Trading Desk · Command Center

Sell the chip, buy the electron — today's AI selloff is funding a fresh power/nuclear trade.

Mixed/churning, mild risk-off tape (S&P -0.4%, Nasdaq -1.0%, semis -3.9%, VIX +4.3%). The stale watchlist all reads NEUTRAL — so the edge today is a FRESH idea off live news, not a re-rank. Three desks independently found the same door: the money leaving chipmakers is flowing toward the thing that powers the AI buildout — electricity.

2026-07-27 · Committee · dynamic opportunity hunt · live prices (Yahoo) · desks last met this session
1 · What the market's saying
The weather right now — indices, the fear gauge, the regime, and the week's landmines.
S&P 500 +0%Nasdaq -0%Semis -2%VIX 19 +0%
Mixed / churning — no clear risk signal. Be selective; don't force a bet.
🟡 Seasonally soft — May–Oct is the weaker half of the year, and the toughest months (Aug–Sep) are just ahead. Keep some dry powder.
Wed Jul 29
FOMC + Warsh presserHawkish-hold base case; a rate HIKE is the live tail. The single biggest risk to the book this week.
Wed Jul 29
MSFT + META earningsHyperscaler capex is the spine of the power thesis — a cut breaks it.
Thu Jul 30
AAPL + AMZN earningsMore hyperscaler capex signal.
Wed Aug 5
SOFI earnings (your holding)Below trend into a print — the committee says EXIT.
Thu Aug 6
CEG earningsThe core opportunity name reports here — a Sept spread spans it. Capped-risk structure matters.
Fri Aug 7
VST earnings + Jobs reportVST (the CEG alternative) reports; monthly jobs the same day.
Tue Aug 18
OKLO earningsThe speculative kicker reports — tiny size means a write-off is a shrug.
Tue Aug 25
CRWD earningsThe watchlist software name — a post-earnings buy candidate, not before.
2 · The call
The committee's calibrated read — structured disagreement, not consensus. One action, sized to your cash.
⭐ Opportunity of the day · confidence MEDIUM
Buy the companies that sell electricity to AI data centers
Chip stocks (like Nvidia) are falling today because investors are nervous about who's going to pay for all the AI spending. But here's the thing everyone's missing: AI data centers need a HUGE amount of electricity to run. So instead of chasing the falling chip stocks, you buy the power companies that sell that electricity — they win whether or not any single chipmaker does. The best one is a nuclear-power company called Constellation (ticker CEG).
Do this — step by step
  1. Free up cash first: sell your three dead positions — UBER, SOFI, and FUBO. That turns stuck money into about $10,000 of usable cash.
  2. WHEN to buy (exact): watch CEG's price. GREEN LIGHT the moment it trades above $275.60 (its 10-day high) and holds — that's the momentum breakout. OR, if instead it dips to about $255 and holds there, buy that pullback. Do NOT buy while it's just drifting between ~$258 and $275 (no edge), and don't buy at all if it closes below $236 (the idea's broken).
  3. HOW to buy (exact order): in thinkorswim, open CEG → the Sep 18 options → build a 'call vertical' / debit spread: buy the 270 call, sell the 280 call (one ticket). Enter it as a LIMIT order at $4.60 debit — that's $460, your max loss. If it won't fill, nudge up, but never pay more than $5.04, and NEVER use a market order (the bad 'natural' price is $5.70). Cheaper alternative: the 280/290 spread ≈ $415.
  4. Optional small long-shot (~$150): buy one cheap OKLO Sep call — a tiny nuclear startup, a lottery ticket sized so losing it all wouldn't sting. Also a LIMIT order. Only add it if you keep the CEG core cheaper so the total stays near ~$500.
  5. SIZE: 1 CEG spread = $460, which fits your ~$4,080 cash with plenty left. Keep the rest dry — especially through Wednesday's Fed.
The trade — grounded in live data (patterns · trend · real option price · fundamentals)
TickerCEG
When (trigger)GREEN LIGHT when CEG breaks and HOLDS above $275.60 (its 10-day high) — the momentum entry. OR if it dips to ~$255 and holds (the 20-day) — the pullback entry. NO entry while it drifts between ~$258–275 (no edge). Stand down if it closes below $236.
OrderBuy 1 CEG Sep-18 270/280 call VERTICAL as a LIMIT order @ $4.60 debit ($460 max loss). Won't pay above $5.04. LIMIT, never market — the 'natural' worst fill is $5.70, don't pay that. Legs: 270c ~$22.80 / 280c ~$18.20.
StopA close below ~$236 (the recent swing low) — the idea's wrong there.
Option (real)Sep 18 · 270/280 call debit spread · $460 max loss · max gain $540 · breakeven $274.60 · IV 47% (live, marketdata.app)
PatternsNo candlestick pattern firing today (TAAPI) — this is a trend/rotation entry, not a pattern trigger.
FundamentalsP/E 24 · profit margin 13% · $98B market cap · beta 1.1 · Utilities (Alpha Vantage)
ExitsTake profit ~double; cut if it loses the 50-day or a hyperscaler cuts capex Wed/Thu.
Why a "spread"?A plain call is expensive and bleeds value fast. A 'debit spread' (buy one call, sell a higher one) is cheaper because the call you sell pays for part of the one you buy. The trade-off: your profit is capped — but your loss is capped too, at what you paid. That capped loss is exactly why it's safe to hold through CEG's Aug 6 earnings, which would otherwise be a wild card.
When to get outTake profit if the spread roughly doubles. Bail immediately if Microsoft or Amazon CUTS its AI spending at earnings this Wed/Thu — that would break the whole reason for the trade.
The desk's original (trader-speak) version
Long the power/nuclear rotation as chips sell off: core = a CEG September call debit spread (~$300–350 max loss, VST as the swap if CEG's chain is thin); kicker = a tiny ~$150 OKLO September call. Enter on any hold of today's levels; keep the rest of the cash dry through Wednesday's Fed.
⚠ Biggest risk to the whole book
A hawkish surprise from the Warsh Fed on Wednesday 7/29 — specifically a rate HIKE (a live tail off 3.50–3.75%). It's the one event that hurts almost everything at once: it spikes real yields (already ~2.43% and climbing), compressing growth stocks AND rate-sensitive utilities — the one weak spot in the power trade. Immediately followed by the mega-cap earnings gauntlet (MSFT+META Wed, AAPL+AMZN Thu): if a hyperscaler CUTS capex, the power-rotation thesis breaks. Hence small, defined-risk, capped-loss only this week — max ~$700 total new risk, nothing naked into a print.
CEGConstellation Energy (core)$270.07 -2% today · RSI 77
OPPORTUNITY — starterconf MED
📅 earnings in 10d (2026-08-06) — inside the danger window
Highest-quality nuclear name powering data centers, and it's HOLDING today while chips bleed — the rotation in action. Short-duration contracted cash flow is the opposite of what rising rates punish, and the catalyst is live TODAY, not weeks out.
What changes our mindA hyperscaler cuts capex Wed/Thu → the whole theme breaks, stand down. If the group already gapped up hard, wait a day.
The riskEarnings Aug 6 — a Sept debit spread spans the print. Use the SPREAD (capped loss), don't buy naked premium; don't oversize.
💵 Sizing to your $4,080 cash: ~15 shares (~$4,051), or a small defined-risk spread.
Real option (live): 2026-09-18 270/280 call debit spread ≈ $460 max loss, IV 47% — confirm in thinkorswim.
OKLOOklo (variance kicker)$41.83 +4% today · RSI 35
SPECULATIVE — tinyconf LOW
📅 earnings in 22d (2026-08-18)
Small-cap nuclear = maximum torque for a win-only game. Green today while chips fell. A ~$150 lotto sized so a total write-off is a shrug.
What changes our mindOnly add AFTER CEG/VST confirm the group is being bought, not bleeding.
The riskSpeculative and volatile; earnings Aug 18. Tiny size is the whole risk control.
💵 Sizing to your $4,080 cash: ~97 shares (~$4,058), or a small defined-risk spread.
Real option (live): 2026-09-18 40/45 call debit spread ≈ $180 max loss, IV 91% — confirm in thinkorswim.
CRWDCrowdStrike$180.10 -2% today · RSI 40
WATCH — triggerconf LOW
📅 earnings in 29d (2026-08-25)
The only genuinely clean uptrend (+63% 3mo, above 50/200-day) — but MACD is rolling over and it walks into its own Aug 25 earnings + the real-rate grind. Good name, bad moment.
What changes our mindHolds its 50-day (~$179) AND clears the Aug 25 print → a post-earnings buy. Loses $179 → stand down.
The riskLong-duration software into rising real yields.
GLDGold$374.58 +1% today · RSI 47
WATCH — triggerconf LOW
The only green name today and the only one with MACD rising — a possible early trend turn and the one non-stock diversifier. But still below both moving averages.
What changes our mindClears/holds above ~$385 with real yields rolling over = the turn is real. Back under $368 = headfake.
The riskRising real rates are a live headwind for gold.
NVDANvidia$196.56 -5% today · RSI 50
AVOID — falling knifeconf MED
📅 earnings in 30d (2026-08-26)
−5% today on the OpenAI $250B 'circular-financing' headline, below its 50-day, MACD down, earnings Aug 26 still ahead. An active, unpriced overhang.
What changes our mindThe financing fear gets digested AND it stabilizes above ~$190.
The riskCatching it here is knife-catching into an open news story.
QQQNasdaq 100$682.05 -0% today · RSI 37
AVOIDconf MED
RSI 37, below its 50-day, tech leading the selloff — buying the epicenter of today's weakness in an index wrapper.
What changes our mindTech stops leading down; MACD turns up.
The riskNo asymmetry — you'd be long the exact thing that's falling.
3 · The idea behind it
The structural findings, and each desk's voice — so you can see who's saying what.
  1. The AI selloff is a fight about WHO PAYS for the buildout, not WHETHER it happens. Hyperscalers (MSFT/AMZN/GOOG) are still spending up; the new bottleneck is electricity, not chips. So the exact panic hammering the chipmakers makes the power companies more valuable — today's red tape is the entry, not a knife. Three desks (Catalyst, Devil's Advocate, Opportunity) found this door independently.
  2. Why power beats waiting for CRWD: CRWD is long-duration software into rising real rates AND its own earnings. Power/nuclear is short-duration, contracted cash flow — the opposite of what rising rates punish — and the catalyst is live today, not four weeks out.
  3. Four of today's five watchlist names (NVDA/QQQ/SPY + CRWD's timing) are the same tech/semi beta getting knifed. NVDA is down ~5% specifically on an OpenAI $250B circular-financing headline — an open, unpriced story. Don't buy the epicenter.
  4. The real-data feed caught the catch: CEG earnings Aug 6, VST Aug 7 (Finnhub). A Sept spread spans them — the reason to use a defined-risk debit spread, not naked calls. Honest confidence on the whole rotation is MEDIUM: it's one day old.
  5. This week is the highest-variance stretch of the quarter (FOMC Wed + MSFT/META Wed + AAPL/AMZN Thu). Place asymmetric bets, not hero bets: max ~$700 total new risk, all capped-loss, nothing naked into a print or the Fed.
Market Scannerlive movers
Whole board reads NEUTRAL/red today — semis −3.9% leading down, GLD the lone green. The clean momentum names aren't buyable here; the edge is off-list.
Economistregime · real rates
Rising real yields (~2.43%) are the axis de-rating long-duration growth. Favors short-duration, contracted-cash-flow assets — which is exactly the power/utility complex — over rate-sensitive software.
Catalyst / Geopoliticalnews → plays
NVDA's drop traces to an OpenAI $250B circular-financing story. The durable read underneath: AI electricity demand is the bottleneck → power & nuclear names. FOMC Wed + hyperscaler earnings are the week's swing events.
Setup Screenertechnical setups
No clean breakout in the tech complex today. The setup that IS working is relative strength in power/nuclear and gold against a falling tape — rotation, not breakout.
Devil's advocateattacks the idea
The one real hole in the power trade: a hyperscaler capex CUT at Wed/Thu earnings would break the 'buildout continues' spine. Also rate-hike risk hits utilities. Both reasons to size small and defined-risk — not to skip it.
Risk officersizing · survival
Permits ~$500 across the opportunity (CEG spread ~$300–350 + OKLO ~$150), max ~$700 total new risk this week, all capped-loss, nothing naked into the Fed or a print. Makes the opportunity survivable rather than blocking it.
4 · Your book
Your account: $4,080 cash · ~$24,669 total. Each position is babysat below — RIDE / HOLD / CUT updated on live momentum.
PYPL PayPal200 sh · -0% today · +11% 3mo · $11,214
RIDE
Trending up and strong — let it run. Raise your stop as it climbs.
UBER Uber100 sh · +3% today · -9% 3mo · $6,818
CUT
Below its trend and heading the wrong way — dead money. Free the cash for a mover.
SOFI SoFi100 sh · +3% today · -8% 3mo · $1,688
CUT
Below its trend and heading the wrong way — dead money. Free the cash for a mover.
FUBO FuboTV100 sh · +4% today · -28% 3mo · $869
CUT
Below its trend and heading the wrong way — dead money. Free the cash for a mover.
Guardrails & blind spots
Risk guardrails (non-negotiable)
  • Small, defined-risk, capped-loss structures ONLY this week — highest-variance week of the quarter
  • Max ~$700 total new risk across all new positions
  • Opportunity of the day ~$500: CEG spread ~$300–350 + OKLO ~$150 (speculative)
  • Use debit SPREADS on the power names — their earnings (CEG Aug 6, VST Aug 7) fall inside a Sept expiry
  • Nothing naked into an earnings print or across Wednesday's FOMC
  • Add the OKLO kicker only after CEG/VST confirm the group is bid
  • Stagger expiries; keep dry powder for after the Fed
What we can't see (no faked precision)
  • Live option prices / implied volatility / skew on CEG, VST, OKLO — so use spreads and do NOT overpay premium; the sizing is directional logic, not a measured fill
  • Real-time order flow and breadth internals confirming the rotation is more than one day old
  • The actual FOMC outcome and whether a hike is really on the table (published odds only)
  • Hyperscaler capex guidance before Wed/Thu — the swing factor for the whole thesis
  • Power-contract pricing detail; ranges only, never point targets
A command center for a fake-money school competition — live market data + a multi-desk committee's reasoning, refreshed when the desks meet. Game strategy, not investment advice, and not a prediction. Verdicts are calibrated bets with stated confidence; what actually hits is variance.