The market just survived what felt like a geopolitical panic attack — Iran / Strait of Hormuz chaos — and is now in that awkward post-crisis recovery where nobody fully trusts the bounce. Meanwhile, a brutal multi-month software and SaaS selloff is either bottoming out or about to get worse, and earnings season kicks off right now (ASML, TSMC, Netflix this week, Goldman already reported today).

Three stories are running simultaneously:

The AI infrastructure stocks are recovering — Nvidia, Broadcom, Alphabet, Marvell up 6–14% in recent days. The SaaS and software basket is still a bloodbath — ServiceNow down 43% YTD, Adobe down 35%, Palantir down 28%, Zscaler down 47%. And the macro backdrop remains uncertain — Iran ceasefire is "fragile" at best, the Fed is still hawkish, and the Strait of Hormuz is only partially open.

The week's central tension: is the software selloff a buying opportunity or a value trap? Several smart analysts say yes, the market is pricing worst-case. The market keeps proving them wrong by going lower anyway. We'll be honest about that tension throughout.

Live and unresolved as of April 14.

ASML and TSMC earnings. This is the biggest single data point of the week, full stop. Multiple fresh sources flag these reports as the make-or-break moment for the entire AI capex supercycle thesis. TSMC raises guidance, semiconductors rip. TSMC guides down, every AI stock takes another leg lower. If you have semiconductor exposure heading into these prints, you should have a plan for both outcomes.

Netflix earnings. Cited by multiple sources as an imminent momentum catalyst. The subscriber + pricing power story is still intact at ~330 million premium subscribers, with 2026 revenue guided to $12.6B per quarter. A beat signals risk-on. A miss in this environment gets punished disproportionately.

Iran ceasefire status. Still genuinely unresolved. A "ceasefire" was announced, markets rallied hard, then it emerged the Strait was still partially blocked, Lebanon strikes continued, and Iran claimed violations within 24 hours. Crude oil volatility is the direct transmission mechanism to inflation fears, energy stocks, and consumer sentiment. Don't treat this as settled.

Goldman Sachs Q1 beat (already reported today). Second-highest revenue, EPS, and net income in firm history. Record equities performance, sharp rebound in investment banking. The punchline: Goldman specifically benefits from volatility, so their blowout quarter is not necessarily a signal that the broader market is healthy. Still, the investment banking backlog being described as "extraordinarily robust" is a quiet bullish signal for M&A-adjacent names.

The SaaS divergence question. Earnings from ServiceNow, Salesforce, Adobe, and Snowflake are coming in the next few weeks. The setup question every analyst is asking: will this earnings season finally force the algorithms to stop selling all software as a monolith and start distinguishing AI winners from AI losers?

AI infrastructure is still the core trade. High conviction. Virtually every credible source from the past week agrees: Nvidia, Broadcom, Alphabet, Marvell, and Micron are the foundational AI plays. The Anthropic $11B Broadcom chip order, Nvidia's Vera Rubin sampling, Micron's HBM4 dominance while Korean competitors scramble for helium — this narrative is tight, consistent, and fresh. Multiple independent sources reaching the same conclusion without copying each other is about as good a signal as you get in this business.

SaaS is oversold on fundamentals, but sentiment is broken. Medium conviction. The CFA analysts, the European DCF channel, and several deep-dive pieces all agree that the market is pricing worst-case AI disruption on companies still growing 20–56% with expanding margins. Adobe's 10% free cash flow yield, ServiceNow at 19x non-GAAP earnings needing only 6.7% annual growth to justify its price — the math looks compelling. The problem: the market doesn't care about your math right now. Cheap has kept getting cheaper for months. This is a "probably right, definitely early" situation.

Microsoft is disconnected from fundamentals and eventually breaks the IGV correlation. Medium-high conviction. Three separate fresh sources independently conclude that MSFT at $370–382, trading at 21x forward earnings while growing EPS at 22%, is absurdly mispriced relative to its history. The cited trigger is a meaningful AI product announcement. The Azure backlog growing 50% to $600B, with 65% of it unrelated to OpenAI — that's a business the market has somehow decided to value below where it traded before the OpenAI partnership began. That's a category error, and eventually category errors get corrected.

Palantir's fundamentals don't justify the selloff. Medium conviction. Tom Nash's breakdown is genuinely thorough: $4.5B revenue growing 56%, operating margins expanding at 191% year-over-year, $7.2B cash vs $400M debt, 98/100 business scorecard. Michael Burry's disruption thesis requires revenue slowdown and margin compression — neither of which appears in the data. The Trump Truth Social endorsement on April 10 shows political tailwind. But sentiment is working against you, and sentiment doesn't care about your spreadsheet.

Bitcoin has a precedent-based risk of further downside before the real bottom. Medium conviction. Josh Nichols' cycle analysis shows that in the comparable prior cycle, after the same MACD cross-up setup, Bitcoin fell another 38% before the final bottom — implying a potential move toward $46K from ~$73K. The invalidation is clear: two consecutive weekly closes above $75,200 (Saylor's DCA average) flips the thesis from bearish to bullish. This is the most internally consistent, well-defined bear case in the batch.

Nike is a value trap. Don't touch it. High conviction. Four separate sources independently conclude this, which is as close to a consensus short as you'll find. No growth. Margin compression. Losing share to On, Hoka, and Adidas. Discounting product at wholesale after abandoning the DTC strategy. New CEO "win now" plan is mid-execution with no proof it works. When four different analysts who don't talk to each other all say the same thing about a stock, maybe believe them.

Nvidia short-term: bull vs. bear. The AI infrastructure bull piece has NVDA at 56.5% upside over 12 months. The MarketBeat Q2 red flags piece warns that the put/call ratio has flipped bearish, helium supply constraints could cause a production miss, and the trailing-to-forward P/E gap creates execution risk. Both can actually be right simultaneously — Nvidia could disappoint in Q2, dip, and still hit 12-month targets. The helium angle is genuinely novel and not widely discussed. Worth watching TSMC's comments on input costs.

Palantir: AI winner or AI loser? Michael Burry argues Anthropic is eating Palantir's lunch. Tom Nash argues the fundamentals show zero evidence of disruption. One retail educator cited "Anthropic holds 73% of new AI business spending" — a stat that sounds made-up or badly sourced, and should not be acted on without verification. The most interesting data point in the whole batch: an actual Claude trading bot reportedly bought ServiceNow this week because it identified itself as "infrastructure for AI agents, not a victim of them." If the company supposedly disrupting ServiceNow is buying ServiceNow, someone's thesis is wrong.

Tesla: Q2 trade risk vs. long-term hold. Josh Nichols is explicitly buying below $300 as a long-term position while trading puts short-term. MarketBeat's analyst is bearish for Q2: EV demand weak, mega pack business slumping, 50,000 unsold cars sitting on lots, Q1 sales achieved via 0% interest and subprime lending. These subprime-financed vehicles could come back as returns. The disagreement is purely time horizon — the long-term bulls acknowledge the near-term problems and don't care. The near-term bears don't have a view on whether Optimus matters in 2029.

SaaS bottom timing. The European analyst (April 14, same-day): "I still don't think we're out of the woods yet" — but reverse DCF shows Salesforce and ServiceNow priced for near-zero growth. ServiceNow deep dive: "10–15% more downside possible before too cheap to ignore." Multiple other analysts are already buying. Everyone sees the valuation case clearly. Everyone disagrees on whether the bottom is in or 15% lower. The earnings catalyst over the next four weeks is the most likely resolver.

The Iran ceasefire (already resolved by reality). Multiple sources from April 10–12 described a deal that appeared settled. As of April 14, the ProfG Markets piece makes clear the "deal" was confused from the start. The Strait wasn't fully reopened. Lebanon strikes continued. This disagreement has been resolved: the ceasefire is fragile, not clean.Treat any geopolitical calm signal with a 24-hour confirmation window before acting.

Every thesis below gets a time label, a trigger, and an honest invalidation. No exceptions.

⚡ ACT TODAY — window closes within 24 hours

ASML / TSM — Earnings catalyst positioning. Guidance from TSMC and ASML this week is binary for the entire AI semiconductor space. If TSMC guides up, buy the dip on anything semi-related. If they guide down, step back from the whole basket. Don't guess — wait for the print and react. Trigger: earnings report mid-week. Invalidation: guidance cut or capex pullback signal. Conviction: High. Source freshness: 🟢

META — Muse Spark momentum continuation. The stock is already up 8%+ on the AI model launch. The DCF at $660 vs. $618 market price, plus revenue re-acceleration to 25% in 2026, plus the return on invested capital at 2x+ cost of capital — the fundamental case for holding or adding here is strong. The risk is a profit-taking reversal after the run. Trigger: already in motion, monitor for continuation vs. reversal. Invalidation: Muse Spark gets universally panned by users, or Meta cuts 2026 capex guidance. Conviction: High. Source freshness: 🟢

NFLX — Earnings this week. Subscriber base growing to 330M+ and 2026 quarterly revenue guide of $12.6B. A beat here is a risk-on signal for the whole market. A miss in this environment gets punished hard. Position sizing accordingly heading into the print. Trigger: earnings report. Invalidation: subscriber miss or weak 2026 revenue guide. Conviction: Medium. Source freshness: 🟢

📅 THIS WEEK — 2–5 day window

MSFT — One announcement from a breakout. Trading at 21x forward earnings while growing EPS at 22%, with Azure growing at 38% vs AWS at 19%, and a $600B backlog growing 28% ex-OpenAI. The catalyst is any AI product announcement that forces the market to stop lumping MSFT in with the IGV software selloff basket. Multiple sources independently put the buy zone at $350–370 and a target of $550+. Trigger: AI product launch or Azure contract win, breakout above $382–390 resistance. Invalidation: No catalyst materializes, stock stays correlated to IGV. Conviction: High. Source freshness: 🟢

AVGO (Broadcom) — Hyperscaler custom silicon dominance. Anthropic $11B chip order, five hyperscaler ASIC customers, VMware cash machine, first 2nm custom AI chip shipping now. The business is arguably better positioned than Nvidia for the custom silicon wave. Trigger: TSMC earnings confirm capex intact, watch for breakout above $370 toward $385+. Invalidation: Hyperscaler capex cuts or Anthropic delays order. Conviction: High. Source freshness: 🟢

NVDA — Wait for TSMC, then decide. Long-term thesis unchanged: center of gravity for every AI dollar, Vera Rubin sampling underway, $230 price targets widely cited. Short-term risk: helium supply constraints, put/call ratio flipping bearish, Q2 execution risk. Don't add ahead of TSMC's print. React to guidance. Trigger: TSMC confirms capex intact + put/call ratio normalizes. Invalidation: TSMC cuts guidance, helium disruption worsens. Conviction: Medium (short-term), High (12-month). Source freshness: 🟢

GOOG — Beneficiary of almost everything. 14% Anthropic stake, Azure-beating cloud growth, Waymo at 500K paid rides per week, Gemini at 18% US market share up from 5%. Up 7.5% last week, but still down 6.8% from 52-week high. The analyst consensus target cluster around $390–400 vs. current $317. Trigger: pullback to $310–315 is the entry, or any Gemini/Waymo expansion announcement. Invalidation: Search market share erosion accelerates meaningfully. Conviction: High. Source freshness: 🟡

MU — Technical range trade with fundamental tailwind. Bounced off the bottom of the range, targeting the top at $445 from ~$418. The Korea/helium disruption story gives Micron a structural pricing power advantage that's multi-quarter, not multi-week. Trigger: close near $418 by Friday → move toward $445. Invalidation: break below $390 support, or broader semi selloff on TSMC miss. Conviction: Medium. Source freshness: 🟡

SKY / CVCO (manufactured housing) — Road Act binary catalyst. The Road to Housing Act passed the Senate 89-10 and is back in the House potentially for a vote this week. Manufactured homes are the stealth winner — the bill would remove 50-year-old restrictions on the industry. Champion Homes (SKY) and Cavco (CVCO) both popped in February on the House version, then faded back. A vote this week is a catalyst. Trigger: House floor vote, potentially this week. Invalidation: Congress delays, or manufactured home provisions get stripped. Conviction: Medium. Source freshness: 🟢

PLTR — Oversold bounce watch. Fundamentals remain exceptional. The Trump April 10 endorsement showed the political floor is real. 4-hour RSI going below 30 is the oversold signal. Don't buy on vibes — wait for the RSI trigger or an earnings confirmation. Trigger: 4-hour RSI below 30 + technical stabilization. Invalidation: Confirmed contract losses to Anthropic, or revenue growth deceleration in next earnings. Conviction: Medium. Source freshness: 🟡

🗓️ THIS MONTH — patience required, thesis live

NOW (ServiceNow) — Waiting for earnings to unlock. 98% retention rate, 20% revenue growth, $28B remaining performance obligation, margins expanding not contracting. Trading at 38x GAAP or 19x non-GAAP. The argument that seat compression will destroy margins has been explicitly addressed by management — and the data shows active user base growing 25%. The Claude trading bot buying ServiceNow while it's supposedly being disrupted by Anthropic is a genuinely interesting data point. But 10–15% more downside is possible if the market continues pricing GAAP. Start small, leave room to average down. Trigger: Earnings report + management confirms per-seat model stability. Invalidation: Revenue below 15%, or actual margin compression appears. Conviction: Medium. Source freshness: 🟡

ADBE — Priced for extinction that isn't coming. 10% free cash flow yield. 97% recurring revenue. 12% revenue growth reaccelerating. 9x non-GAAP forward earnings. Semrush acquisition adding 48% ARR growth and already generating AI revenue that tripled in a year. The company is buying back 6–7% of market cap annually. Average historical P/E was 30x; it traded at 51x at the peak. The market is now pricing the worst-case. Trigger: Earnings hold above current revenue growth trend, Semrush AI ARR exceeds $50M. Invalidation: Revenue growth decelerates below 8%, or Canva steals material enterprise share. Conviction: Medium. Source freshness: 🟡

HIMS — Transitioning from GLP-1 compounder to healthcare platform. The AI operating system announcement is directionally significant even if details are thin. Post-Zavva and Eucalyptus acquisitions, 4–5M subscribers makes this the largest healthcare subscription app in the world. The Eli Lilly oral GLP-1 pill distribution deal is reportedly incoming. If they become the aggregator — the "app you click when you have healthcare questions" — the 10x upside case gets real. Trigger: AI feature rollouts over "next few months" + Eli Lilly partnership announcement. Invalidation: FDA enforcement against compounding; Eli Lilly deal goes to competitor. Conviction: Medium. Source freshness: 🟡

BTC — Bear case with defined invalidation. Historical precedent from the prior cycle: after this exact MACD setup (cross-up with two prior legs down), Bitcoin fell another 38% before the final bottom — implying a potential move toward ~$46K from $73K. The "if then" is critical: two consecutive weekly closes above $75,200 flips the thesis from bearish to bullish. If BTC drops to $46–50K, that's the load-up zone. If it holds above $75,200, the bear thesis is wrong — pivot and go long. Don't be a perma-bear hog. Trigger: Weekly close below $70K OR above $75,200. Conviction: Medium. Source freshness: 🟡

MRVL (Marvell) — Asymmetric AI infrastructure pick. The Nvidia NVLink Fusion partnership for optical interconnects is the kind of exclusive positioning that shows up in 10-year returns. Data center revenue is now 74% of the business, up from 40%. Celestial AI acquisition adds another layer. Smaller and more volatile than Broadcom, but potentially more upside if the Nvidia partnership scales. Trigger: TSMC/ASML guidance confirms AI infra spending, Nvidia partnership revenue ramp H2. Invalidation: Nvidia renegotiates partnership; Celestial integration fails. Conviction: Medium. Source freshness: 🟢

SOFI — Digital bank at value pricing. Upgraded to Buy at $17.50 on March 19, now at $16.50 — $1 lower than the upgrade. 30% member growth guidance for 2026. Legacy bank branch erosion continues playing into SoFi's hands. 27x forward P/E looks genuinely reasonable for a company with this growth profile. The key insight: SoFi doesn't need the big banks to fail — it just needs them to keep cutting branches and understaffing physical locations. Trigger: Member growth report next quarter. Invalidation: Credit quality deterioration, or legacy banks accelerate digital pivot meaningfully. Conviction: Medium. Source freshness: 🟢

NBR (Nabors) — Oil drilling services, binary on April 29 earnings. PE of 5x in oil drilling services. Iran = elevated oil prices = accelerating drilling services demand. Already trading above consensus targets, but Citigroup and Susquehanna both raised targets above current price in April. The April 29 earnings print will be the moment to either add or exit. This is a geopolitically-sensitive trade, not a structural one. Trigger: April 29 earnings. Invalidation: Iran ceasefire becomes real and oil drops sharply. Conviction: Low — only appropriate as a small, geopolitically-hedged position. Source freshness: 🟡

TSMC/ASML guidance cut — THIS WEEK. Highest-probability single-event risk on the calendar. If Taiwan Semi guides down citing demand softness or supply constraints, every AI infrastructure stock reprices lower simultaneously. The entire bull thesis on Nvidia, Broadcom, Marvell, and Micron hinges on hyperscaler capex commitments remaining intact. One bad earnings call unwinds weeks of recovery in a single session.

Iran re-escalates — ONGOING, UNPREDICTABLE. The ProfG Markets piece is the most honest on this: nobody, including people physically in the room with Trump, can predict what he does next. The ProfG piece also notes suspicious trades around the ceasefire announcement that haven't been fully explained. If the Strait of Hormuz closes for months rather than weeks, energy inflation returns, the Fed can't cut, and growth stocks take another multiple compression hit.

SaaS earnings deliver the bear case — NEXT 4 WEEKS. If ServiceNow, Salesforce, or Snowflake report actual seat compression, slowing consumption, or lost contracts to AI agents — the market's worst fears get confirmed and the next leg down in software begins. The software bulls are making a fundamentals argument that only works if the fundamentals hold. Earnings season is the test.

Fed pivot doesn't come — 1–3 MONTHS. Multiple sources implicitly assume some form of Fed accommodation arriving to rescue growth stock multiples. Goldman's "final hurdle" comment about the Fed transition is a quiet warning. Sticky energy inflation could prevent cuts entirely. Without multiple expansion, the software valuation argument only works if the businesses compound their way to reasonableness — which takes years, not weeks.

Bitcoin breaks through $60K — 1–2 MONTHS. If Nichols' bear case plays out, crypto fear contagion hits risk appetite broadly, particularly for retail investors holding both crypto and growth stocks simultaneously. Not a macro-level crisis, but a retail sentiment shock.

On the SaaS buying opportunities, we're probably misjudging how long institutional algorithms will continue selling all software as a monolith before distinguishing winners. "Cheap can get cheaper" is not just a cliché — ServiceNow looked like the obvious buy to multiple smart analysts when it was at $110, then $95, then $90. Being early by 3 months and 25% in a bearish sentiment environment is functionally identical to being wrong. Patience and small position sizing are the only honest responses to this situation.

On the AI infrastructure bull case, we might be underestimating the helium supply shock. If Korean fab costs spike 30%+ and production timelines slip, analyst 12-month targets become detached from near-term reality. Backlogs don't generate revenue — production does.

On the macro situation, the "market already priced it in" argument is compelling but historically unreliable during genuine geopolitical escalation. The forward P/E falling from 23x to 19x is progress, not an all-clear. If the Strait of Hormuz disruption lasts longer than markets are pricing, the earnings growth narrative gets revised down and the multiple compression continues.

The meta-mistake we're most likely making: treating a geopolitical situation as resolved when it isn't, and front-running an earnings-season rally that hasn't happened yet. Goldman's record quarter was powered by volatility — that's a very specific kind of bull market that doesn't lift all boats.

Not financial advice. Probabilistic thinking, scenario planning, and risk management only. If you act on any of this and lose money, that's between you and your brokerage account.