WSTS: 2026 semiconductor market forecast +90% to US$1.51 trillion
MONSTER upward revision from $975 billion for 2026.
Memory alone expected +250% to over $800 billion this year.
Americas +112%
Asia Pacific +87%
Europe +58%
2027 seen +27% to $1.91 trillion
When European Politicians Confuse Wordplay With Insights
I chose to live in Europe because I like the continent. I wish nothing but the best for its people and its economies.
That is why, when I see a senior European politician @carlbildt repeat this dangerously misleading line — essentially defending predatory Chinese exports — I feel not just disappointed, but disillusioned.
Comparing China’s industrial overcapacity to Germany’s cars, France’s wine, Sweden’s trucks, or Italy’s fashion is not analysis. It is wordplay.
According to a 2025 IMF working paper, the estimated fiscal cost of China’s industrial policy support reached about 4.4% of GDP in 2023: roughly 2.0% of GDP in cash subsidies, 1.5% in tax incentives, 0.5% in land subsidies, and 0.4% in preferential credit.
Apply that to China’s 2025 GDP of RMB 140.19 trillion, and you get roughly RMB 6.17 trillion, or USD 900B, in annual industrial-policy support. Over the past decade, even using a rough 4% of GDP estimate, this runs into trillions of dollars — roughly the size of Germany’s annual GDP.
And that still leaves out suppressed wages and benefits, policy-directed capital, regulatory favoritism, local-government support, and the broader political system that forces capital, labour, land, and energy to serve state industrial priorities.
So no, this is not Germany making good cars or Italy making beautiful clothes, or France subsidizing its farmers.
Cantonese has a saying: 蚊髀同牛髀 — comparing the thigh of a mosquito with the thigh of a cow.
If you call both simply “thighs” and pretend they are the same thing, you are either stupid, or deliberately misleading.
DID YOU LISTEN ANON?
Reuters: New Sivers x GFS strategic collaboration.
$SIVE has now announced its lasers will be integrated into reference designs built on Globalfoundries Silicon Photonics Platform.
For pluggable optical transcivers, CPO, and SiPH.
This is fundamentally the most groundbreaking news for Sivers in history.
As Broadcom, Nvidia, Marvell, AMD, and anyone who goes through GFS silicon photonics has Sivers embedded as a default laser route.
I personally think this news alone should easily 2x or 3x Sivers market cap over the medium term, given how fundamental this is to their revenue.
To have Sivers be the standard laser route for the many hyperscalers that use the world's leading photonics foundry.
▶ Marvell CEO says copper wall is moving inside the rack, and copackaged optics is the only way through
• Marvell CEO Matt Murphy emphasized at Computex 2026 that the next bottleneck in AI infrastructure is not compute or memory but connectivity.
• The shift from copper to optical interconnect is already underway and is expected to trigger a large scale demand cycle within the semiconductor industry.
• He highlighted Marvell’s sophisticated engineering capability, integrating advanced CMOS DSP, fourth generation SiPh, and SiGe based broadband analog technology through its Coherent optical modules.
• Marvell’s first 102.4T switch dedicated to AI data centers, the Teralynx T100, is built on a 3nm process, draws under 1,000W, and delivers up to 25% lower power than competing solutions.
• The T100 routes signals through copper traces on the PCB to optical modules on the front panel, whereas a CPO switch connects optical fiber directly to the package and removes copper wiring entirely.
• The reach of copper cable is inversely proportional to bandwidth: at 100Gbps per lane it can carry signals about 5m, but at 200Gbps this shortens to roughly 2.5m, and at 400Gbps copper can no longer make connections even within the rack.
• Each time the “copper wall” moves one step, the number of connections that must shift to optical increases at least tenfold, which is expected to drive explosive demand across the optics industry, and the Taiwan supply chain is already expanding to respond.
• The number of connections inside a rack is roughly ten times the number of connections between racks, so conventional pluggable optical modules alone cannot address the power and space limits; CPO solves the connectivity problem by integrating the optical engine directly into the switch and compute package.
• Nvidia’s Vera Rubin platform has already adopted Spectrum-X Ethernet Photonics, the first CPO based switch to enter mass production, a case showing that the CPO transition has moved beyond proof of concept into actual commercialization.
• As optical connectivity extends into the server itself, compute, memory, and network resources can be disaggregated and dynamically configured per workload, enabling a shift from a fixed server architecture toward operating the entire data center as a single integrated system.
• He stressed that the CPO transition is impossible without Taiwan’s manufacturing ecosystem, explaining that Marvell has accumulated high volume PAM4 production experience, field data, and supply chain infrastructure including ASE.
• Over the past decade Marvell has invested a total of $36 billion to acquire companies such as Inphi, Cavium, and Celestial AI, expanding its connectivity portfolio.
• He emphasized that Marvell is the only company able to address the full connectivity stack of an AI data center, from millimeter scale inside the package to kilometer scale between data centers.
$MRVL
PODCAST: ARE WE IN A BUBBLE?
1/2 hikes from Central Banks this year?
Alarmed by selloff in Govt bonds?
@darioperkins@freyabeamish
Apple: https://t.co/ddrdQIJo75
Spotify: https://t.co/nNBp03bbyq
Just some mobile shower thoughts around $XFAB and train of thought:
1. 800vdc $NVDA push look for GaN/SiC players / power semis.
2. $NVTS and other fabless/fab-lite beneficiaries of $NVDA push probably use foundry models
3. care more about Western supply chains over Asia, want to build up Western capabilities + Western premiums.
4 China has a lot of capacity, maybe risk into 2028, but again it’s building up Western supply chains
5. XFAB only high volume SiC foundry in America (others like $ON or Infineon are vertically integrated)
6. advanced 6in SiC, 8in GaN on Si, building out 8in GaN
7. Maybe likely they’re developing 8in SiC from CHIPS backing, just not public material
8. check SiC revenue -> 152% Y/Y growth okay. Probably something markets missed, since blended looks worse from automotive slump, that should come out recovery
9. $NVTS and others turns out to use $XFAB. $POWI cites $XFAB in filings, among others.
10. both are $NVDA power semi explicit partners, great exposure indicator to 800vdc power semi players.
11. US Dpt. of commerce cites $XFAB as only high volume SiC foundry in the US, $50M PMT
12. validation from US Gov about critical component in supply chains is amazing
13. EU CHIPS Act gives $XFAB $128M EU, for foundry (MEMS, AI, etc), okay turns out they’re critical MEMS player
14. So that’s validation from EU gov about critical component in supply chains, dual continent subsidies
15. So now we know $XFAB is a critical MEMS foundry so you get SiC capabilities, GaN development, and MEMS upside
16. they also got $47.6M EU funding for leading Silicon Photonics supply chain in EU. So that’s EU funding on multiple angles.
17. Turns out, I know all the players there from smartphotonics from $GFS deck.
18. $NVDA and $NOK are qualifying them for silicon photonics HVM. I think this is just a government backing angle for success in EU photonics so likely to succeed… kinda like how Us gov encourages everyone to use $INTC.
19. Okay chips act 2 is coming out next week… so they’ll probably get funding there or more revenue commitments
20. 1.28 p/b, now that’s probably just book cost? Likely coming out of $SOI type legacy drag cycle.
21. Did some modeling around actually replacement values, true replacement p/b cost likely ~.5/.7.
22. Getting business for free, while having upside from SiC near term into Silicon Photonics / GaN as main growth past H2 2027.
Thoughts: derisked by p/b values + replacement value. maybe like 20% downside from macro.
However, critical dual continent importance. So downside risk seems low, but upside is compelling.
Lot of capex likely backstopped by upcoming chips act catalyst + national security concerns.
Maybe 2.5-4x rerating seems possible/likely.
Not a 10-20x, but recovery from depressed valuations from silicon photonics upside with SiC / GaN bridge.
TLDR: likely trading lower than replacement value, dual continent subsidies likely subsidize capex.
Gov grants shows importance to Western supply chains, photonics longer term upside, SiC/GaN demand likely near term upside and bridge.
Don’t control any recent volatility, should shake out anyone not really confident in the thesis though.
CHIPS act 2 from EU is coming up, $XFAB was listed in earlier blueprints for optical ecosystem, so should get a boost after that comes out as near term event catalysts.
So now is the risk reward seems compelling, we’ll see if this is right or not
What's happening in the MLCC market
First off, MLCC as a whole is a $15B market. MLCCs for servers were a $1.3B market in 2025 ($600m for AI servers, $700m for general servers)
The AI server MLCC market is growing at 80%+ CAGR, and the general server MLCC market will also accelerate due to agentic AI increasing CPU demand (around 30%-40% CAGR)
We will see negative growth in the smartphone/mobile MLCC market for at least 2026-27.
Humanoids are another future high-growth market for MLCCs
Book-to-bill ratio for most MLCC suppliers is over 1 now
Reasons for price hikes-
High Nickel & Silver are affecting all segments
There is a supply-demand mismatch in the high-end (high capacitance, high voltage) segment, which is used in autos & servers
High-end MLCC lead time is over 20 weeks
Spot/distributor prices have increased by 20%-40% for low capacitance & consumer device MLCCs due to hoarding and double booking, especially in China
OEM contracts have not seen large price hikes yet
What's happening now:
Rapid capacity expansion happening across the industry
Murata expects blended ASP prices to remain flat (ASP going down in consumer electronics, expansion in AI server market)
Tier 1 players like Murata, Taiyo Yuden, SEMCO building capacity to serve AI server MLCC market
This will create opportunities for Tier 2/3 and Chinese suppliers to expand in the mid to low end market (Macronix effect)
Future:
MLCC production equiment & raw materials suppliers will be the biggest beneficiary of this CAPEX boom
MLCC producer stocks have performed well, and it is finally spilling to raw material/equipment producers
I expect them to outperform MLCC producers now
US Treasuries are experiencing the worst bear market in history:
The US Treasury Total Return Index has now been in a drawdown for 69 consecutive months, the longest streak in over 100 years of data.
The previous record stretch that ended in 2019 lasted for ~30 months.
This is also only the 3rd time in history that a drawdown has exceeded 20 months.
During the current drawdown, the US Treasury Total Return Index fell as much as -18% from 2020 to 2022.
Since then, it has recovered some of its losses, but it is still down -6% since 2020.
Meanwhile, the 20+ year Treasury ETF, $TLT, is down -40% since its April 2020 peak.
US bonds are more unpopular than ever.
Strait watch 👀
4million barrels of crude went through the strait today
BBG: It’s the first time in a week that four million barrels of unsanctioned crude have been seen crossing the waterway, with ships leaving in bunches and then lower volumes passing in subsequent days. So while Tuesday’s number is high, it may be offset by a decline in the coming days. The two tankers were carrying crude from Saudi Arabia and the United Arab Emirates respectively.
The EU is 6% of global emissions. Committing economic harakiri in response to climate concerns is pure vanity. Europe can't save the planet because the continent doesn't have any meaningful influence on the outcome.