My daddy lost 3K in SOXS yesterday so I have no choice but to use divine power against SOXL.
People invested with conviction that SOXL will fall and SOXS will rise. In Jesus’ name, let the charts move in their favor.
SOXL closed around $173 yesterday, up ~11.6% on the day, and is up roughly 864% over the past year. The 52-week range has been $15.10–$191.29 , so you’re sitting near the upper end of that range.
Bullish case from here:
Both short and long term moving averages are giving buy signals, with the short-term MA above the long term MA a positive setup. Support levels sit around $169 and $118.
The semiconductor macro tailwind is strong global chip sales could approach $1 trillion in 2026, driven by generative AI chips nearing $500B in revenue and massive data center buildouts.
Who’s holding SOXL into NVDA earnings
HUH?
Here’s the translation:
BTC Partial SOXL 9/18 $100 Put
Today SOXL surged 12% — closing the profitable portion of the SOXL $100 sell put leg.
This corresponds to closing the $120 put leg that was part of the original short put butterfly:
BTO SOXL 9/18 $120/100/85 Put Butterfly
Today my core holding AMD surged over 7%.
Evercore ISI raised its price target on AMD to $579 from $358.
Reason: reportedly a massive surge in CPU demand. Another CPU stock, INTC, also jumped 6% today.
Reportedly Intel used a new memory design for its GPU to work around the HBM shortage:
Leaked Intel Crescent Island PCB with Xe3P GPU, 160GB LPDDR5X and 16-pin power
At market open, there was a very aggressive large options order on Intel:
BUY CALL INTC $195C 07/17/26 — $2.2M @ $2.31 (+68.4% OTM) 05/20 09:50 ET
I have a position in Intel but it’s not large.
After hours tonight, Nvidia reports earnings — a major catalyst that will significantly influence the broader market direction.
Perhaps a prelude of the earnings to come.
Hello, I'm the guy that's been posting the "Why is SOXL up/down today?" posts for the past couple weeks, and at the end of every single one I include the same short message "SOXL to $500 EOY". Understandably this has garnered attention in the comments many times so I'm making this post for future posts to link to so I don't have to explain myself each time. I don't expect everyone to come away from this agreeing with me, just not having to ask why I think it and *maybe* not think I'm a complete goddamn nutjob for thinking it.
First, a little math. At the time of writing SOXL sits at ~$150. To reach $500 that would be about +236%. The value of SOXL is essentially derived from 3Xing the daily movement of SOXX (technically SOXX and SOXL both follow the NYSE Semiconductor Index, but for the most part they are in alignment). Because it is 3x leveraged, you might think that would mean SOXX has to go up 80% but actually, just as leverage tends to depreciate, it also compounds. If SOXX goes up 1% a day for 41 days it would be up (1.01^41)-1=+50%, but SOXL would go up (1.03^41)-1=+236% so SOXX would actually only need to go up about +50%. That alone should make the whole thing seem a lot more realistic, and also explain why SOXL can comfortably be up so much from $10.
Secondly, I have to say that it has nothing to do with cycles. I try to be respectful to everyone but this type of pattern or shape finding investing philosophy seems silly outside of day trading. Any supposed cycle should at least be derived from the SOXX chart but the main patterns people refer to don't really exist there. Although with SOXL being a leveraged index it's often not obvious what the price is actually derived from. That's why I started the daily posts, to give people something to go off of beyond that which to me looked like reading tea leaves. Now on any given day you can see exactly what stocks are pulling SOXL up or down, and what the relevant news stories are to that effect. Although I'm not unbiased I try to just relay the numbers and the news and save my stupid takes on the news for the end if anyone cares for my analysis.
So what is my analysis or investing philosophy?
Secular growth. AI demands semiconductors. The better AI gets, the more value semiconductors can provide, the more demand there will be. I don't know or care who's going to win the AI war. I don't know who's going to win out in the semiconductor industry. What I am pretty freaking confident of is that the industry is going to win, so I don't bet on any of the players; I bet on the house, because the house always wins.
This is contingent on AI not being a bubble. If it's a bubble then show me the air. Like in the dot-com bubble there was a shitload of companies getting valued with no income or profits or real business model (example: Enron) or in the real estate bubble there was a bunch of mortgages being given to people with no way of paying them. Where's the bubble for AI? Anthropic is cash flow positive. OpenAI is making tons of money too. Their cash flow is a little ambiguous but it seems mainly to be based on consumer subscriptions with growing industrial subscriptions. I could potentially see a scenario where industry fails to see any value and drop subscriptions, but that seems more speculative than just about anything I'm including here. I can't imagine the consumers dropping subscriptions en masse, once you get consumers on a subscription they tend to keep it or move to a similar service.
Okay so maybe you don't think AI is a bubble per se but you think it's going to hit a wall or not be very useful. After all with SOXL being leveraged the value can't just stagnate it has to keep going up. Well I'm a software developer and it's already making my job much easier. It can easily program in proprietary languages that I'm pretty confident don't appear in its training data. Even at its current level of intelligence there's tons of industry value to be extracted from it that there just hasn't been the time to set up yet, because a couple years ago it could maybe write a function or a class but now it can do a couple days of work easily. That's deep mostly untapped value.
But I don't think it's even going to stop here. All trends point to it getting much better very fast. A common retort is that meme of the guy saying "My 3-month-old son is now TWICE as big as when he was born. He's on track to weigh 7.5 trillion pounds by age 10" with the idea being that the trend is too early to call exponential or that the exponential trend is going to hit a wall at some point. Well this has been going on for some time now, METR published the claim that AI task time horizons double every ~7 months almost exactly 14 months ago now, or 2 doublings. So where are we at on that?
First let's go back to the baby. You can work out that the baby's weight at birth was 7500000000000*(1/(2^(12*10/3))) = 6.82 pounds. At the time of the post the baby was 3 months old and 13.64 pounds. If the joke prediction were correct then 6 months later, 2 predicted doublings, the baby would weigh 4x as much or 54.56 pounds at 9 months old, which is basically impossible and such a baby would be studied extensively for science.
Back to the AI time horizon. Back in March 2025 METR's frontier estimate was that AI could complete programming tasks up to 50 minutes long, with the record being held by Claude 3.7 Sonnet. Now the record is held by Claude Mythos, but they aren't entirely sure what the time horizon is as it's getting hard to build accurate tasks that take so long, but they do say that it is "likely at least 16 hrs". So it turns out METR was actually wrong. Since their observation 14 months ago AI's ability to do long tasks has not increased 4x, it has increased 16x. They have observed that the doubling time has accelerated to ~4 months instead of the original ~7 months. If the 9-month-old baby grew at that rate from 3 months it would weigh 218.28 pounds. That baby wouldn't just be studied, it would need to be abducted by the SCP Foundation.
Looking beyond 2026, if we achieve AGI like everyone in the industry is aiming for, then $500 is just the beginning. A humanoid robot with AGI can do literally anything a human can do. I don't know what the economy looks like with that but it is clearly valuable.
With all that being said I could totally be wrong. Even if my entire thought process is generally correct, it could still just stagnate for the time being or even go up but only to $350, or other political/macroeconomic factors hurt the economy broadly and SOXL is just hit in the crossfire. Or maybe my thought process is wrong and AI adoption stagnates or drops, or AI takes off but chips just don't become more valuable for one reason or another. Hell, I could be basically proven wrong by tomorrow afternoon if Nvidia underperforms earnings. There are certainly a lot of paths to failure. But I plan on keeping the "SOXL to $500 by EOY" line in there even if it does end up looking hopeless later on because I believe in keeping myself accountable, and also because it would be funny.
SOXL to $500 by EOY
The 30‑year Treasury yield just spiked to 5.19%.
That’s the highest level since before 2008.
Semiconductors are basically the poster child for “long‑duration assets.”
Their valuations depend on earnings far into the future.
When the 30‑year spikes:
Future earnings get discounted harder
AI‑capex becomes more expensive
Risk appetite shrinks
Half of the S&P 500’s gains this year came from four stocks.
Higher yields draw capital away from equities into bonds, and analysts warn this reprices growth stocks. The long-duration asset / valuation compression argument is sound.
BMO’s head of U.S. rates said if 30-year yields push through 5.25%, there will be a “more durable pullback” in equity valuations
Markets are now pricing a 55% probability of a Fed rate hike by year-end.
• Speculators have built their largest net short position in Nasdaq 100 futures since 2023.
• S&P 500 gamma exposure is at extreme levels, risking a massive unwind if support breaks.
The biggest threat
Been bearish on SOXL lately, but short term (next 1-2 days) I’m actually rooting for the bulls here.
15m chart starting to look better:
- RSI climbing
- MACD turning green
- Momentum curling upward
- Buyers defending lows
Not calling for a massive breakout yet, but this looks like possible short-term momentum reversal instead of another straight dump.
If NVDA/news flow cooperates, bulls might finally get a little run here 🤝
📝 Not financial advice. Just following momentum.
Main story: broad semi weakness is still mostly macro/positioning, not one company-specific blowup. Today’s pressure is coming from higher Treasury yields, inflation/oil/Iran concerns, and traders de-risking ahead of Nvidia earnings Wednesday. Fresh semi-specific stories are Samsung strike risk in memory, Google/Blackstone launching a TPU-based AI cloud venture, and Nvidia earnings/China access uncertainty.
Micron (P/E: 32.9) Samsung and its union are still negotiating ahead of a possible 18-day strike; the strike could threaten memory supply, meanwhile Mizuho raised Micron’s target to $800 on strong DRAM/NAND pricing and AI-server demand.
Intel (P/E: N/A) Intel and Qualcomm have shown early takeover interest in Tenstorrent, an AI-chip startup. Benchmark raised Intel’s target to $140.
Nvidia (P/E: 54.0) is still the main market catalyst. Wednesday earnings will test whether Nvidia can maintain AI dominance as inference competition rises. There is also fresh China uncertainty: Jensen Huang said he expects China to open over time, but there was no immediate H200 approval breakthrough from China.
Analog Devices (P/E: 76.0) ADI is in advanced talks to buy Empower Semiconductor for about $1.5B cash, which would push ADI further into AI power-management chips for AI processors and data centers.
ASML Holding (P/E: ~49.1) has fresh High-NA EUV news. ASML expects the first chips made with its next-gen High-NA machines within months; Intel, Samsung, and SK Hynix are moving ahead, while TSMC still sees the machines as too expensive for now.
AMD (P/E: 134.5) got fresh analyst support but is still trading with the broader high-expectation AI unwind. Citi raised AMD’s target to $460 from $358 while keeping a Neutral rating.
Qualcomm (P/E: 21.0) is part of the same Tenstorrent story as Intel. The relevance is strategic: Qualcomm may be looking for more AI/inference exposure beyond phones and edge devices.
| Ticker | Company | SOXX weight | Today | Est. SOXX impact | Est. SOXL impact |
|---|---|---|---|---|---|
| MU | Micron | 9.35% | 4.73% | 0.44% | 1.33% |
| MRVL | Marvell | 5.85% | 6.31% | 0.37% | 1.11% |
| ALAB | Astera Labs | 1.71% | 11.43% | 0.20% | 0.59% |
| INTC | Intel | 6.71% | 1.94% | 0.13% | 0.39% |
| CRDO | Credo | 1.48% | 6.11% | 0.09% | 0.27% |
| NXPI | NXP Semiconductors | 3.64% | 1.52% | 0.06% | 0.17% |
| UMC | United Microelectronics | 0.72% | 4.35% | 0.03% | 0.09% |
| TXN | Texas Instruments | 4.01% | 0.63% | 0.03% | 0.08% |
| NVDA | Nvidia | 7.10% | 0.35% | 0.02% | 0.07% |
| ARM | Arm Holdings | 0.54% | 3.04% | 0.02% | 0.05% |
| SWKS | Skyworks Solutions | 0.62% | 1.72% | 0.01% | 0.03% |
| ADI | Analog Devices | 3.33% | 0.29% | 0.01% | 0.03% |
| STM | STMicroelectronics | 0.72% | 1.28% | 0.01% | 0.03% |
| TER | Teradyne | 2.84% | 0.26% | 0.01% | 0.02% |
| MTSI | MACOM | 1.40% | 0.32% | 0.00% | 0.01% |
| TSM | TSMC | 2.74% | 0.00% | 0.00% | 0.00% |
| MCHP | Microchip Technology | 2.88% | -0.01% | 0.00% | 0.00% |
| RMBS | Rambus | 0.77% | -0.61% | 0.00% | -0.01% |
| KLAC | KLA Corporation | 3.26% | -0.20% | -0.01% | -0.02% |
| ASML | ASML Holding | 2.63% | -0.30% | -0.01% | -0.02% |
| ENTG | Entegris | 1.12% | -1.36% | -0.02% | -0.05% |
| ASX | ASE Technology | 0.95% | -2.13% | -0.02% | -0.06% |
| NVMI | Nova | 0.91% | -2.35% | -0.02% | -0.06% |
| LRCX | Lam Research | 3.36% | -0.89% | -0.03% | -0.09% |
| ON | ON Semiconductor | 2.51% | -1.28% | -0.03% | -0.10% |
| AMAT | Applied Materials | 4.69% | -0.71% | -0.03% | -0.10% |
| AMD | Advanced Micro Devices | 8.87% | -0.58% | -0.05% | -0.15% |
| MPWR | Monolithic Power Systems | 3.68% | -1.84% | -0.07% | -0.20% |
| AVGO | Broadcom | 7.45% | -1.63% | -0.12% | -0.36% |
| QCOM | Qualcomm | 4.05% | -3.07% | -0.12% | -0.37% |
| SOXX | iShares Semiconductor ETF | 100.00% | 0.95% | 0.95% | 2.84% |
| SOXL | Direxion Daily Semiconductor Bull 3X Shares | N/A | 2.50% | N/A | 2.50% |
My Take: SOXL really dug itself out of hole just in time for the post lmao. While I post trailing P/E just because it's the standard it's worth noting ahead of NVDA earnings that their forward P/E is only ~25 right now leaving a lot of room to grow, which may sound weird to say for the largest company in the world but the fact is plenty of the other Trillion dollar companies have higher P/E's notably Walmart at 45 and Tesla at 190, which would easily justify the 50% growth needed to reach my infamous price target:
Main story: broad semi selloff driven by macro pressure, not one company-specific blowup. The fresh broad-market issue is still rates, oil, inflation, and Iran-war uncertainty. This morning brought some relief as yields and oil pulled back, but tech was still one of the weaker groups. Friday’s market was hit after Trump left China with no major trade breakthrough and no tangible China help on ending the Iran war. Nvidia earnings on Wednesday could be very important to the market overall.
Analog Devices (P/E: 77.0) is one of the few green names. Fresh coverage over the weekend framed ADI as benefiting from AI data-center power, optics, and test exposure ahead of its May 21 earnings. That helps explain why ADI is holding up better than most of the group today.
Intel (P/E: N/A) has fresh analyst support, but the stock is still weak. Citi reportedly raised its Intel target to $130 from $95, and Benchmark raised its target to $140 from $105, but Barron’s said the stock was still on track for a fourth straight down day after a massive AI-driven run.
Lam Research (P/E: 51.8) actually had positive fresh news. Morgan Stanley upgraded LRCX to Overweight, citing improved NAND forecasts, shipment outperformance, margin expansion drivers, and positioning in DRAM/packaging transitions. The stock is still down because the whole equipment group is getting hit.
Nvidia (P/E: 54.6) has fresh bullish analyst commentary ahead of earnings Wednesday. KeyBanc and D.A. Davidson reportedly raised their Nvidia targets to $300, citing Blackwell/Rubin demand and Nvidia’s supply-chain positioning. Nvidia earnings on Wednesday is the next major catalyst for the AI chip trade.
AMD (P/E: 136.7) also got a fresh Citi target hike to $460 from $358, but it is still trading down with the broader AI/CPU semi unwind. The same Barron’s piece tied AMD, Intel, and Arm to the AI CPU/server growth trade, but today the market is selling high-expectation chip names anyway.
Applied Materials (P/E: 42.5) has the cleanest fresh negative company-specific item. Morgan Stanley downgraded AMAT to Equal-weight, saying 2027 growth looks more in line with the market, China share concerns remain, and the valuation discount may not narrow near term. That makes AMAT’s drop look partly company-specific, not just sector selling.
Micron (P/E: 33.1) is tied to the fresh Samsung strike update. Samsung and its union are extending talks to avoid a major strike that could disrupt memory-chip supply chains. A court also partially limited strike actions, which reduced the worst-case disruption risk, but the strike threat is still relevant for memory names. MU being the largest SOXX holding makes even a modest drop a major SOXL drag.
| Ticker | Company | SOXX weight | Today | Est. SOXX impact | Est. SOXL impact |
|---|---|---|---|---|---|
| ADI | Analog Devices | 3.18% | 0.44% | 0.01% | 0.04% |
| SWKS | Skyworks Solutions | 0.55% | 0.65% | 0.00% | 0.01% |
| ARM | Arm Holdings | 0.53% | -0.31% | 0.00% | 0.00% |
| UMC | United Microelectronics | 0.67% | -1.44% | -0.01% | -0.03% |
| STM | STMicroelectronics | 0.71% | -1.74% | -0.01% | -0.04% |
| NXPI | NXP Semiconductors | 3.43% | -0.48% | -0.02% | -0.05% |
| TXN | Texas Instruments | 3.85% | -0.85% | -0.03% | -0.10% |
| RMBS | Rambus | 0.76% | -4.42% | -0.03% | -0.10% |
| MCHP | Microchip Technology | 2.82% | -1.28% | -0.04% | -0.11% |
| ENTG | Entegris | 1.15% | -3.14% | -0.04% | -0.11% |
| NVMI | Nova | 0.97% | -4.62% | -0.04% | -0.13% |
| MTSI | MACOM | 1.41% | -4.58% | -0.06% | -0.19% |
| ASML | ASML Holding | 2.65% | -2.45% | -0.06% | -0.19% |
| QCOM | Qualcomm | 3.72% | -1.78% | -0.07% | -0.20% |
| ASX | ASE Technology | 0.99% | -6.94% | -0.07% | -0.21% |
| TSM | TSMC | 2.71% | -2.60% | -0.07% | -0.21% |
| KLAC | KLA Corporation | 3.29% | -3.02% | -0.10% | -0.30% |
| ON | ON Semiconductor | 2.54% | -4.13% | -0.10% | -0.31% |
| LRCX | Lam Research | 3.39% | -3.11% | -0.11% | -0.32% |
| AVGO | Broadcom | 7.29% | -1.76% | -0.13% | -0.39% |
| TER | Teradyne | 2.95% | -4.77% | -0.14% | -0.42% |
| ALAB | Astera Labs | 1.70% | -8.55% | -0.15% | -0.44% |
| NVDA | Nvidia | 7.05% | -2.07% | -0.15% | -0.44% |
| MPWR | Monolithic Power Systems | 3.74% | -3.95% | -0.15% | -0.44% |
| CRDO | Credo | 1.63% | -10.48% | -0.17% | -0.51% |
| AMD | Advanced Micro Devices | 8.88% | -2.12% | -0.19% | -0.56% |
| INTC | Intel | 6.73% | -2.95% | -0.20% | -0.60% |
| AMAT | Applied Materials | 4.68% | -5.83% | -0.27% | -0.82% |
| MRVL | Marvell | 5.92% | -5.38% | -0.32% | -0.96% |
| MU | Micron | 9.97% | -5.33% | -0.53% | -1.59% |
| SOXX | iShares Semiconductor ETF | 100.00% | -3.14% | -3.14% | -9.43% |
| SOXL | Direxion Daily Semiconductor Bull 3X Shares | N/A | -9.78% | N/A | -9.78% |
My Take: When overall SOXX news is positive and overall market news is negative and there's a red day, that obviously isn't good but I'm not too worried. Not every day can be green, not every month can be April. More than likely this will look like a slight divot in the overall bull market. On the off chance there were a recession caused by the broad market forces driving the market down today then I'm confident all of these companies would do fine.
SOXL to $500 EOY
Found this analysis from a trader I follow and thought it was worth sharing with this community. Translating and paraphrasing his commentary:
SOXL continued selling off today, dropping more than 10%. He took partial profits on his bearish position but did not try to bottom-fish.
His macro framework right now is what he calls the “Three Highs”:
Is it better to sell at a loss or wait for possible rebound and then sell?
Do we have a bottom or resistance threshold at 140?
My guess is our new bottom resistance must be around 140x
What the lowest you all think it could go this month?
110?
Or perhaps it could reclaim its glory and go higher?
SOXL closed at 20% drawdown today.
When it goes down another 10% or above I will start DCA with little amount.
After that when it goes another 20% I will DCA more amount than before but I don't think it will happen next 2 to 3 weeks.
NVDA and AVGO earnings pending.
Bird poop
Hot CPI
Trump tweets
NVDA uncertainty
Buy the dip bro. 😭
Lower Highs ↓
Lower Lows ↓
175
↓
170 ← rejection
↓
165 ← another rejection
↓
160
Mu’s barrier stronger than semiconductor demand itself