects is looking [at] some of the memory new architecture.

And I think you just saw the news — I hire my good friend Seok-Hee Lee. He used to run SK Hynix, right? So you kind of know something that I’m thinking about. We’re not ready to unfold it.'

Discounted exposure to SiTime

Given the (short-lived) success of the Kingboard Holdings idea, here is another hold co idea for value investors.

SiTime is the leader in micro-electro-mechanical system (MEMS) oscillators which have been replacing quartz oscillators in a variety of applications such as computing, consumer, networking, communications, automotive and industrial systems.

MEMS win on durability, shock resistance, and rapid configurability, while quartz crystals generally lead in pure phase noise, low jitter, and ultra-low power consumption. It’s a relatively small market which has been growing rapidly thanks to AI and datacenters, with the serviceable available market (SAM) quadrupling between 2019 and 2027, a CAGR of 27%.

Read more

In July, Sandisk and Kioxia drew down dramatically, more than six months earlier than when NAND prices are widely thought to peak (~H2 2027). This was largely attributed to some investors being forced to sell significant holdings, as well as the “second derivative”, the rate of change of NAND price increase, slowing down.

Since then, Jukan received a bit of heat for this tweet which was viewed 3.6 million times. Those angry about Jukan’s views fail to appreciate that being bullish about memory market fundamentals does necessarily mean being bullish about the price action in the short or medium term.

This is why we can, without contradiction, be bullish about s/d dynamics while also have reservations about pressure on the stock because of a turnover in the shareholder base and because it will take some time for people to persuaded that “this time is different”.

In this note, we summarise one of Dr Xiaoyu Ma’s presentations at the Future of Memory and Storage Convention 2026.

Dr Xiaoyu Ma is a Senior Software Engineer at Deepmind. He created the concept of high bandwidth flash with David Patterson in 2023, and has published numerous important papers.

Dr Ma explains the demand vectors driving the LLM inference / memory crisis. We also flag some honest questions about NAND arising from his arguments.

Subscribe now

First, Dr Ma’s explanation of the basic flow of LLM inference.

There are two stages, prefill and decode. Prefill is the stage where you process all the input tokens (the basic units of text or data) all at once.

This can be hundreds, thousands, or even up to millions of tokens all at once. Prefill is inherently parallel, so it’s compute-bound and runs great on TPUs and GPUs.

The second decode stage, where we generate tokens step by step, one token at a time, is problematic. This sequential step is called autoregressive, and that's the fundamental reason why the decode phase is memory-hungry.

There’s a data structure called KV cache that connects the two stages. The size of the KV cache is proportional to the input sequence length plus the output sequence length. While the auto-regressive decode stage creates a memory bandwidth challenge, the KV cache creates both a bandwidth and capacity challenge.

These challenges are made a lot of more difficult by the six trends below.

First, Mixture of Experts (MoE). Transformers began eight years ago as dense architectures; the industry has since shifted to sparse MoE, moving from low tens of experts to hundreds or even thousands in today’s state-of-the-art models.

The consequence is a blow-up in the size of model weights.

Editor’s note: Model weights currently sit in HBM, so this point is not relevant to SSDs. However, if HBF is successfully commercialised there is a good argument for storing model weights on HBF. HBF could tighten supply for NAND in the same way that HBM has tightened supply for DRAM. We’ll park this argument for now until we have better visibility on HBF - keep a lookout for the Sandisk investor day on 13 August.

Second, multi-modality.

Models have moved from text generation to audio and video generation, and per-token storage scales accordingly: a text token is a couple of bytes, an audio token roughly a thousand bytes (~1,000x text), and video tokens hundreds of bytes up to megabytes — another hundred-to-thousand-fold increase. More bytes per token means a proportionally larger KV cache for the same sequence length.

Editors note: This is indeed a well understood bull case for NAND demand. We are also in the very early innings of this playing out. See tweet below.

Third, reasoning.

Read more

It was quieter than usual here last week - my travel and gearing up for the Future of Memory and Storage conference coincided with a very unusual week in the markets. I think the further drawdown was anticipated, as evidenced by the numerous comments in the chat about how we needed to go lower, though the manner and time of the slide ending was not.

Looking ahead, there are many hot topics that will benefit from a discussion with the big memory players and other memory analysts next week.

We will also get to meet some smaller names like Macronix and Penguin Solutions.

There will obviously be some notes circulated next week but I am also interested in how the Funda platform will draw on everything we learn at these conferences going forward, whether you specifically query it about the conference or not.

SK Hynix / Samsung Electronics: -7/8%

Nvidia has been making Rubin Ultra less powerful than promised - when first shown off in 2025, it was supposed to have a big 4-chip package with 1TB of memory.

That's been cut back step by step, and following two weeks of rumours, it appears to ends up with just 192GB — actually less memory than the standard Rubin it is meant to replace, though the memory runs slightly faster, and 576 GPUs can be linked together instead of 72.

I have been hearing debate about whether this is demand destruction due to price increases or supply constraints, and whether HBM despec causes DRAM price to dip because more DRAM wafers are freed up.

It appears to have at least spooked the market. SemiAnalysis also has a note out highlighting that spend is being reallocated from HBM to interconnect, to connect the 576 GPUs.

Kioxia: MS survey of 73 investors

Interesting results:

Read more

These comments were apparently made a few times over the course of the last two weeks. The note summarising the comments is reproduced below:

Shawn Kim bearish take on memory

NAND

-Module makers - inventory 13 weeks, covid peak was 15 weeks

“Demand evaporated”

- Rate of change of price hikes slowing to 5% qoq in Q4 - less of a need to pull forward, opportunity cost of pulling in demand much lower

- CSPs say they have healthy inventory - one of the CSP Tencent said covered for 90% - NAND spot price dropping last 2m - Think demand and supply converging

-CXMT supply ramping up quickly

- DRAM - if NAND goes down, sell DRAM also

- HBM - capping at +40% (say based on Trend force)

- Joe Moore checks why different? - Joe checks with US CSPs, US too bullish on AI

Funda argues that Mr Kim draws from information in the Chinese market, but it is the US NAND market that is driving pricing.

Mr Kim’s view also runs counter to the sentiment at a small industry event I attended yesterday, where I spoke with folks in South East Asia who deeply involved in standing up data centers as quickly as they can:

Read more

We’re committed to presenting a balanced view, even if it comes at the expense of clicks. As we entered the second half of the year, my conversations with experts and investors surfaced questions that were indicative of a more cautious posture. In particular, there was a growing concern that frontier labs may not be as good a business as advertised, coupled with uncertainty about the implications for the AI buildout.

We then saw more model competition in the form of Grok 4.5, a surge of investor interest in hyperscaler capex plans, followed by concerns about memory costs being too high.

Funda quickly addressed the Meta Cloud concerns, but it seemed to dominate the conversations with clients for an entire week.

Funda’s view on Grok catching up is that it is bullish for compute, and the firm continues to argue that fundamentals are unchanged.

I have had the benefit on sitting in on some calls and reviewing transcripts of others, and as recently as this week we have an expert reiterating memory tightness out to 2028, coupled with an expectation that any retreat in NAND pricing thereafter would be modest. We’re looking forward to making this repository of information more powerful through the new search function. In theory, you should be able to ask any question of the Funda platform and get a view informed by Funda’s proprietary data.

TSMC: Committing 100b to Arizona and raising capex guidance

TSMC - the most prudent and well informed market participant - raised capex guidance.

The reaction to relatively good earnings from TSMC and ASML suggests, and this is my personal view, that we need to go lower and that the next checkpoint will be just before Google’s earnings next week.

When asked to break down capex, CC Wei called out testers:

We try very hard to make sure that our CapEx number is correct, but with the flexibility between the front-end and the back-end. Sometimes we have a bottleneck. So we put more money to buy the bottleneck tools.

For long term, I mean, that's the back end, is about 10 to 20%… actually, I'm very honest to tell you that as time goes by, some of the customers' product need more tester. You cannot believe that. I mean, so the tester in shortage, so we have to put more CapEx in the tester or in the packaging or in other areas.

…

the most important reason [for capex guidance increase] is because of the demand continued to increase, and we feel the pressure from the customer to drive TSMC, not drive actually, to cooperate with TSMC for the capacity increase. That's one of the major reasons. The second reason is inflation. Now we buy the tools with inflation price, okay?

Xi promises AI for all (BBG)

China continues to paint itself as the mirror image of the United States, capitalizing on the recent Fable export controls.

The US and China appear to be running two different AI races, and the uncomfortable question this raises is whether the US should continue inflating its training costs to train next generation models when good enough models can evidently be trained under the constraints faced by the Chinese.

Google Gemini Launch Delayed (BBG)

Some dirty laundry being aired by 10 current and former employees in this article.

…encouraging leadership of every department to move in the same direction is like trying to boil an ocean, one ex-employee said. When mandates shift or efforts end up duplicated in multiple departments, it gets even more difficult to maintain a cohesive strategy, current and former employees said. It’s also a challenge for any one offering to get the resources it would need to succeed, and to gain traction in the market, they said.

Montage Technology positive pre (PR)

In the first half of 2026, the Company expects to achieve a substantial increase in operating results, primarily benefit from the AI industry trend and strong market demand. On the one hand, as the penetration of DDR5 products rose with continuous iterations of their sub-generation, shipments of the Company’s DDR5 RCD chips have increased significantly, with the portion of shipments of Gen3 RCD chips and Gen4 RCD chips further rising. On the other hand, revenue from new interconnect products including MRCD/MDB, PCIe Retimer, CKD and CXLMXC chips has grown significantly. In the first half of 2026, revenue from interconnect chips of the Company reached approximately RMB3,111 million, representing an increase of around 26.4% period-to-period, in particular the revenue from interconnect chips in the second quarter reached approximately RMB1,694 million, representing an increase of around 28.2% period-to-period and of around 19.5% quarter-on-quarter.

Highlights from the chat this week

Tsmc/ASML great earnings, terrible reaction. All of the data is showing that this factor unwind is historical about -30% (varies by broker) in 17 trading days. Typically, this drawdown is -22% over 33 days with the worst being a -40 in 2021.

I think what caught ppl off guard was the lack of a fundamental reason for the quick drawdown

Lack of long only buying exacerbating the unwind:

Read more

Despite those [Meta Cloud] efforts, Meta is still hungry for even more computing power, the spokesperson said. It is still moving forward with plans for expensive new data centers and recently inked major computing deals with CoreWeave Inc., Alphabet Inc.’s Google and Oracle Corp., among others.

In fact, just days before the original Bloomberg article on Meta Cloud, Meta leadership has been making under-the-radar comments about the need for capacity and perceived ROI.

Read more

In news that appears positive for sentiment on Anthropic and OAI:

Chinese authorities have held meetings with top tech firms over the past month ​about potentially restricting overseas access to China’s most advanced AI models, including those yet to be released, three people familiar …

Read more

Remarkable 23% swing from the low today.

Kioxia announced the production of 10th Gen 3D Flash at Kitakami 2 today as well as the shipping of samples to customers.

I understand UBS also published a note forecasting 28%/11% ASP increase in Q3/Q4 ‘26.

Perhaps more important was the slew of notes defending Meta Cloud, which seemed to outweigh the Reute…

Read more

Memory weakness?

This is thought to be a reference to Jerry Tworek at Core Automation.

Memory: Customary cooling announcement weighs on the legacy memory names, including Macronix.

Digitimes:

GigaDevice Technology Group Co. said in a risk alert filing on June 30, 2026, that the memory chip industry has historically shown clear cyclical volatility. It added that current product prices are already at historically high levels, and that the trend of continued steep increases is unlikely to last. As supply and demand move back toward balance, major product prices could fall sharply, weighing on its storage business margins and overall profitability.

China Times.

Memory: NAND overweight at the new SA / TEMA DISK ETF (PR)

MLCC: Yageo price increases

To add to Jukan’s tweet: the price increase will extend beyond distributors to include direct sales customers such as ODMs and EMS, which is expected to significantly boost future revenue and profits (Source)

Nittobo (Nikkei Asia)

Sumco +17%

Think we can attribute strength in silicon wafer names to a sell-side note today…

Read more

Exponential View: AI demand is “revenue-validated”.

• Generative AI ecosystem revenue has already surpassed $175 billion annualized.

• It is just about covering the infrastructure bill.

• The investment debate is not ‘is there demand?’ but ‘can falling token prices drive enough volume and margin to earn back the AI CapEx cycle?’

  • Token volume growing 14x YoY.

Memory / Semicap:

• We are expecting more details of capex plans from Samsung and SK Hynix this week.

• On the All-in Podcast out over the weekend, Gavin Baker said he and Elon think had the DRAM / HBM bottleneck is the most important bottleneck:

The bottleneck that matters is DRAM. This is the most important bottleneck simply because memory capacity and bandwidth are foundational to the performance of every AI model. So, this is the most important bottleneck. Elon is focusing the Terrafab on memory because he sees it as the most important bottleneck.

Not lasers, not capacitors, not power power supply semiconductors, not NAND, not HDDs. DRAM.

Both points bullish for semicap.

Frontier Models and Open Source Models

Separately, Gavin on All-In Podcast on Anthropic:

I think Anthropic is worth $3 trillion today.

They're going to end this year well over 100 billion.

What's the 28 number? Is it 200? Is it 300 billion? It's probably not going to trade at 10 times that number. And it will be very profitable at that scale because it'll be inference dominated and people reporting they have 85% gross margins on inference

Side-note, the Funda Anthropic ARR tracker is at $63.7B:

Gavin on GLM 5.2:

I do think how good GLM 5.2 is has challenged some of my beliefs. And there was a great post from a TPU engineer that for sure distillation has happened. There's been an immense amount of distillation. No question.

But now that this model is so good, it is good enough to do its own RL and you know the kind of cat may be out of the bag.

Now I don't think we really know how good Mythos is. We don't really know how good the next OpenAI model is. We haven't seen the next SpaceX model. So maybe that gap opens up again.

But either way, I profoundly believe the future is composable models and you're going to every enterprise. You're going to have a what Andre Karpathy called a council of LLMs.

What I mean by a composable model when you have kind of um, you know, kind of a symphony of models working together with kind of the frontier models being, you know, maybe the conductors.

(Essentially he means most of the work being done by open source and the most difficult queries going to frontier models).

A common misconception that a lot of people have is that open- source models are, you know, somehow bad for AI. They're awesome for the AI infrastructure providers.

They just shift economic value from the margins of the frontier labs to the infrastructure. And that's not bad for AI.

But I do think there's still a role for these frontier models and it may be true to date frontier tokens are capturing 90% of the economic value and open source tokens are probably 80% plus of tokens processed and those ratios may be here to stay but I just think composable models are the future.

Sacks:

So we now have a Chinese openweight model that is as good as the currently available models from OpenAI and Anthropic. And look, this is a point I’ve been making really since I joined the administration is that we are in a very competitive situation with China.

I’ve been saying this from the beginning. Our whole AI strategy from the get- go was about winning this AI race, defining it as a race, as being globally competitive.

And we cannot afford to do things unnecessarily that slow our companies down.

But look, my view on it is we should not reward Daario by giving him exactly what he's always craved, which is some sort of labyrinthine government approval process that does reward regulatory capture. So I hope that very soon now do think that as long as Anthropic has resolved the jailbreak issue then I do think they should be allowed to come back to market and similarly for open AI. I don't think we should be delaying them unnecessarily.

CXMT

SemiAnalysis also sees Naura as a big CXMT WFE winner:

There is a long discussion of CXMT in the 27 June thread in the Substack chat, including thoughts on beneficiaries and what Apple is trying to achieve.

Subscribe now

Neocloud contract pricing

Read more

Separately, Funda is ramping up its expert interview offering this month. A snapshot of some of the interviews published last week:

Led by the same analysts writing the reports, we are aiming to complete a hundred calls a month, summarised in a weekly update to clients.

The firm prides itself on the quality of the channel checks that feed into the reports as well as how quickly it responds to discourse dominating the AI trade (e.g. TurboQuant and InP export restrictions).

I’m particularly looking forward to being able to address more hot button topics by tapping on our growing expert network. For example, we could run a call on each of the topics discussed in the email below. Watch this space.

Please reach out to sales@funda.ai for more details.

Rigaku: 3,100 yen target from JPM

I believe the “Growing Demands” column describes an X-Ray application that Optical / CD cannot adequately perform.

PCB Supply Chain: Optical module upgrades benefit Compeq, ZDT and Unimicron (Commercial Times)

The increasing scale of AI clusters is driving demand for high-speed optical communications. With 800G optical modules entering mainstream deployment and 1.6T products being rapidly adopted, high-end PCB designs are moving towards higher layer counts, lower-loss materials, and mSAP (modified semi-additive process) manufacturing. Industry experts point out that the upgrade of optical modules not only boosts shipments but also raises the technical threshold and unit price of PCBs, benefiting supply chain companies such as Complux (2313), Zhen Ding-KY (4958), and Unimicron (3037).

Another bullish piece on Unimicron.

Kingboard:

Read more

Optics: Weakness yesterday attributed to Dongshan expansion plans

Sorry, don’t have a better source than Futu.

Some concerns about VCSEL as well, and I’d expect Funda will have something to say about that soon.

Kingboard Holdings / Laminates

Doubled since it was flagged to us on 12 May. Price hikes + vertical integration make this quite a popular name with Asian investors.

On top of that, AT&S will potentially be this newsletter’s first 10 bagger (more on that below the paywall).

The “bottleneck trade” is very much alive, contrary to sniping on X. We’ll follow up on our Musashi-adjacent idea soon.

Subscribe now

Semicap: Penang / Kulim takeaways

I was in Penang last week and have been reporting live to the Substack chat, which has been very entertaining.

Takeaways below:

Read more

Back for a conference in Singapore before heading to Penang / Kulim for the weekend.

Marvell: Production capacity secured to meet 50% DC growth target this year and 55% in 2027 (Digitimes)

Following on from S&P inclusion we got this Digitimes interview:

• COO says that current revenue forecasts were foreseen years ago; have been providing five year projections to suppliers since 2021; built trust with suppliers by meeting forecasts.

• If Marvel can secure more supply today, revenue growth would be higher.

• Supply chain constraints actually good for smoothing out cyclicality.

• “Products incorporating Marvell IP come in many forms, sometimes an XPU, sometimes a network interface controller, sometimes an accelerator device, or a storage accelerator, or a new memory computing device. There's a misconception that Marvell is in the XPU business, but the company is actually in the I/O business.”

  They are going to need all the production capacity they can get to meet Mizuho’s expectations for the ASIC ecosystem:

  ---

  Macronix / Yageo : May revenue out today

  

  Macronix

  

  Yageo

  ---

  AP Memory / Intel

I’m told there was some enthusiasm about silicon capacitators at Computex, though the focus seemed to be on UMC and PSMC.

“Intel’s EMIB-T packaging will see large-scale adoption in 2027, while Google TPU v8e and Meta’s in-house CPU have already confirmed their use.”

From the chat this morning, AP Memory missed May revenue expectations:

Notably, sitting at the top of that list, Kingslide is hosting investor meetings in New York this week. Kingslide was also called out in MS’s Computex note yesterday, observing that Kingslide is the unsung hero found everywhere on the Computex floor.

Rigaku: Bruker also flags H2 inflection, suggests doubling of business by end of 2026

Read more

It has been a very productive week at Computex - Fabian and I met Semi-Twitter celebrities and supply chain sleuths, and refined our booth-hopping technique.

It will take some time to piece together and digest all our notes. In the meantime:

• The FundaAI team is tracking implications for the likes of Marvell, Intel, AMD, Qualcomm, Micron, Sandisk and Kioxia. Also in the background we have analysts dropping notes on Coherent Lite, NPO (institutional report), and Cybersecurity.

• Funda also tracks revenue growth at frontier AI labs, now getting more attention because Sam said token costs is becoming a “huge issue”.

• We have at least one new pick for this newsletter arising out of Computex, could well be our pick for June. It has barely run up despite being in a very hot space right now.

  Subscribe now

Memory: SA plot twist; Jensen in Korea; Only YMTC and Kioxia adding Nand capacity

So the SA note yesterday carried the rather bearish headline:

Thanks for the Memories - VR NVL72 SOCAMM DRAM Capacity per rack Cut in Half.

But ends with:

We believe the reduction in SOCAMM content for Vera Rubin is being driven by constrained DRAM supply across all three memory suppliers.

Also not sure what the market is doing selling off Hynix and Samsung when Jensen is Korea and when the media feels like it needs to report what he is eating:

Samgyeopsal (pork belly) and soju are the likely dinner menu items.

Separately, Jukan reported today that SK Hynix is not adding Nand capacity:

FPS: +8% yesterday

We hit the double on FPS yesterday since we first flagged it in the chat. Thanks again to Guasty Winds for the tweet about it dropping below $30.

We have also been helped by the White House designating electrical infrastructure as essential to national defence as well as this chart:

We are going to comp folks for great contributions like this going forward.

Victory Giant: -11% as US weighs subsidies to boost domestic PCB manufacturing

Some attributing to this story.

From the chat:

Read more

In particular, pleasantly surprised to learn today that silicon capacitators are levered to TPUs.

As always, thank you to those who flagged these to me.

MLCCs

MLCCs continue to be a US investor favourite.

A few notes out this week circling around the idea that the AI server is creating a structural, multi-year boom in high-end MLCCs.

GS put it most bluntly: they think this AI-driven MLCC cycle "will be the largest and longest in history," and that we're "still in the early stages."

Their note of a sell-side meeting with Murata Manufacturing held on 27 May:

1. Murata previously assumed AI infrastructure investment would peak around 2028, now sees possibility it extends to around 2030, partly because power shortages stretch the build-out timeline. On top of that, management flagged “edge AI” devices — autonomous driving/software-defined vehicles and humanoid robots — as a second growth wave over the next few years, playing to Murata’s strength in compact, lightweight components.

2. Pricing - Rather than hiking prices on identical parts, Murata sets fresh (higher) prices on each new-generation product as platforms change annually and bills-of-material climb. Management frames this as mix improvement; Goldman calls it a de facto price hike.

3. Silicon capacitors — Management acknowledged that TSMC, SEMCO, and Murata all have silicon capacitor technology. For PKG substrates with embedded capacitors, silicon capacitors are thin, enabling good surface flatness and high connection reliability with circuits. On the other hand, MLCCs are superior in terms of capacitance and customer usability. The company suggested that it is also developing MLCCs with high connection reliability and will be able to handle either approach.

Yageo

From the chat:

…its not true that Yageo is getting bid because of the spillovers going to tier 2 players.

Read more

JPM continues to see semi trade extending its longevity supported by:

• Agentic AI benefits gradually materialize in real life applications across industries.

• Emerging adoptions in physical AI, like robotics, autonomous driving.

Both driving CSPs and enterprises to invest more.

JPM continues to like AI beneficiaries in Asian tech supply chains, naming TSMC, ASE, Alchip, Unimicron, SEC, and SK Hynix.

Citi with similar comment noting NVIDIA suggested the tightest constraints are not in GPU silicon alone, but broader AI system deployment capacity. Key bottlenecks now include HBM supply, advanced packaging, optical interconnects, liquid cooling systems, rack integration, and power infrastructure.

FundAI made these observations more than a month ago here.

Taiyo Yuden: +11% on Friday after sell-side roundtable on Thursday.

GS, MS and Citi published after the meeting with Buy (¥7,100), EW (¥4,700) and Sell (¥2,900). The stock ended the week at ¥9,100.

• MS says Taiyo Yuden’s MLCC market share (11.3%) is third place behind Murata (40.8%) and Semco (22.5%)

• AI server MLCC sales grew ~30-35% YoY in FY3/26 and guided to accelerate to +80% YoY in BOTH FY3/27 and FY3/28 (a ~3.2x jump in two years).

  • However, servers were only ~6% of total company sales in FY3/26 and GS adds that MLCCs for servers were just over 9% of total MLCC sales.

• Taiyo Yuden forecasts 32% CAGR for AI server MLCC demand 2026-2030 (4x over five years), which the company itself frames as conservative.

• No MLCC price hikes this fiscal year, but FY3/28 is a possibility. GS and Citi tie say the company will hike if there is ~95% capacity utilization.

• Mix shift toward smaller, higher-capacity parts is the ASP driver. Continued migration to small-case high-cap products is lifting ASP and carries high margins (GS gives the specifics: 47µF moving from 2012/1608 to 1005, 100µF from 2012 to 1608 in 2H 2026).

• Capacity expansion held at ~10%/year. Management cautious about going higher given past overbuild.

• Management acknowledged attempts to reduce MLCC count/capacitance per board via vertical power delivery and component-embedded substrates. The company (and Citi) view this risk as small.

None of the analysts appear prepared to chase and up pay up for this.

We hear investors lean towards Yageo and Murata, but happy to listen to counterarguments.

Mitsui Kinzoku: +6% after medium-term plan briefing (FY2025-27) after close on Thursday

Mitsui Kinzoku holds a near-monopoly — ~100% global share — in carrier/ultra-thin copper foil, sold under the MicroThin brand. It is a required ingredient for building the ultra-fine circuit boards used in advanced chip packaging and high-speed optical transceivers. Mitsui also sells a high-speed foil (”VSP”) for AI servers and switches.

Read more

As always, if you can’t meet in person, always happy to hear from you in DMs about what you are interested in hearing more about.

This is the usual weekly roundup of what we have been discussing online, in the chat, DMs or on X.

Read more

Lots of questions about whether the boat has been missed so I thought to summarise the bull case being circulated by key opinion leaders and reproduce some highlights from the chat.

AI server demand for a tiny but critical component — the capacitor — is inflecting hard, and supply can't keep up:

• GPU power consumption is exploding across generations, and more power means proportionally more capacitors required per rack.

• New rack power architectures require more expensive, harder-to-manufacture capacitors, lifting average selling prices even before counting unit growth.

Read more

We feature one such idea contributed via email today.

This is said to be one of the cheapest ways to have copper clad laminate (CCL) exposure. You can own the majority of a CCL company through a holdco, with the holdco’s four other divisions covering almost the entire market cap of the holdco.

The tightness of the CCL market, and its attractiveness to investors seeking AI exposure, is evident from this story translated in Jukan’s tweet:

Subscribe now

Read more