SharkWater Trading  •  Nuclear Desk • Small Modular Reactors • Sector Selloff

Three Reasons, One Bad Friday: The SMR Complex Loses 16 to 23 Percent in a Week

September 12, 2026

Bottom Line Up Front

On Friday, September 11, 2026, NuScale (NYSE: SMR) fell 15.67 percent after UBS downgraded it to Sell with a price target cut to $6.00 from $10.00. Oklo (NYSE: OKLO) fell 9.18 percent the same day after filing an 8-K for a new $1 billion at-the-market share program, replacing one it had just fully spent at an average price of $55.64 a share. X-Energy (Nasdaq: XE) fell another 5.74 percent with no new catalyst at all, just the hangover from Piper Sandler's September 9 Sell call.

Three different mechanisms, same week, same sector. Since Monday's close, OKLO is down 16.4 percent, XE is down 22.0 percent, and SMR is down 23.0 percent. This is not one company's bad news anymore. It reads like a sector repricing, and the risk sits in different places for each name.

NuScale: UBS Puts a Number on the Doubt

UBS downgraded NuScale to Sell from Neutral on September 11, 2026, cutting its price target to $6.00 from $10.00. This note itself was not independently reviewed here; the rationale below is sourced to press coverage of the note (Benzinga, Yahoo Finance, 24/7 Wall St.), not the UBS document directly, and is labeled as such.

The reasoning, as reported: a construction timeline stretching past five years, roughly $700 million of projected cumulative cash burn over three years against a cash and investments position of $1.9 billion as of Q2 2026 (a figure the company itself has disclosed), a lack of firm customer commitments, setbacks on the Romania RoPower project, and limited progress on a Tennessee Valley Authority agreement. The stock closed at $8.61 on September 11, down from $10.21 on September 10, a one-day move of 15.67 percent. Some coverage suggests the decline accelerated through the session rather than opening at that level, which fits a note landing mid-day and repricing the stock in real time.

Oklo: The ATM Machine Refills Itself

Oklo filed an 8-K on September 11, 2026 (accession 0001104659-26-106897, Items 1.01/1.02/9.01) disclosing a new Equity Distribution Agreement for up to $1.0 billion in at-the-market Class A common stock sales, with agent commissions of up to 1.5 percent of gross proceeds. The same filing terminates Oklo's prior $1 billion ATM, dated May 13, 2026, which the company had already fully used: approximately 17.97 million shares sold for roughly $1 billion gross, at an average price of $55.64 a share.

A boat that keeps needing to refuel isn't necessarily sinking. But if it comes back to the dock for another full tank every few months, you start asking how far it's actually getting on each one.

The stock closed at $36.22 on September 11, down 9.18 percent from Thursday's $39.88. That is a steep one-day drop for a financing announcement rather than an operational setback, and it suggests the market is pricing in a pattern, not a one-time event. Oklo remains a pre-revenue company funding a capital-intensive build-own-operate strategy; a second consecutive $1 billion ATM inside roughly four months is dilution investors should expect to keep seeing as long as that model holds, priced or not.

X-Energy: Still Bleeding From Last Week

X-Energy closed at $14.93 on September 11, down 5.74 percent from Thursday's $15.84, with no new filing or company news found for the day. This extends the decline that started with Piper Sandler's September 9-10 note, which rated Oklo Buy and X-Energy Sell in the same call, arguing that X-Energy's asset-light licensing model shifts project execution risk onto its customers. X-Energy is now down roughly 22 percent since Monday's close with nothing new driving Friday's leg lower. That is arguably the most uncomfortable position of the three: no fresh catalyst, just continued selling on a thesis that has not been rebutted.

The Week in Numbers

Ticker	Sept 8 close	Sept 10 close	Sept 11 close	Friday move	Since Sept 8
OKLO	$43.31	$39.88	$36.22	-9.18%	-16.37%
XE	$19.15	$15.84	$14.93	-5.74%	-22.04%
SMR	$11.18	$10.21	$8.61	-15.67%	-22.99%

Source: stockanalysis.com (aggregator, not a live execution feed). Oklo's September 11 8-K is the one primary-source filing behind this table; the UBS note and its rationale are relayed through secondary press coverage, not the note itself, and should be treated accordingly.

The Bull Case

• The underlying demand thesis hasn't moved. Nothing this week changes AI-driven data center power demand or federal policy support for advanced nuclear. This looks like a repricing of execution risk, not a repudiation of the sector's premise.
• Oklo's dilution is disclosed, not desperate. A fully-utilized ATM followed immediately by a new one is a financing choice by a well-capitalized, pre-revenue builder, not a distress signal. The company chose to raise growth capital in public markets rather than debt or a distressed placement.
• A three-name, one-week drawdown of this size can overshoot. Sector-wide selloffs on stacked bad news often run further than the news alone justifies, especially in thinly-traded, story-driven names. A stabilizing catalyst, such as the South Korea-US nuclear cooperation package firming into a signed agreement, could reverse sentiment quickly.

The Bear Case

• UBS just put a specific, quantified bear case on paper for SMR. A five-year-plus construction timeline against $700 million of projected cash burn is a concrete claim, not vague caution, and it came from a sell-side desk willing to attach a $6 price target to it.
• Oklo's capital needs look structural, not one-time. Two consecutive $1 billion ATMs in about four months, on a company with no revenue, is a pattern worth taking seriously. Every future raise dilutes existing holders at whatever price the stock happens to be trading.
• X-Energy has no offsetting catalyst. A stock falling on no news, purely as leftover momentum from an analyst call nearly a week old, is a stock where nobody has stepped in to defend a price level. That is not a floor being built; it is an absence of buyers.

The SharkWater Take

Three sourced, dated catalysts hitting three related stocks in one week is not a coincidence I'm willing to wave off as noise. This looks like the market repricing execution risk across the whole SMR group at once, and I don't think it's finished. I would not treat Friday's drop in any of these three as a dip to buy. Oklo's back-to-back ATMs are the one I'd watch closest: dilution that repeats on a fixed schedule stops being a one-time financing event and starts being a cost of doing business, and that cost comes straight out of existing shareholders. If there's a name here worth revisiting, it's whichever one shows the first sign of a catalyst breaking the pattern, a signed customer contract, a completed construction milestone, or the South Korea package actually getting signed rather than just "nearing." Until one of those shows up with a primary source behind it, I'm watching this complex from the dock, not the boat.

Tight lines. — SharkWater

Educational and informational purposes only. Not personalized investment advice. All figures sourced as noted and accurate as of publication; the UBS rationale is relayed through secondary press coverage, not the original note. Options involve substantial risk of loss. The author may hold positions in securities discussed. Do your own work.

SharkWater Trading  •  Nuclear Desk • Analyst Calls • Small Modular Reactors

X-Energy Is Down 17 Percent Since Piper Sandler Said Sell. Oklo's Buy Rating Isn't Holding Up Much Better.

September 11, 2026

Bottom Line Up Front

Piper Sandler initiated Oklo (NYSE: OKLO) at Buy with a $55 price target and X-Energy (Nasdaq: XE) at Sell with a $9 price target, in a note dated September 9. Since that close, XE has fallen from $19.15 to $15.84, down roughly 17 percent, while OKLO has fallen from $43.31 to $39.88, down roughly 8 percent, over the same three sessions.

Both stocks are still unwinding a September 8 nuclear-sector spike that multiple outlets called catalyst-free before Piper Sandler ever weighed in. The Sell call on X-Energy is behaving the way a Sell call should. The Buy call on Oklo is not yet behaving the way a Buy call should, and that gap is the actual story here.

The Call

Piper Sandler analyst Dimple Gosai split the small modular reactor sector down the middle. Oklo got a Buy rating and a $55 target, a little under 38 percent above Thursday's close. X-Energy got a Sell and a $9 target, roughly 43 percent below Thursday's close. The stated reasoning, per coverage of the note, comes down to financing structure: Oklo runs a build-own-operate model the analyst called "bankable by design," while X-Energy licenses its reactor technology to customers under an asset-light approach that shifts project execution risk onto those customers instead of keeping it on X-Energy's own balance sheet.

That is a real, testable thesis about who eats the risk when a first-of-a-kind reactor project runs long or over budget. It is not a comment on which company has the better technology, and this desk isn't in a position to referee that engineering question. What we can track is whether the market is pricing the difference the way the ratings imply it should.

What the Tape Says

So far, it is pricing one half of it. X-Energy has fallen in a straight line since the call, exactly what a Sell rating with 43 percent downside implied should happen if the market agreed. Oklo has also fallen every session since, including a 6.32 percent drop Thursday that accelerated rather than stabilized. A stock carrying a fresh Buy rating with 38 percent of implied upside is not supposed to be having its worst single day of the week two days after the call.

Ticker	Sept 8 Close	Sept 9 Close	Sept 10 Close	Cumulative
OKLO	$43.31	$42.57 (-1.71%)	$39.88 (-6.32%)	-7.9%
XE	$19.15	$17.26 (-9.87%)	$15.84 (-8.26%)	-17.3%

Source: stockanalysis.com, an aggregator, not a primary source. A separate aggregator (MarketBeat) puts XE's September 9 close at $18.03 (-5.8 percent) rather than $17.26 (-9.87 percent); the two did not reconcile this week, so treat the exact size of X-Energy's single-day drop as approximate. The direction and the relative underperformance versus Oklo hold under either version. Oklo's Thursday volume ran about 1.17 times its 20-day average, not a capitulation print.

A split rating is a bet that two boats caught in the same swell will handle it differently once the wave passes. Right now both are still pitching. Only one of them is supposed to be.

The One Filing That Matters

Away from the rating, the only fresh SEC paper on either name this week is a Schedule 13G filed September 4 by Jane Street Group, LLC, disclosing a new passive stake of 5.2 percent in X-Energy. It states no activist intent and no plan to influence control. That is a market-making and index-flow signal, not a fundamental one, and it predates the Piper Sandler call by five days. Nothing in EDGAR corroborates or contradicts the thesis itself for either stock.

The Bull Case

• Piper Sandler's own math still points up. A $55 target against Thursday's $39.88 close is real, stated upside from a firm that just put a differentiated thesis in writing, not a vague "AI nuclear theme" call.
• Oklo's decline hasn't come on panic volume. Thursday's drop ran close to average turnover. That is consistent with sellers unwinding the same September 8 spike everyone else is unwinding, not with new information breaking against the stock specifically.
• The policy backdrop keeps getting louder, even if it isn't confirmed. Reporting on a US-Korea nuclear cooperation package, now sourced to the Wall Street Journal and a Korean outlet and not just one anonymous-sourced piece, describes a deal in the $100 billion range covering up to eight reactors. It names no companies on this watchlist and remains tentative, but it is evidence the sector narrative has more than one leg to stand on.

The Bear Case

• A Buy-rated stock accelerating its losses two days after the rating is a bad look, not a rounding error. If Piper Sandler's differentiated thesis were actually driving the tape, Oklo should be outperforming, not just losing less.
• The whole rally this is unwinding was never explained in the first place. September 8's spike across Oklo, X-Energy, and NuScale was described by multiple outlets as catalyst-free, tied to short-covering and sector rotation rather than any company-specific news. A thesis about financing structure doesn't need to be wrong for a name to keep falling simply because the move it's riding on top of was never real.
• X-Energy's Sell rating implies another 43 percent of downside from here. A stock already down 17 percent in three sessions that still carries that much stated downside, from the same analyst who just called it, is not a name to catch on the way down without a reason the price has actually stopped falling.

The SharkWater Take

I believe the Sell call on X-Energy before I believe the Buy call on Oklo, and the price action is telling me exactly that. X-Energy is falling the way a stock falls when a real thesis lands on top of a fake rally. Oklo is falling the way a stock falls when the fake rally hasn't finished unwinding yet, Buy rating or not. Until Oklo actually decouples, holding a green day or outperforming X-Energy on a bad tape rather than just losing less of it, I'm not treating Piper Sandler's split as confirmed by the market, only asserted by the analyst. I'd rather watch X-Energy keep validating the bear side of this call from the sidelines than buy Oklo on a thesis the stock itself hasn't started agreeing with.

Tight lines. — SharkWater

Educational and informational purposes only. Not personalized investment advice. Price targets referenced above are Piper Sandler's, as reported, not SharkWater's own forecasts. All figures sourced as noted; several are aggregator-sourced (stockanalysis.com, MarketBeat) rather than exchange-primary and are labeled accordingly, and two aggregators disagree on X-Energy's September 9 closing figure. Options involve substantial risk of loss. The author may hold positions in securities discussed. Do your own work.

SharkWater Trading  •  Data Center Desk • Grid Infrastructure • IREN

IREN Clears a 2 Gigawatt Grid Milestone in Texas, and the Stock Gave the Pop Back in a Day

September 10, 2026

Bottom Line Up Front

IREN's own press release, dated September 8, 2026, says its Sweetwater Hub, 2,000 megawatts split between Sweetwater 1 (1,400MW) and Sweetwater 2 (600MW), was conditionally included as Base Load in ERCOT's Batch Zero interconnection process. The stock rose 5.04 percent to $46.93 that day, then gave it back and more on September 9, closing $45.37, down 3.32 percent, as a broader risk-off session hit nearly every capital-intensive name in the sector.

The word doing the most work in that release is conditional. ERCOT can still change its mind. IREN discloses no revenue number for Sweetwater at all, and a $6 to $7.5 billion annual recurring revenue figure now circulating in secondary coverage is an outside estimate, not company guidance. The more interesting fact might be what IREN chose not to say.

What Actually Got Announced

ERCOT's Batch Zero is part of the interconnection screening process that determines whether a large new load, like a data center campus, gets treated as base load or something less certain. Base load status matters because interconnection queues, not capital, are the real bottleneck on AI data center buildout right now. A project can have the money lined up and still sit for years waiting on a grid slot.

IREN's release states the classification directly: Sweetwater's combined 2,000 megawatts, across two phases, has been "conditionally included" as Base Load. The company's own language is explicit that "ERCOT's classifications remain conditional and subject to ongoing approval processes." The first slice, 300 megawatts gross out of Sweetwater 1, is targeted for delivery in the fourth quarter of 2027. No dollar figure appears anywhere in the release.

The company also states, again in its own words, that it only adds projects to its publicly announced portfolio "following the execution of the relevant grid connection agreements." That is a disclosure policy, not a footnote. It means the 5.8 gigawatts IREN has announced to date is a floor, not a ceiling, on whatever it actually controls in the interconnection queue.

A conditional base load classification is a boat waved up from the harbor waiting line to the loading dock. It still has to tie up and actually take on cargo before anything ships. But a boat still bobbing out in the queue never even gets that chance.

The Numbers, and the Number That Isn't There

Item	Figure	Status
Sweetwater 1	1,400MW	Company-disclosed, conditional ERCOT base load
Sweetwater 2	600MW	Company-disclosed, conditional ERCOT base load
Combined Sweetwater Hub	2,000MW (2GW)	Company-disclosed
First delivery slice	300MW gross, Q4 2027 target	Company-disclosed timeline
Disclosed total pipeline	5.8GW	Company-disclosed as of this release; likely understates actual queue position given the grid-agreement-first disclosure policy
Sweetwater annual recurring revenue	"$6-7.5 billion" (widely cited)	THEORETICAL. Sourced to secondary analysis (Motley Fool, Sept. 10), not IREN's own release, which contains no revenue figure

Source: IREN press release, "IREN's 2GW Sweetwater Hub Included as Base Load in ERCOT Batch Zero," GlobeNewswire, September 8, 2026, 07:01 ET. The revenue figure is explicitly not from that release and should not be treated as company guidance.

The Pipeline You Can't See

IREN's competitors in this space tend to disclose developmental capacity long before a grid agreement is signed. Hut 8, for comparison, has publicly disclosed 5.4 gigawatts of developmental capacity using that looser standard. IREN's policy of waiting for an executed grid connection agreement before naming a project means its public 5.8 gigawatt figure is almost certainly conservative relative to what it is actually pursuing in the queue. That is a real, if unquantifiable, positive. It is also unverifiable by definition, since the company will not name the rest of it until the agreements exist.

The Bull Case

• This is a real grid milestone, not a marketing claim. The interconnection queue is the actual chokepoint in this sector, and a base load classification, even a conditional one, is genuine progress that most competitors cannot point to on a specific, dated basis.
• The disclosure policy implies upside that isn't in the public number. Waiting for signed grid agreements before naming a project is a disciplined standard, and it means the disclosed 5.8 gigawatt pipeline is probably a floor, not the whole picture.
• There's a real date attached. A 300 megawatt first delivery targeted for the fourth quarter of 2027 gives the market something concrete to hold the company to, rather than an open-ended promise.

The Bear Case

• Conditional means not done. ERCOT's own language reserves the right to revisit the classification. Nothing here is a signed, binding grid connection agreement yet, and the company's own disclosure standard treats that distinction as the one that matters.
• There is no company-sourced revenue number, and the one being repeated everywhere isn't real. A $6 to $7.5 billion annual recurring revenue figure is circulating in secondary coverage attached to this announcement. IREN did not say that. Treating an analyst's extrapolation as company guidance is exactly the kind of number this desk won't repeat as fact.
• The market's own reaction argues against urgency. The stock popped 5 percent on the news and gave that back the very next session on unrelated macro pressure. If this were viewed as a genuine re-rating event, it likely would have held up better against a garden-variety risk-off day.

The SharkWater Take

I like the discipline in IREN's disclosure policy more than I like this specific announcement. Only naming capacity after a signed grid agreement is the right way to avoid the vaporware problem that shows up everywhere else in this sector, and it means the real pipeline is probably bigger than 5.8 gigawatts. But conditional is conditional, there's no revenue number attached to Sweetwater from the company itself, and the tape gave the entire move back within a day on news that had nothing to do with IREN. That combination tells me the market isn't convinced this is a re-rating event yet either. I'm not buying this specific headline. I'd want to see an actual executed grid connection agreement, or Sweetwater 2 confirmed with a real contracted revenue figure attached, before this turns into a position instead of a data point worth tracking.

Tight lines. — SharkWater

Educational and informational purposes only. Not personalized investment advice. All figures sourced as noted and accurate as of publication. Options involve substantial risk of loss. The author may hold positions in securities discussed. Do your own work.

SharkWater Trading  •  Income & NAV Desk • Closed-End Funds • Premium/Discount

VCX's Own Prospectus Admits a 76 Percent Premium, Then Asks You to Buy More Shares

September 9, 2026

Bottom Line Up Front

Fundrise Innovation Fund (NYSE: VCX) filed an N-2ASR shelf registration on September 9, 2026, and the prospectus states its own number: net asset value per share of $21.70 against a market price of $38.12 as of September 8, 2026, a premium of 75.67 percent. The same filing shows shares traded as high as $289.51 against an $18.97 NAV during the second quarter, a premium of roughly 1,426 percent.

The shelf lets the fund issue an indeterminate number of new shares on a continuous basis. That is good for the manager, who collects fees on a bigger asset base. It is a real risk for anyone buying today, because a premium of this size has nowhere to go but down over time, and the fund itself says so in its own risk factors.

What the Filing Actually Says

Fundrise Innovation Fund began trading on the NYSE on March 19, 2026. It holds a portfolio of private, venture-stage companies, priced the way any venture fund prices private holdings: by internal marks, not by a public tape. The shares, on the other hand, trade every day on a public exchange, priced by whoever wants in or out that day.

Those two prices are supposed to converge over time. They have not. The fund's own N-2ASR, filed today, puts the September 8 gap at 75.67 percent. It goes further and discloses the second-quarter range: a NAV of $18.97 against a trading band of $76.88 to $289.51, meaning the stock changed hands for as much as fourteen times what the fund said its underlying assets were worth.

The shelf itself registers new common shares and rights for sale on an "immediate, continuous or delayed basis," with no dollar cap stated in the prospectus. That is standard shelf language, but paired with a premium this size, it means the fund can keep printing shares near the inflated price for as long as buyers show up, growing assets under management without doing anything to the underlying portfolio.

A boat riding a swell looks like it is climbing. It is not gaining any real height over the seafloor. When the swell passes, it comes back down to the water it was always floating on.

The NAV-to-Price Gap, By the Numbers

Period	NAV / Share	Market Price	Premium to NAV
September 8, 2026	$21.70	$38.12	75.67%
Q2 2026 low	$18.97	$76.88	~305%
Q2 2026 high	$18.97	$289.51	~1,426%

Source: Fundrise Innovation Fund N-2ASR, filed with the SEC September 9, 2026 (accession 0001213900-26-098443). All three figures are the fund's own disclosures, not aggregator estimates. Not independently confirmed: whether the $21.70 September 8 NAV reflects a same-day mark or is carried forward from the fund's June 30, 2026 NPORT-P filing, which reported an identical figure. Treat the NAV as of-quarter until that is confirmed.

The Bull Case

• Real liquidity for a normally illiquid asset class. Venture-stage private company exposure is usually locked up for years. VCX trades every day on the NYSE, and that convenience is worth something to buyers who would otherwise have no way in or out.
• The premium reflects genuine demand for the underlying names. Retail access to pre-IPO companies like the ones VCX holds is scarce. Scarce access to a popular asset class often prices at a premium, the same way a hard-to-get concert ticket does.
• New capital from the shelf can go to work in the portfolio. If the fund deploys shelf proceeds into more private positions rather than just diluting existing holders, NAV per share itself could grow over time, even if the premium compresses.

The Bear Case

• A 76 percent premium is not a durable valuation, it is a crowd. The fund's own prospectus warns that closed-end vehicles "frequently trade at a discount from their net asset value." A fund trading at multiples of NAV is the same instrument, just further from home.
• Continuous share issuance at a premium is a one-way street for the manager, not for you. Every new share sold near $38 while NAV sits near $22 grows fee-generating assets under management. It does nothing to close the gap for shareholders already in, and it adds supply right as the stock has already been volatile.
• The NAV itself is a private mark, which cuts both ways. If the underlying venture portfolio is actually worth more than $21.70 per share, the premium is smaller than it looks. If it is worth less, or if any holding takes a markdown, the gap when it closes could be worse than 76 percent, not better.

The SharkWater Take

I am not buying VCX at a 76 percent premium to its own stated NAV, and I would not sell it short into it either, because a fund that has traded at fourteen times NAV before can stay expensive longer than a short position can stay solvent. This is not a setup, it is a warning label, and the fund wrote the label itself. The number worth watching from here is not the stock price, it is how much stock actually gets sold off this new shelf and at what premium. If Fundrise prices a real offering meaningfully below today's tape, that is the market doing the fund's job for it. Until then, this is a name to watch from the dock, not from the boat.

Tight lines. — SharkWater

Educational and informational purposes only. Not personalized investment advice. All figures sourced as noted and accurate as of publication. Options involve substantial risk of loss. The author may hold positions in securities discussed. Do your own work.

SharkWater Trading  •  Digital Infrastructure • Power • Sector Map

Two Ways To Own The AI Buildout, And Only One Of Them Needs A Tenant

September 7, 2026

Bottom Line Up Front

Everything called "AI data center infrastructure" is really two different businesses stacked on top of each other. Layer one is the powered shell landlord, mostly converted bitcoin miners, whose entire value turns on whether a creditworthy tenant signs. TeraWulf has 401 MW to Anthropic on a 20 year lease worth roughly $19 billion. Core Scientific reports $24 billion or more of contracted capacity. Cipher has two hyperscaler leases. Keel has zero signed leases as of this writing.

Layer two is the supplier, and it gets paid on aggregate build volume rather than on any one landlord winning. GE Vernova closed Q2 2026 with a $176 billion backlog and $24.2 billion of quarterly orders, up 88 percent organically. Quanta carries a $53.4 billion backlog. The risk is not the same risk. Layer one is a binary on a signature. Layer two is a bet on volume, priced accordingly.

The Stack, In Plain Terms

The binding constraint on AI compute right now is not silicon. It is energized land with a live grid interconnect. That is why a cohort of bitcoin miners suddenly became infrastructure companies. They already owned the scarce thing.

Layer one takes that power position and rents it as a powered shell or a colocation contract. Layer two sells the turbines, switchgear, transformers, chillers, and field labor that turn a permitted site into a building full of racks. Layer two invoices whether or not any individual layer one developer ever finds a tenant.

A landlord with no lease is a finished dock with no boat tied to it. The pilings are driven, the water is deep, the slip is legally yours. It earns nothing until somebody throws a line.

Layer One: The Landlords

Rank these by signed contracts, not by pipeline gigawatts. Pipeline is a claim. A lease is a filing.

Name	Contracted Status	Counterparty
WULF TeraWulf	401 MW critical IT, 20 yr, ~$19B; up to ~$33B with both 5 yr extensions. FY2025 leases totaled 522 critical IT MW.	Anthropic
CORZ Core Scientific	~590 MW, 12 yr take or pay, $10B+ revenue potential, 80% to 85% anticipated margin. $24B+ total contracted capacity.	CoreWeave, others
CIFR Cipher Digital	300 MW gross to Fluidstack (207 MW critical IT), 10 yr, ~$3.8B, with $1.73B Google backstop. Separate 15 yr, 300 MW AWS lease.	Fluidstack / Google, AWS
IREN IREN Limited	Not a lease. A 5 yr GPU cloud services contract, ~$9.7B through 2031, 20% prepaid, 200 MW critical IT at Childress. IREN also bought ~$5.8B of GPUs from Dell.	Microsoft
KEEL Keel Infrastructure	No signed lease. Three priority sites near full permitting, $819M liquidity as of Aug 7, 2026, negative $23.7M adjusted EBITDA per quarter.	None yet

Sources: TeraWulf Q2 2026 Form 8-K and Form 10-Q (filed 2026); Core Scientific Q2 FY26 earnings presentation, EX-99.2 to Form 8-K dated July 28, 2026; Cipher Mining Form 8-K exhibits dated September 25, 2025 and November 20, 2025; IREN Form 8-K exhibit dated November 3, 2025; Keel Infrastructure Q2 2026 results release dated August 10, 2026. The dollar value of the Cipher AWS lease is not stated in the Cipher release I reviewed and is carried in secondary coverage only. See data gaps below.

Note the IREN row carefully, because aggregators file it next to the others and it does not belong there. TeraWulf, Core Scientific, and Cipher are landlords. IREN sells compute. It owns the data center and the GPUs, which is why its headline contract value is so much larger per megawatt, and also why it carries hardware refresh risk the landlords do not.

What A Megawatt Actually Rents For

Headline contract values are not comparable until you divide by capacity and term. Do that and the spread gets interesting.

Contract	Value	MW	Term	$M per MW-yr
WULF / Anthropic	$19.0B	401	20 yr	2.37
CIFR / Fluidstack	$3.8B	207	10 yr	1.84
CORZ / CoreWeave	$10.0B	590	12 yr	1.41
IREN / Microsoft, gross	$9.7B	200	5 yr	9.70
IREN / Microsoft, net of GPU capex	$3.9B	200	5 yr	3.90
KEEL	None	0	n/a	0.00

DERIVED BY SHARKWATER. Simple division of stated contract value by stated capacity and initial term. Not a company reported metric. No discounting, no ramp schedule, no escalators. Megawatt definitions are not uniform across issuers: WULF, CIFR, and IREN figures reference critical IT load, the CORZ figure references leased power across five sites. Cross-check: Core Scientific separately discloses roughly $850M average annual colocation GAAP revenue on the CoreWeave contract, which computes to $1.44M per MW-yr against $1.41M above. IREN separately discloses ~$1.94B targeted annualized run rate on 200 MW, which computes to $9.70M per MW-yr and matches exactly.

The takeaway is not that TeraWulf negotiated better than Core Scientific. It is that a twenty year lease to a frontier lab and a twelve year take or pay to a neocloud are different instruments with different credit behind them, and the per megawatt spread is roughly 68 percent. When someone tells you a developer has "2 gigawatts of pipeline," that number is worth nothing until you know what a megawatt rents for and who is signing.

Layer Two: The Suppliers

These companies do not need to know which landlord wins. They ship into the aggregate.

Function	Names	Hard Number
Power generation	GEV, CAT	GE Vernova Q2 2026: $176B total backlog, $24.2B orders up 88% organically, gas equipment backlog and slot reservations 100 GW to 116 GW in one quarter, targeting 125 GW by year end. Electrification data center orders above $5B year to date.
Electrical distribution	ETN, POWL, HUBB, NVT	Eaton Q2 2026 revenue $8.53B, up 21.4%, adjusted EPS $3.15.
Power and cooling in the white space	VRT, MOD, TT, AAON	Vertiv 2026 guidance as of July 29: net sales $13.8B to $14.2B, adjusted diluted EPS $6.65 to $6.75. Deferred revenue $1.815B at Dec 31, 2025 to $3.634B at Jun 30, 2026.
Field construction and MEP	PWR, EME, FIX, MTZ, DY	Quanta Q2 2026: revenue $9.56B, record backlog $53.4B, RPO $33.6B, 2026 revenue guidance raised to $39.3B to $39.7B.
Baseload power owners	CEG, VST, TLN	Signing PPAs directly with hyperscalers to bypass interconnect queues.

Sources: GE Vernova (NYSE: GEV) second quarter 2026 Form 8-K dated July 22, 2026; Quanta Services (NYSE: PWR) second quarter 2026 results, Exhibit 99.1 to Form 8-K dated July 30, 2026; Vertiv (NYSE: VRT) guidance update dated July 29, 2026 and June 30, 2026 balance sheet. Eaton (NYSE: ETN) Q2 2026 figures are from secondary coverage of the July 31, 2026 release and were not verified against the filing. Baseload power row is descriptive and carries no cited figure.

The Vertiv Line Nobody Read

Vertiv is the cleanest single expression of the buildout, at roughly three quarters of revenue from data center customers. It is also the best illustration of why headline numbers deserve suspicion.

In Q2 2025 the company led its release with backlog: $8.5 billion, book to bill about 1.2 times. Entering 2026 it was pointing at roughly $15 billion. The Q2 2026 release, published July 29, contains no backlog figure and no orders figure at all. Revenue of $3.274 billion came in under consensus, and the stock fell as much as 17 percent that day.

Here is the line three rows below the headline. Deferred revenue, which is cash customers have already paid for equipment not yet delivered, went from $1.815 billion at December 31, 2025 to $3.634 billion at June 30, 2026. Backlog is a management defined metric. Customer cash on the balance sheet is not. Those two facts point in opposite directions and the second one is audited.

I am not resolving that for you. I am telling you the disclosure changed and that the reappearance of a backlog figure in the Q3 release is the single most informative thing Vertiv can publish.

The Bull Case

• The contracts are long, large, and increasingly credit enhanced. Twenty year terms, take or pay structures, Google backstopping $1.73 billion of a tenant's obligations at Cipher. This is not spot GPU rental. It is contracted infrastructure cash flow.
• Supplier backlogs give multi year visibility that does not depend on picking a winner. A $176 billion GE Vernova backlog and a $53.4 billion Quanta backlog are orders already booked, not a forecast of orders.
• Power is a genuine physical bottleneck and it has a lead time. GE Vernova is booking gas turbine slots into 2031. Interconnect positions take years and litigation to replicate. Scarcity is not narrative here, it is queue position.
• Hyperscalers keep reaching outside their own footprints. Microsoft, AWS, Google, and Anthropic have all signed with third party developers rather than build everything first party. That is revealed preference about how tight capacity is.

The Bear Case

• Layer one without a lease is a cash burning developer. Keel runs negative $23.7 million adjusted EBITDA per quarter before development capital and has declined to restate its three lease target. The liquidity buys time to sign, not time to build.
• There is a duration mismatch buried in the stack. Landlords underwrite seven to nine year paybacks and sign ten to twenty year leases. The neoclouds sitting between them and the end demand run GPU contracts of two to five years. Somebody is carrying that gap.
• Supplier multiples already discount the visibility. Valuations across the electrical and cooling names sit near historical highs. You are paying for the backlog, which means the backlog has to convert on schedule.
• Disclosure quality is deteriorating at the edges. Vertiv stopped printing backlog. Developers publish pipeline gigawatts that mix energized capacity with load studies. When companies change what they show you, that is information.

The SharkWater Take

I would rather own the picks than the claims, and I would rather own a contracted landlord than a hopeful one. Those are two separate judgments and I hold both.

Layer two is where the buildout gets expressed without a counterparty bet. GE Vernova and Quanta have booked orders in hand and a physical lead time protecting the position. That does not make them cheap. It makes the failure mode a slower schedule rather than a zero, and slower schedule is a survivable outcome. That is the trade I would size normally.

Inside layer one I draw a hard line at the signature. TeraWulf, Core Scientific, and Cipher have executed documents with named counterparties, disclosed terms, and in Cipher's case a backstop from Google. Those are infrastructure businesses with an execution problem. Keel is a real estate option with a burn rate. It may well work, and I said so when I wrote it up in August, but it is a different instrument and it deserves a different position size. Do not let a shared sector label collapse that distinction.

And I would not treat IREN as a landlord comp. Its $9.7 billion headline is nine times the per megawatt economics of a Core Scientific lease because it includes the GPUs. Net of the $5.8 billion Dell purchase, the number is closer to $3.9 million per megawatt year, still the best in the group, earned by taking hardware obsolescence risk onto its own balance sheet. That may be a fine trade. It is not the same trade.

What I Could Not Reconcile

• Cipher AWS contract value. Secondary coverage cites roughly $5.5 billion over 15 years. The Cipher release I reviewed states 300 MW of capacity delivered in 2026 but not a dollar figure. I left it out of the derived table rather than compute against an unverified number.
• Megawatt definitions are not uniform. "300 MW" at Cipher's AWS site is not clearly stated as critical IT load versus gross capacity. At Barber Lake the same 300 MW gross corresponds to 207 MW of critical IT. That distinction moves per megawatt math by more than 40 percent and issuers are inconsistent about it.
• Eaton's data center backlog. A secondary source cites 307 GW of US data center backlog representing 15 years of work at current build rates. I could not verify that against the filing and have excluded it. The Q2 revenue and EPS figures above are also secondary.
• A Core Scientific AMD agreement. A July 28, 2026 8-K summary references a 2.5 GW AI capacity agreement with AMD. I did not read the underlying terms and have not included it in any figure here.
• Keel's lease status. "No signed lease" reflects the absence of any announcement I could locate through September 7, 2026. It is not a confirmed EDGAR negative.
• No prices, market caps, or valuation multiples appear in this post. Any I could source would be stale by the time you read it, and the argument does not need them.

Tight lines. — SharkWater

Educational and informational purposes only. Not personalized investment advice. All figures sourced as noted and accurate as of publication. Contracted revenue figures are management disclosures subject to delivery, commissioning, and counterparty performance, and are not guaranteed. Per megawatt calculations are the author's arithmetic on stated figures, not company reported metrics. The author may hold positions in securities discussed. Do your own work.