CIFRStock Deep DiveNeutral

With an 8.5 billion yuan long-term contract in hand, why did CIFR drop 13% in one day?

$850 million in convertible bonds has fueled dilution anxiety, with thick contracts and thin cash, and CIFR still waiting for rent payments.

October 13, 2026
Share
NextPick signal · CIFR
Strong sellissued 9/24
Down since signal-37.4%
💰 Money flowing out📍 Below 50-day MA $16.48

NEXT PICK · Market Insights

With an 8.5 billion yuan long-term contract in hand, why did CIFR drop 13% in one day?

Tuesday, October 13, 2026

$850 million in convertible bonds has fueled dilution anxiety, with thick contracts and thin cash, and CIFR still waiting for rent payments.

An $850 million convertible bond caused CIFR to evaporate nearly 13% in a single day

On October 13, Cipher Digital ($CIFR) closed at $11.34, down 12.77% in a single day, with trading volume expanding to about 3.8 times the daily average over the past three months.

The reason it trended wasn't because of a financial report crash or losing a major client, but because the company announced the issuance of $850 million in zero-coupon convertible bonds and swapped some old bonds for common shares.

This is a typical "financing-driven decline"

New debt delays dilution until 2032, but swaps drag dilution straight to today.

A company holding long-term contracts with AWS and Fluidstack, with a planned power pipeline of 5.3 GW, has a market value of only about $4.7 billion, down more than 20% this year, and more than 60% from its June peak of about $30.

So here's the question

Is this cut cutting off a bubble, or is it discounting an "AI infrastructure option"?

Latest price
$11.34
▼ -12.77%
Quantity ratio
3.8 times
▲ Daily average for the past three months
Convertible bonds
$850 million
— Zero interest / maturity in 2032
$CIFRCipher Digital Inc. $11.34▼ -12.77%

From Bitcoin miner to AI landlord, Cipher is actually selling electricity

Cipher Digital was formerly known as Cipher Mining, which was renamed in February 2026. Removing the word "mining" from its name is itself a strategic statement: The company is transforming from a Bitcoin miner into an HPC data center landlord who builds, supplies power, and collects rent for hyperscale cloud providers.

The cost of transformation is written on the income statement. In fiscal year 2025, revenue was $223.9 million, but net loss reached $822.2 million; Entering 2026, mining operations contracted as planned, with Q1 revenue of $34.8 million, Q2 further dropping to $24.8 million, and new business rental income only began to be booked in the second half of the year.

What truly supports valuation is the contract book

In the second half of 2025, the company signed a 15-year lease worth about $5.5 billion with Amazon AWS, and a roughly $3 billion, 10-year hosting agreement with Google-backed Fluidstack.

Google also secured Fluidstack's $1.4 billion lease obligation and obtained warrants corresponding to about 5.4% equity, effectively betting its credit on the Cipher campus.

In Q1 2026, the company signed a third lease for an AI data center campus with an investment-grade hyperscale tenant. According to management, the signed lease target is an average annual net operating income of about $793 million—this figure is more than three times the revenue for fiscal year 2025.

If you compare Cipher to a mall still under renovation, then AWS and Google have already signed the lease for their flagship store, but the lights in the mall aren't all on yet—how much are you willing to pay upfront for a signed contract that hasn't started paying rent yet?

The price-to-sales ratio of over twenty times is betting on the rent three years from now

With a market value of $4.71 billion, compared to less than $200 million in revenue over the past four quarters, CIFR's price-to-sales (P/S) ratio is in the 20-plus range and 26.90 times in research reports, while the software industry averages only 5-10 times.

The price-to-earnings (P/E) ratio cannot be calculated due to ongoing losses, with earnings per share (EPS) of -$2.87 and return on equity (ROE) of -73.32%—there is hardly a solid figure to support on the income statement.

The valuation model in the report presents three scenarios

Under the low scenario, revenue in 2028 is $320 million, with a P/S ratio of 8 times valued at $2.56 billion; In the medium scenario, revenue is $390 million, or 12 times, or $4.68 billion; In the high scenario, revenue is $520 million, or 20 times, or $10.4 billion.

Today's market cap of about $4.7 billion is almost exactly stuck in the Chinese scenario; If only fiscal year 2025 revenue is applied at an 8x transition premium, the reasonable market cap would be only about $1.8 billion, with a current premium exceeding 175%.

But this model, based on historical revenue, may underestimate the weight of the contract book—once the target of about $793 million in annual net operating income is realized, the revenue curve will be much steeper than the 45% compound annual growth rate. The research report did not perform DCF due to a lack of reliable discount rate, which itself shows that the valuation anchor is not in the financial statements, but in the construction progress schedule.

Cash flow is another mirror. Free cash flow for fiscal year 2024 is -$875 million, for fiscal year 2025 is -$207.9 million, and in Q1 2026, it unusually turned positive at $91.5 million.

However, in Q2, it recorded an operating loss of $78.5 million and a net loss of $267.5 million, of which $150.5 million came from non-cash warrant revaluation.

Share capital will increase from about 405 million shares at the end of 2025 to about 415 million shares in Q2 2026. The report's judgment is that "external capital will be needed within one year"—Today's $850 million convertible bond actually fulfills this judgment ahead of schedule.

So, has the negative news been paid off? The financing "boot" has been implemented, but the spot dilution from the share swap will only be fully calculated after pricing; And with a surge in volume rent, the stock price clearly hasn't been willing to buy in early.

During the pre-breakout phase, the market was the first to vote against it

The report positions Cipher at the "pre-breakthrough" stage: the Black Pearl campus has delivered its first batch of HPC capacity two months ahead of schedule, with leasing starting in August; the Barber Lake campus is planned to start leasing in October; both power resources and long-term contracts are in hand, but revenue scale and profitability have yet to cross the turning point, and cash flow has yet to stabilize and turn positive.

But moving averages tell a different story

The closing price of $11.34 is about 32% lower than the 50-day moving average of $16.74, about 38% lower than the 200-day moving average of $18.28, and the 50-day moving average has fallen below the 200-day moving average. Fundamentals are moving forward, but stock prices are falling — the market is clearly saying: "Pay rent first, then talk valuation." "

While others were looking for electricity, Cipher was waiting for the meter to start running

According to the research report, Hut 8 (HUT), another mining company transforming, has about 1.2 GW of installed power capacity, with a price-to-sales ratio of less than 5 times, making the market more cautious about its revenue prospects.

Nebius (NBIS) is pursuing a GPU cloud route. Research reports estimate its price-to-sales ratio is about 3-5 times higher, with more flexible pricing, but electricity mainly relies on third-party procurement, resulting in higher costs.

The report states that CoreWeave is "unlisted and hard to compare," which needs correction: CoreWeave (CRWV) was listed on NASDAQ in 2025 and is a benchmark in the Neocloud sector. Every earnings report it releases affects the sentiment of the entire Nebius, Cipher, and Hut 8 sectors.

Cipher's differentiation lies in its 5.3 GW of power line planning, long-term contract endorsements from AWS and Google, and facilities tailored for tenants that are difficult to relocate once occupied—tenant switching costs are extremely high, which is the strongest moat for data center landlords.

The report also mentions network effects and economies of scale

The larger the power and tenant base, the more GPU cloud providers can be attracted, and the lower the unit operating cost. However, the 10-K does not disclose core technology patents, with barriers mainly coming from power infrastructure and operational experience.

The problem is, in the AI computing power gold rush, there are many people selling shovels. What is truly scarce is not the shovel, but the land that can be connected to the power grid—how much of the land Cipher holds can actually get electricity on time?

With an $80 billion pool, Cipher is targeting the $3 billion spoonful

The report estimates that by 2025, the global HPC data center market (TAM) will be about $80 billion, with the United States accounting for about 45%, with demand mainly coming from hyperscale cloud providers such as Amazon, Google, and Microsoft.

The serviceable market (SAM) for HPC infrastructure services is about $15 billion, while Cipher leverages the available market (SOM) for electricity and location in Texas and surrounding areas to be about $3 billion.

The report assumes a compound annual growth rate of about 45% in HPC demand from 2026 to 2028, with company revenue growing from about $25 million in Q2 2026 quarterly to about $1.2 billion in 2028.

The prerequisite is that electricity asset utilization rises to 80%, 5-7 tenant contracts are signed, and construction progresses at 2 new stations per month. Every assumption is aggressive; if any link fails, the ceiling becomes the floor.

Contracts signed, production capacity ahead of schedule, and the bulls hold a hand

Bulls don't need to believe that "AI will always thrive"; they only need to believe that contracts will be executed:

▲ Bull Case
① Signed about $5.5 billion 15-year leases for AWS and about $3 billion for Fluidstack for 10-year agreements, offering extremely high revenue visibility
② Google secured $1.4 billion in lease obligations for Fluidstack and holds approximately 5.4% equity warrants, with deeply bound interests
③ Black Pearl was delivered two months in advance and leased starting in August, while Barber Lake is scheduled to start renting in October, with execution beyond expectations
④ The leased target annual net operating income is approximately $793 million, more than three times the revenue for fiscal year 2025
⑤ Free cash flow in Q1 2026 was $91.5 million, turning positive for the first time and reversing the $875 million loss in fiscal year 2024
⑥ 5.3 GW Power Pipeline Planning is a scarce resource at a time when "power shortage" has become the biggest bottleneck for AI
⑦ The zero-coupon convertible bonds are locked in no interest burden before 2032, with funds raised in Barber Lake and HPC pipelines, no longer used for mining

Reports are bleeding, dilution is accelerating, and bears are justified

Bears focus on a simple fact: no matter how beautiful the contract is, cash flow must be scraped through season by quarter:

▼ Bear Case
① Q2 2026 revenue is only $24.8 million, about 22% below the market expectation of $31.7 million, and down from $43.6 million year-over-year
② Q2 net loss was $267.5 million, EPS -$0.65, a 133.9% increase from Q1's $114.3 million loss
③ In fiscal year 2025, revenue growth will be -49.27%, net profit growth will be -586.41%, with mining shrinking faster than HPC replacement
④ The price-to-sales ratio is over twenty times, far higher than Hut 8 and less than five times, with an ROE of -73.32%. The enterprise value multiple cannot be calculated due to negative EBITDA
⑤ Share capital increased from about 405 million shares to about 415 million shares, and combined with this share swap, dilution is shifting from the "future tense" to the "present" tense
⑥ Revenue is highly dependent on the top two tenants; any delay or default on either side would directly impact cash flow
⑦ The stock price has fallen more than 20% this year, down more than 60% from its June high. Today marks the 12th decline in 15 trading days

Starting rent, pricing, and financial reports—these three factors determine the next step

In the coming months, CIFR's stock price will most likely revolve around three factors: whoever pays first will have the upper hand:

• Convertible bond pricing and conversion scale (recent): The conversion price and number of shares determine the depth of dilution; pricing preference exhausts negative news, while deviations add further pressure
• Barber Lake campus starts renting in October (Q4 2026): The first Fluidstack campus begins renting, which is a key signal testing execution and is positive
• Q3 2026 Financial Report (expected November): Black Pearl's rental income is reflected for the first time. If HPC revenue a surge in volume falls short of expectations, it will amplify bearish sentiment
• Texas data center approval suspension (ongoing from August): The governor requires PUCT and ERCOT to audit and approve new projects, serving as a moat for approved projects but uncertainty for pipeline approvals

From $30 to $11, the market is pricing in the "time difference."

Over the past year, CIFR has almost fully recreated a rollercoaster of the AI infrastructure narrative: in the second half of 2025, two long-term contracts for AWS and Fluidstack were launched consecutively, their stock prices were repeatedly revalued, and in June 2026, they reached a high of about $30.

The turning point came during earnings season. Q1 revenue was $34.8 million, Q2 revenue was $24.8 million, both below expectations, and warrant fees in Q2 further amplified losses to $267.5 million, causing the stock price to weaken noticeably after the earnings report.

The convertible bond in October combined with the share swap pushed it to its lowest close since September 2025.

The review of the research report is very objective

The assessment of power assets and regulatory moats is correct, but the timeline for revenue and cash flow improvements is viewed too optimistically by the market.

⚠️ Risk Notice

• AI investment cycles are slowing, hyperscale tenants are cutting capital expenditures, and HPC demand is declining
• Park construction overruns or delays, delaying rent start and revenue recognition
• Rising electricity costs or unstable supply erode leasing profits
• Revenue is concentrated among the top two tenants, and a single customer default has a huge impact
• Texas regulatory audits are ongoing and the progress of projects affected by natural disasters such as extreme weather
• Convertible bonds and share swaps have led to ongoing equity dilution, and financing may still be needed going forward

🟡 Neutral

Contracts are genuine and capacity is accelerating, but dilution and cash gaps have yet to bottom.

💬 Discussion

Negative financing news has materialized; don't hold heavy positions before rent is delivered; Post CIFR to the official account for real-time buy and sell positions.

Data source

• Source: NextPick real-time snapshot + company earnings and SEC announcements + mainstream financial media coverage, data as of the close of the US East Asian Week on October 13, 2026.

Disclaimer: This article is for reference only and does not constitute investment advice. Markets carry risk — invest with caution.