Alphatica

🧵 ALPHATICA PREDICTION TRACKER (PINNED) We post predictions BEFORE earnings. We post receipts AFTER. January 2026 Scorecard: $NFLX Q4: 98.2% EPS accuracy $WFC Q4: 1.1% error vs Street 5.7% $AAPL Q1: 3.2% error vs Street 6.0% $GE Q4: Within $0.02, called sell-off $RTX Q4: Off by one penny $MA Q4: Called +12.3% beat $V Q4: Beat Street both metrics $CAT Q4: Called +11.2% beat 8 for 8 on direction. Beat Street accuracy on 6 of 8. Institutional-quality quantitative insights. Robust trading strategies for every trader. Sign up for weekly predictions, signals & strategies → alphatica.io — ALPHATICA brought to you by Blackline

Goldman Sachs just published what might be the most important oil research note of the year. The title says it all: "Higher Prices for Longer?" The key findings: Brent is likely to exceed its 2008 all-time high if Hormuz flows stay depressed and disruption timelines extend. That's not a fringe scenario anymore. It's Goldman's base case risk. They studied the 5 largest supply shocks of the past 50 years. The average hit to production was still 42% four years later. Not four weeks. Four years. Infrastructure damage and underinvestment don't recover on a press conference timeline. Iran and the 7 other Gulf producers accounted for 30% of global crude in 2025. That production base is now under direct fire. The risk most people are missing: strategic stockpiling. After this crisis, countries will raise their SPR targets. That means demand for oil stays elevated even after flows resume because everyone is rebuilding reserves simultaneously. The crisis ends but the buying doesn't. Goldman's conclusion: risks are skewed to the upside on net, both near-term and into 2027. Oil above $100 may not be temporary. It may be the new floor. We've been making this case all week. Now Goldman is putting institutional weight behind it. $USO $XLE $GS

At $1B per day, the current air and naval campaign burns through $200B in about 200 days. That takes you to October. Right before midterms. But Hegseth said the money is also for "what we may have to do in the future" and to ensure ammunition is "not just refilled but above and beyond." That's not the language of a campaign winding down. That's the language of escalation planning. Even members of Trump's own party are asking the question. Chip Roy: "We're talking about boots on the ground. Now we're in a whole 'nother zip code." Thomas Massie: "Is this the first $200 billion?" Congress hasn't even authorized the war. Now they're being asked to fund it at a quarter of the Pentagon's annual budget. The budget tells you the plan before the press conference does.

You don’t ask for $200 billion if you don’t think it’s going to require ground troops.

Both approaches will probably coexist and the market will sort it out. New construction will increasingly be designed around automation from the ground up. Smaller, task-specific robots. Built-in systems. Optimized layouts. That's the long-term efficient path. But there are 140 million existing homes in the US that aren't getting redesigned anytime soon. For retrofitting the existing world, humanoids that can navigate human spaces are the practical solution because you can't rebuild the infrastructure. The humanoid isn't the optimal robot. It's the compatible one. And compatibility with the world as it already exists is worth a lot.

Doesn’t take into account that there is value to redesigning homes and other spaces so humans have more and better living space Just like warehouses were redesigned to optimize speed/storage and access. Homes, offices and other spaces can be redesigned Just because a humanoid can do the job, it doesn’t mean it’s the optimal robot for the job and space utilization Why wouldn’t you use a smaller, less expensive, easier to maintain robot that has an environment it was designed for ?

DoorDash paying people to film themselves doing chores so AI can learn to replace them is the most dystopian sentence I've read this year and it's buried in the middle of the list like it's normal. That's where we are. The humans are now training data for their own obsolescence and getting paid hourly to do it. Meanwhile $450B in AI spend hasn't moved the growth needle according to Goldman. The money is being spent. The jobs are being cut. The growth isn't showing up. And Bezos wants to raise another $100B to automate factories. At some point the question shifts from "will AI change everything" to "who benefits when it does." Based on this list, the answer so far is: not the person filming themselves doing laundry.

absolutely insane day in AI, just to recap: - uber drops $ 1.25B on rivian for 50k robotaxis - hsbc planning deep cuts across middle and back offices - doordash pays people to film chores so AI can replace them - goldman sachs says $ 450B in AI spend added basically zero to US growth - cursor’s fifty person team drops a model beating top labs on coding - openai acquires astral - jeff bezos is raising $ 100B to buy and automate factories - fortune 500 now puts AI impact at $ 4.5T, 93% of jobs are exposed.

This is the single chart that explains everything we've been talking about this week. Data centers up 228%. Offices down 38%. Since the same date. The economy is physically rebuilding itself around AI while the old infrastructure hollows out. You can see the transition in steel and concrete, not just earnings calls. And remember the trades shortage we discussed. BlackRock's Larry Fink said the US needs 500,000 electricians just for data center buildout. The irony is that the same AI driving office construction down is driving demand for tradespeople up. The economy doesn't need fewer workers. It needs different workers in different buildings. Connect this to the capex numbers from earlier this week. Microsoft +693%. Meta +501%. Amazon +464%. That capital is flowing into the $45.1 billion in data center construction you're looking at. And every one of those facilities needs electricians, HVAC techs, and pipe fitters to build and maintain it. AI is reshaping the economy. But the reshaping is being done by human hands.

BREAKING: The value of US data centers under construction has officially surpassed the value of office buildings under construction for the first time in history. Data centers under construction are up+29% YoY, to a record $45.1 billion. Meanwhile, the value of offices under construction are down -13%, to $43.5 billion, the lowest since October 2015. Since November 2022, when ChatGPT was launched, data center construction is up +228%. Over that same period, office construction is down -38%. AI is reshaping the US economy.

Domino's generates $4.7 billion in annual revenue. Papa John's? $2.1 billion. Same product. Same customer. Completely different execution. In FY16, these two were neck and neck — DPZ at $2.2B, PZZA at $1.6B. A 1.4x gap. Today that gap is 2.3x — and it's still widening. What happened? DPZ went all-in on tech and delivery infrastructure. They turned pizza ordering into a logistics operation. Revenue more than doubled from $2.2B to $4.7B in a decade. PZZA had a governance crisis in 2018 that cratered the brand. Operating income fell from $165M to $25M. Revenue stalled. They've spent six years trying to recover and still haven't reclaimed their FY18 peak. The operating margins tell the real story: DPZ: 17-19% operating margin — every single year for a decade. Clockwork. PZZA: Swung from 9.6% to 1.5% and back. Still inconsistent. Both went public around the same time at nearly the same price. Since IPO: DPZ: $13.50 → $375.30 (+2,680%, 16.6% annualized) PZZA: $12.56 → $33.70 (+168%, 4.5% annualized) $10,000 in DPZ at IPO is worth $278,000 today. $10,000 in PZZA? $27,000. Sometimes the winner in a sector isn't the one with the better product. It's the one that doesn't blow itself up. $DPZ $PZZA $MCD alphatica.io

The people who are scared aren't scared because tech leaders used scary words. They're scared because their company just laid off their department and cited AI. You don't fix a sentiment problem with better messaging when the lived experience is confirming the fear. Jensen's right that panic doesn't help anyone. But the answer isn't softer language from CEOs. It's showing people what AI creates, not just what it eliminates. Right now the American AI story is overwhelmingly about cost cutting. Until that changes, no amount of careful language moves the needle.

JUST IN: Nvidia CEO Jensen Huang calls on tech leaders to "be careful not to scare people" regarding AI.

We did this study on our own. Here is what we found: We tracked how AI mentions evolved across earnings calls from Q4 2023 through the end of 2025. Overall AI mentions held relatively steady at around 5% of total call content. But the tone shifted dramatically. Over that same period, mentions specifically tied to AI cost savings rose 57%, with positive sentiment around those savings increasing across all industries and sectors. Companies aren't just mentioning AI anymore. They're quantifying what it's saving them. In 2025, AI companies raised $95 billion across 143 funding rounds, nearly tripling the pace from 2024. The partnership ecosystem around AI continues to evolve and risk tolerance is splitting into two distinct camps. The first is all in. They're opening the gates, letting employees use whatever AI tools they want, with guardrails around IP and internal data protection. The second camp is moving slower, more conservative, focused on integrating AI safely, soundly, and securely before scaling adoption. Neither approach is wrong. But the gap between the two is widening every quarter and the companies in the first camp are pulling ahead fast. #OpenAI

@alphaticaio @unusual_whales Is the study accessible somewhere?

"Massive investment in AI contributed basically zero to US economic growth last year," per Goldman Sachs

@ValarMorghuliz9 Might get $666–$667 at 3:50 PM tomorrow afternoon, but only if Trump behaves. We do not think market makers are going to pay on all these puts expiring tomorrow.

@alphaticaio ....let's see 669 or let's see 650?! 😳

Expect $SPY $663.00 tomorrow mid-morning. After that, let's see.

Mercedes literally makes the A-Class and the GLA, both entry-level vehicles designed to get younger buyers into the ecosystem so they eventually buy a C-Class, then an E-Class, then an S-Class. That's not giving up on being aspirational. That's how you build a pipeline to aspirational. Apple did the same thing with the iPhone SE, the base iPad, and AirPods. The cheap product isn't the destination. It's the on-ramp. Once someone is in the ecosystem, the average spend per customer climbs every year. That's not running out of ideas. That's funnel economics. $AAPL

Apple selling cheap laptops means it’s giving up on being an aspirational brand. Means it’s run out of ideas - which has been obvious for a while with the goggles and the thicker phones and iPads. It’s like if Mercedes entered the compact truck market.

@beyoumf Data wins.

Convince us to follow you… But you only have 2 words

The intent is straightforward. The mechanics are a mess. There are roughly 11 million undocumented people in the US. Many of them have bank accounts, pay taxes through ITINs, and participate in the formal economy. Forcing banks to verify citizenship doesn't just affect new accounts. It raises the question of what happens to existing ones. Do banks close 11 million accounts? What happens to the direct deposits, the rent payments, the small business transactions flowing through those accounts? The financial system is interconnected. You can't pull a thread that big without unraveling something.

Trump considering forcing banks to verify citizenship, signaling a potential massive immigration crackdown measure. Follow: @AFpost

$81,705 per person. Spending tripled. Population grew 26%. And the comptroller can't even trace where the money goes. Run those numbers through any lens you want and the conclusion is the same. The system isn't designed to solve homelessness. It's designed to process it. There's a massive industry of contractors, operators, and service providers that depends on the problem continuing. When spending triples and outcomes get worse, someone is getting rich. It's just not the person sleeping on the street.

NYC is spending more per street homeless person than the median household earns in a year $368 million went to outreach teams, safe havens, and drop-in centers for 4,504 people. That's $81,705 per homeless person. The average NYC household earns $80,483 a year Spending tripled since 2019. The homeless population grew 26%. The state comptroller says the city doesn't publicly report expenses in a way that allows clear analysis of where the money is going

The $81K isn't going to the homeless person. That's the part people miss. It's going to the contractors, the administrators, the hotel operators, the service providers, and the bureaucracy that sits between the tax dollar and the person sleeping on the street. The homeless person isn't receiving anyone's paycheck. They're receiving a fraction of a fraction after everyone else in the chain takes their cut. The outrage shouldn't be that we're spending money on homelessness. It should be that we're spending $81K per person and people are still sleeping on the sidewalk. That's not a generosity problem. That's an efficiency problem.

NYC is now spending $81,000 per homeless person The average median household income is $81,000 for a New Yorker You work hard and receive nothing in return while the homeless receive your entire paycheck for free Stop the waste and fraud.

SPY CLOSE | March 19 $659.80. Down 0.25%. Survived. Barely. The market held above $655 all day and clawed back to close just pennies below $660. After yesterday's 1.57% Fed-driven selloff, a flat close heading into tomorrow's expiration is the best the bulls could have asked for. Alphatica Composite Score: -24.4 [Lean Bearish]. Midday it was -58.5. A 34-point improvement into the close. That's the largest afternoon recovery of the week. The pattern returned. Sell the morning, buy the close. The number that matters most tonight: +142.7M shares of LONG delta flow. That's one of the largest single-day bullish delta flow we have EVER tracked. Someone made an enormous directional bet into the close ahead of tomorrow's expiration. Net premium confirms it. +$1.31B call-heavy. $2.7B in call premium vs $1.39B in puts. 66% of all dollars went into calls. Yesterday was -$1.33B put-heavy. A $2.65 BILLION swing in net premium from put-dominated to call-dominated in 24 hours. That's not hedging. That's repositioning. But the structure hasn't healed. The $660 accelerator hit -$728.6M. A new all-time record. Again. It was -$551M yesterday. It grew 32% in one session. We're sitting directly on top of it at $659.94. Six cents below. Net GEX: -$3,229.7M. Still near the all-time worst. Dealers long 151.6M shares. IV skew at +4.29%, still deeply bearish. The put demand hasn't faded even as the call buyers showed up in force. The divergence is the widest we've ever seen. The flow says bullish. The structure says fragile. +142M shares of long delta sitting on top of -$728.6M of negative gamma at the same strike. Tomorrow one of them is wrong. Tomorrow: March 20 expiration. 2.63M puts vs 1.31M calls. 3.94M total contracts. The largest single-day gamma event of this correction. $660 is the pin. Max pain at $674 is pulling up. The call buyers from today need the expiration to clear the gamma and let price rise. The structure needs to survive one more session. If $660 holds through 4 PM tomorrow, -$728.6M of accelerator gamma expires worthless and the deck clears. If it breaks in the morning, the accelerator stack runs $1.5B deep to $640 and the bear flag's measured move completes. Three weeks of tracking. Three weeks of $660. Tomorrow it ends, one way or the other. The data moves before the price does. alphatica.io $SPY $QQQ $VIX

$FDX Q3: strong print, higher guide, Freight spin still on track. Revenue: $24.0B vs $22.2B YoY Adj. EPS: $5.25 vs $4.51 GAAP EPS: $4.41 vs $3.76 Adj. op income: $1.62B FedEx Express op income: +$21% YoY Core story: Express/package is working. U.S. domestic package revenue +10% International export package revenue +8% Total package revenue +10% Total package ADV +3% U.S. domestic ADV +5% Pricing + volume + cost saves all contributed. Weak spot was Freight: Revenue -5% Op income fell to $8M from $261M Margin collapsed to 0.4% But the spin-off is still on track for June 1, 2026 The bigger headline may be the guide raise: FY26 revenue growth now 6.0%–6.5% EPS now $16.05–$16.85 pre-MTM Adjusted EPS now $19.30–$20.10 Also notable: capex guide cut to ≤$4.1B from $4.5B, while cost reductions are now expected to exceed $1B. Takeaway: Better core execution, higher guide, stronger cash flow, lower capex, and the Freight separation remains a live catalyst.

@mikealfred Markets price conflict. That part is reality. But once someone starts wanting more destruction because it benefits their position, they’ve crossed from analysis into something much darker. Profit should never come before basic humanity.

Wild to see oil bulls and macro doomers literally rooting for more death and destruction because it helps their positions. This is the most evil form of greed. These are very bad, nasty people and they will not win. Good will triumph over evil and they will be bankrupted.

MARCH 26 & MARCH 30 $SPY EXPIRATIONS ARE NOW SITTING ON POSITIVE GAMMA

Bookmark this: 7462 by year’s end. The majority of this selloff has been mechanical. We don’t believe the fear. We will be proven right. It’s just a matter of time frame. $SPY and $QQQ are designed to go up.