haYN Capital

Is Claude slowing down the past week for anyone else???

$SOI Soitec February 17 2026 CEO Interview English Translation Host: Our guest now in the big live interview on Boursorama is Pierre Barnabé. Good morning. Pierre Barnabé: Good morning, David. Thank you for having us on Boursorama. I'm the CEO of Soitec, the world leader in innovative materials for semiconductors. Host: Every time I have you on, I ask you to explain concretely what you do, so we don't lose anyone. You are a very specific link in the semiconductor value chain — chips are everywhere, in phones, in cars, in AI. You're very upstream. Explain simply what you do. Pierre Barnabé: Very simply. I'll try to keep it very simple. The vast majority of semiconductors are built on silicon bases. Silicon is a material made from sand, to put it simply. We heat sand to very, very high temperatures to form ingots, and then we slice them into wafers. Host: And semiconductors are manufactured on those wafers? Pierre Barnabé: Exactly. So we've managed to make that silicon wafer "intelligent" using French technologies derived from the CEA that we've been exploiting for 30 years. We combine these materials — which are inert at the base — to make them intelligent. That's the unique value Soitec brings. It means the components manufactured on these substrates are much higher-performing and much less energy-consuming. Host: So in the semiconductor chain, Soitec is right at the very beginning — you sell to what are called foundries, like TSMC. What does a foundry do? Pierre Barnabé: A foundry manufactures processors by depositing oxides of different elements through physicochemical processes. Those processors are designed by what we call fabless companies — they design and imagine the chip. The foundries then have clients like Nvidia, Qualcomm, for example. Host: Good. So that sets the scene for what you do concretely. You published your results at the start of the month — they were down by a third? That's your third quarter, October to December, which is slightly offset in your calendar? Pierre Barnabé: Yes, it's offset. €160 million. How do we read that number? I think your view is that you're at an inflection point — the glass is half full. Pierre Barnabé: It's more than a point of view — it's a reality that has just recently been recognized. We had a very strong period through 2021–2022 where we equipped virtually all smartphones in the world with a single product. Soitec supplies 100% of the world's smartphones. After Covid, the stock hit extremely high levels in 2021. Then there was a strong drop in mobile phone consumption, with a lot of overstocking at our clients. We had to completely transform the company — from one that lived on a single product whose activity was divided by four... Host: That was the strategic error — not yours necessarily, but the one you've corrected — building the entire success on a single product. Pierre Barnabé: The strategic error would have been not to diversify, which I did. The answer was to move into four other products and build a company now based on five major product lines, also benefiting from two new products serving the AI world. That's what makes the difference. Host: So — are smartphones recovering? Pierre Barnabé: Smartphone activity is slightly improving. It represents about one-third of our RF-SOI revenue, whereas it used to be quasi-100% — maybe 80–90%. It's fallen sharply, but we've managed to diversify. We're gradually destocking. By end of this year, the smartphone segment will be fully destocked, and we can expect an inflection point there — but fortunately the other segments are compensating. Host: The automotive market — cars are big chip consumers, right? Pierre Barnabé: Yes, because the car is intelligent and getting more so — more secure, more autonomous, more connected. How many chips are in a car? Host: Depending on the model, you can get to 10,000–15,000 processors for top-of-the-range models, but an average model has around 3,000–3,500 processors. Pierre Barnabé: You supply 100% of smartphones, but not 100% of cars? Host: Not 100% of cars, but quite a lot, particularly around energy management — for the electric drivetrain, toward the battery or the cockpit. Pierre Barnabé: Do you supply the OEMs directly or the foundries and fabless companies? Pierre Barnabé: We supply foundries and fabless companies who then supply the manufacturers. So among our major clients we have STMicroelectronics, Infineon, NXP, as you know them all. Host: Smartphones, automotive, industry — and AI. Is that where growth is coming from? That's about one-third of revenue today? Pierre Barnabé: Yes. Our AI activity is growing 27% year-on-year, reaching €54 million — so €54 out of €160, roughly one-third. That's the future. When you think Soitec, world leader in semiconductor materials, with the AI wave, you'd expect them to be right at the top — and it's growing very fast. But we're not going to end up as a single-product company again around AI. Pierre Barnabé: Absolutely not. I can tell you that lesson is well learned. This company will remain multi-product. Five today, and we're working on five more — we've doubled our R&D effort and we have five products currently in development that should come through. Host: What do you do concretely for AI? Pierre Barnabé: We have two product types. First, FD-SOI (Fully Depleted Silicon-on-Insulator) — it converts analog to digital and vice versa. It's very well suited for, say, embedded radar in vehicles — low energy consumption, capable of computing edge signals. That's the FD-SOI product. And then we have Photonics SOI, which converts electrons to photons, enabling data to travel through data centers at the speed of light. Host: The magic word — photon. Pierre Barnabé: Photon to photon. And the other magic word is GPU to GPU — the famous graphics cards. Right now, we supply 100% of new-generation AI data centers with our photonic wafers. 100%. Host: So you supply Google, everyone — all data centers, even if not directly as a client? Pierre Barnabé: Not directly, but we supply photonics-SOI wafers that equip the transceivers connecting servers. And we have 100% of the business because this technology cannot be competed against — it's the most efficient. It allows data to travel at the speed of light, server-to-server today, and tomorrow chip-to-chip (GPU to GPU). Host: How much will that represent for Soitec? Pierre Barnabé: AI is growing 27% year-on-year and will weigh increasingly on our activity. And we'll find AI appearing in more and more of our markets — even smartphones will increasingly embed AI. So AI will become a larger and larger part of what we do overall. Host: When does durable growth return to Soitec? 2027? 2028? You're guiding for +20% quarter-over-quarter in the current quarter? Pierre Barnabé: Yes, we'll be +20% from Q3 to Q4. We see a sequential inflection point. We're building a platform for durable growth. For a full fiscal year of growth — in coming years, we'll be growing, but for now we communicate quarter by quarter. Host: Why only quarter by quarter? Pierre Barnabé: Because it's not simple. There's still fog. Smartphones — we're waiting for inventory correction. Automotive — still uncertainties. AI is going very well. You have to combine all of that. What is certain is that we have an inflection point, we've built a growth platform, and we will grow durably again in the years ahead. Host: About a quarter of your revenue comes from China. Chinese players will try to undercut you — lower prices. How do you manage that, especially since your Chinese sales are well-oriented? Pierre Barnabé: True — four years ago we did almost nothing in China. Now it's over 20% of revenue, so it's become very important. In China, we sell innovation, we sell value. Chinese customers buy value — they buy Soitec because our products are well ahead of any competitor, Chinese or otherwise. We'll keep supplying China without transferring any technology. The key is: innovate, innovate, innovate. That's why we doubled our R&D. Run faster than everyone else. And the Chinese, ultimately, buy quality products and are willing to pay for them when they add value to their end products. Host: Net result — what does it look like? Pierre Barnabé: We haven't yet communicated the net result. There's also an inflection point there. Are you positive or negative? We'll be... at an inflection point. That's all we'll say. On 2024, we had a fairly solid net result because revenue was relatively high. After the sharp revenue decline, we managed to contain costs and protect margins through diversification — that's what allowed us to hold the line. Host: The stock — you were at €240 in November 2021, today you're at €35. Down 70% last year, up 60% YTD. Is this the bottom of the cycle? Pierre Barnabé: We're passing through the bottom of the cycle. You had a single product with massive overstocking — a division by four of that activity. The four other products have broadly compensated. Now we're in that transformation, at exactly that inflection point. Once factory utilization rises, profitability follows. On photonics specifically — we have far more demand than we can supply. That will drive revenue growth and strong profitability improvement in coming years. Host: Why are you leaving? You announced your departure last October. Pierre Barnabé: For personal reasons. Transforming a company like Soitec from where we were — it's an enormous amount of effort, intensity, travel. Having seen the turnaround and the positioning for growth — five products, five new ones coming, outstanding teams, significant innovation investment — I felt the moment had come, once my duty was accomplished, to take care of myself. Host: Laurent Raymond will succeed you in April, coming from Infineon. What will his main challenge be? Pierre Barnabé: Accompanying this growth — making sure that all of our products that are at the $100M scale grow to $200M and beyond. His challenge will be to see demand flowing in from customers on new products and being able to meet it. And then, from the five new products we're developing, making sure a few become commercially successful. That's the mission. He has teams that are totally determined and highly motivated to support him. Host: What do you say to investors? Those who bought at €240 and are sitting at €35. And those who might be looking at this now as a potential turnaround play? Pierre Barnabé: You know I can't comment on the share price. But I will say: this is a company that has been profoundly transformed — now based on five products, across all geographies including China at 20%+ of revenue, with 20 clients making up 90% of sales versus just a handful four years ago. A company that has been restructured, stabilized, with a growth platform now — and the profitability to match. I'll say no more. Host: Where will you be in a few months? Until March 31st, it's all about Soitec. After that — we'll have you back. Pierre Barnabé: The appointment is made. Thank you. Host: Interview with Pierre Barnabé, CEO of Soitec, listed on the Paris Bourse on the SBF 120. Thank you. Summary: Soitec CEO Pierre Barnabé discusses the company's pivot from a single-product (RF-SOI for smartphones) model to a five-product portfolio. Key themes: smartphone destocking nearing completion, photonics-SOI capturing 100% of new-gen AI data center transceiver demand, FD-SOI growing in automotive, Q4 guidance of +20% sequential growth, and a sequential inflection point. Barnabé announces his departure for personal reasons; successor Laurent Raymond (ex-Infineon) takes over in April 2026. youtu.be/eapOZ9IBqi4?si…

$SOI SOITEC Likely Current Customers based off geography mix Singapore (~50% of H1 revenue — €115.5m): This is almost certainly GlobalFoundries. GF operates one of the world's largest fab complexes in Singapore (Fab 7) and is the dominant RF-SOI foundry globally. The sheer concentration — half the company's revenue from one country — strongly suggests GF is the single largest customer, probably by a wide margin. GF is the go-to foundry for Qualcomm, Skyworks, Qorvo, and others on RF-SOI process nodes. UMC also has Singapore fabs and does some SOI work, but likely a much smaller piece. France (~31% — €71.4m): STMicroelectronics — Crolles fab is literally down the road from Soitec in Grenoble. ST is the highest-profile adopter of FD-SOI globally and has built multiple process nodes around it. This is almost certainly the bulk of France revenue. Possibly some X-FAB or internal qualification volume, but ST dominates. United States (~12% — €26.7m): Two likely customers here. Tower Semiconductor (Newport Beach fab) — your confirmed biggest SiPho foundry customer, so this is the photonics-SOI revenue showing up. And Skyworks directly — the doc references a multi-year POI supply agreement for the SKY5 platform. Possibly some volume to Intel Foundry Services if they're pulling any SOI. Other (~7% — €17.2m): Likely a mix of TSMC (Taiwan — COUPE platform, probably still early/small), Japan (Shin-Etsu royalties or direct sales to Japanese IDMs), and possibly some Samsung or Dongbu HiTek volume.

Me thinks the SK HYNIX US ADR may be the catalyst that starts to close the gap between forward multiples of memory companies vs photonics companies. Memory companies are priced very cylical as if the cycle ends before 2028/2029 and Photonics are priced as secular growth. I think reality is somewhere in the middle for both of these themes. SK Hynix is a trillion dollar company imo $EWY $HY9H

Use it mainly to identify my directional bias long/short, and gross/net exposure levels. I have it color coded as well. Red = Seriously think about sitting in cash/reducing exposure/shorts Yellow = Long but picking spots more carefully not using margin and not going long assets which rely mostly on liquidity like cryptocurrencies or high beta shitcos Green = Using leverage / employing trend following strategies etc...

@killapabkai How are you interpreting this in terms of exposure and asset choice? Or, more simply, how is it driving your decision making?

Finalizing my automated Macro Liquidity assistant built with Claude Code Basically just took all the data points I track and gave AI my Macro Liquidity Framework to make my own little macro intern. Enjoy! Pasting full text below in case this image is too small too read. All four liquidity pillars are cracking simultaneously — stagflation is draining money supply growth while policy stays frozen and cross-border flows face a fragile dollar and stretched carry trades. Growth Growth is decelerating into a stagflationary trap with no escape valve. Atlanta Fed GDPNow at 2.33% SAAR has just been revised down 48bps, driven by a sharp consumption deceleration to +1.92pp contribution as $112 WTI acts as a regressive income tax — and the NY Fed's more conservative 2.07% read suggests the true underlying pulse is even weaker. Critically, real disposable income growth of only +1.75% YoY is being outpaced by borrowing costs at 3.50-3.75% FFR, meaning the primary engine of credit-fueled money supply expansion — real income growth exceeding debt service costs — is broken; consumers and businesses have no excess real income to leverage up against. Inflation Inflation is not contained and is actively re-accelerating on the Fed's preferred measures. Core PCE at 3.06% and supercore at 3.51% are both trending higher, with PPI running at 3.37% YoY for six consecutive months of sequential gains that will pass through to CPI and PCE in coming months; the benign headline CPI print of 2.66% is the outlier, not the signal. With UMich 1Y inflation expectations at 4.00%, real purchasing power is being eroded faster than nominal incomes can compensate, simultaneously choking real demand and locking the Fed out of any easing — the exact dual poison that kills liquidity expansion. Policy Policy is unambiguously restrictive and incrementally tightening at the margin despite no Fed action. Real yields remain deeply positive (5Y TIPS 1.27%, 10Y TIPS 1.88%), the Fed is parked at 3.50-3.75% with 17/19 participants flagging upside inflation risk and only one 25bp cut penciled in for late 2026, and HY OAS has blown out 44bps over three months to 327bps — the fastest credit tightening in this cycle, indicating financial conditions are self-tightening beyond what the Fed is directly engineering. The 5.8%-of-GDP fiscal deficit provides a mechanical demand floor that prevents a full growth collapse but simultaneously keeps inflation too hot for the Fed to ease, creating a locked policy cage with no exit. Cross-Border Flows Cross-border flows remain nominally positive — foreign portfolio equity inflows of $206B in Q3 and total portfolio inflows of $487B are genuinely supportive of US asset prices — but the regime is deteriorating at the margin in ways that matter. The trade-weighted dollar has fallen 3.2 points over three months to 117.5, weakening the carry-and-safe-haven incentive that has anchored foreign demand for US assets; USD/JPY pushing to 155 has stretched the yen carry trade to a point where a single BOJ signal could trigger simultaneous foreign equity liquidation and dollar selling. The current account deficit narrowing to -$226B (from -$440B) reduces the mechanical financing pull on foreign capital, meaning future inflows must be conviction-driven at a moment when dollar momentum is fading and geopolitical uncertainty is rising. Liquidity Synthesis All four pillars are either negative or deteriorating: growth is slowing with real incomes unable to support credit expansion, inflation is re-accelerating and preventing the monetary easing that would refuel money supply growth, policy is restrictive with credit spreads self-tightening and QT ongoing, and cross-border inflows — the last supportive pillar — are sitting on a fragile foundation of a weakening dollar and an overstretched yen carry trade. The net result is that broad money supply growth (M2) has no credible expansion engine: lending is constrained by real borrowing costs that exceed real income growth, the Fed cannot inject liquidity, fiscal stimulus is being offset by crowding-out and term premium expansion, and the international capital bid is momentum-dependent in a fading-dollar environment. A discretionary equity PM should be running meaningfully below normal gross exposure — this is not a market to press; the risk-reward of adding beta in a stagflationary liquidity contraction is deeply asymmetric to the downside, with the single-cut Fed path at high risk of being repriced to zero. Key Risk A sustained WTI spike above $120 on Middle East escalation simultaneously accelerates core PCE above 3.25%, eliminates the Fed's last projected cut, blows HY spreads through 400bps, and triggers yen carry unwind — a quadruple liquidity shock with no policy offset available. STAGFLATION LOCK: Fed cannot ease (core PCE 3.06%, supercore 3.51%), growth is cracking (GDPNow 2.07-2.33% and falling), HY OAS +44bps in 3 months, and USD carry trade stretched at USD/JPY 155 — reduce gross equity exposure materially.

@siyul This is one of my gold miner positions I sold a little over a week ago and I am happy I sidestepped this drawdown and I am smiling watching these drawdown while I patiently wait to start building positions again.

$BTG at $4.05, ~20% below my Jan writeup. $4, last seen in Nov'25 when Gold was ~$4k/oz. Despite a brutual 15% off ATH, Gold still sits above $4.5k/oz, and B2Gold's story remains largely unchanged. I'm adding to an already-full position since yday. "Like it at $5, and love it at $4", true to my own word.

Put in orders to buy $MAI.V and $WWI.AX Thesis is we have reached peak hawkish pricing imo SOFR & FF Futures pricing no cuts for 2026 and 1 in 2027. FED doesn’t have the balls to hike in the face of supply side inflation Most hawkish scenario is they sit on their hands and don’t cut for 2026 which is what is currently priced Max hawkish pricing = max real yields & max DXY I think real yields and dollar only go down from here which are 2 of the 3 main drivers of gold Downside is more degrossing if this war lasts long enough to cause severe economic pain but I don’t think Trump/US has the stomach to cause severe global economic pain all bc of Iran. I think a path to de-escalate most likely appears before max economic pain occurs

Think it may be time to buy back my gold miners soon. SOFR and Fed Fund futures pricing in no cuts for 2026 and only one in 2027. I think we may be near peak hawk pricing. Still assessing situation Will delete if wrong

The main drivers of gold are 1. Lower Real rates 2. Lower dollar 3. Liquidity Geopolitical risk is a reason to buy gold but NOT the main driver of prices AND Correlations go to 1 in selloffs

Bags will be made once the confoict ends. FED realizes Iran inflation will be transitory and they must cut + US & EU fiscal deficits increase to increase defense spending + Dollar continues to fall + deglobalization trend continues increasing gold reserves compared to dollar reserves

Getting really excited about this pullback in the metals and miners. Some extremely attractive companies down 20-30%+ over the past 1-2 months. Great time for those with lots of cash ready to deploy once bases start forming. Love it.

@calvinfroedge Rising Real yields Rising $DXY general degrossing

So why are people selling gold? This one is really hard for me to grasp Only thing I can come up with is rotation to energy

@shravanrayhaan Yes but correlations go to 1 in a selloff so I like to be able to try and avoid the selloffs and buy back cheaper

@killapabkai Dont think this should impact thor explorations that much

Now 100% Cash Sold all gold miners. Hormuz induced cost-pull inflation not good for gold. - higher oil prices = higher AISC - rising dxy = not good - positive real rates = not good - cost-pull inflation without accomodative CB response = negative for liquidity = not good CB buying and geopolitical risk act as some protection to how deep trough may be. Best Case: i may side step a cyclical drawdown in what i believe is a secular trend for gold Worst Case: Gold chops around here until we get more info on real rates, dxy, liquidity *Interesting Px action with BTC slightly up while Gold chops around. Any theories?

Might look for another potential oil long here. You don't move assets from Japan to the Middle East if you are thinking about de-escalating anytime soon.

@sentdefender so I redundancy plan for trading my 6 figure portfolio as a 22yr old more than the US Military redundancy plans for a large-scale military operation 😐

The Pentagon and National Security Council significantly underestimated Iran’s willingness to close the Strait of Hormuz in response to U.S. military strikes while planning the ongoing wide-spread strike operation against Iran, according to multiple sources familiar with the matter who spoke to CNN. Top officials with the Trump Administration acknowledged to lawmakers during recent classified briefings that they did not plan for the possibility of Iran closing the Strait of Hormuz in response to strikes, under the belief that closing the vital waterways within the strait would hurt Iran more than the United States.

anyone else confused how there is commentary on X talk about buying dips when NFP being revised down + Private Credit showing signs of risks + Global supply shock induced stagflation or potentially recessions due to Strait of Hormuz closure with no clear path to de-escalation I am bullish Gold here as I believe FED will be more worried about supply-shock induced recession rather than inflation and bonds will bid if Strait causes meaningful global slowdown which is bullish for lowering real yields + rising geopolitical risk Despite this I am also considering selling my gold miners as well as correlations go to 1 in a melt down.

Cut most of my exposure now 50% cash. Only positions I still have on are my gold miners, small positions in US petrochemicals + SPY put and CL calls $BTG mines relatively AISC sheltered from an increase in oil prices $WGR, $WWI, $MAI, and $ITRG are about to ramp production -> production starts during a time where oil may likely already normalize In regards to CL calls my view is that the US admin is trying to artificially calm oil prices through misinformation on X and through official statements. I believe $85 CL reflects a mix of that misinformation and being over optimistic about TACO. I don't see a clear path to deescalation for either sides and I am betting on this conflict affecting Oil Supply longer than $85 CL suggests.

@mrholty1 Nah i cut exposure for most of my portfolio past couple of days still monitoring the Iran situation.

@killapabkai did you ever get anything to get more comfortable on this. a 2% position for me.

$SLS seems interesting. Initiated a small poker hand type position. SP: $4 Acquisition Price: likely $20+ possibly up to $50 (based on multiple of peak sales for drug pipeline) Market Implied Probability of success using $20 share price: ~20% Data from Phase 3 for GPS seems like probability of success may be over 20% Still doing some research. Would love to get some thoughts on people who know more than me

Adding some $INTC and $AMKR here Nice pullback and needed to add CPU and advanced packaging exposure. $INTC Thesis: - In the coming Q's I think we start to see revenue acceleration as CPU demand continues/accelerates - Agentic AI increasing CPU usage increases in popularity as a theme similar to memory $AMKR Thesis: - also benefits from increase in CPU packaging via Intel partnership