By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel generally views the 15-year, 529 MW lease with AMD as a significant pivot for Core Scientific (CORZ), potentially generating $14B in revenue. However, they express concerns about the deal's reliance on future cash flows, high capital intensity, and potential execution risks.
Risk: Refinancing risk due to debt maturities clustered 2025-2027 before the first dollar of AMD revenue arrives.
Opportunity: Securing 529MW of long-term capacity with a Tier-1 hyperscaler like AMD provides a predictable, multi-year cash flow profile.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Core Scientific (NASDAQ: $CORZ) is partnering with chipmaker Advanced Micro Devices (NASDAQ: $AMD) on a new artificial intelligence (A.I.) data centre.
The deal involves AMD taking out a 15-year lease for 529 megawatts of U.S. A.I. capacity, which Core Scientific said could generate more than $14 billion U.S. of revenue for it.
Core Scientific is a former Bitcoin (CRYPTO: $BTC) miner turned A.I. data centre operator.
More From Cryptoprowl:
The deal is expected to support AMD customer deployments starting in 2027. AMD is leasing compute power at Core Scientific's data centres in Texas and Oklahoma.
In a joint news release, the two companies said that they plan to collaborate on data centre design and the deployment of AMD microchips, processors, and software.
As part of the deal, AMD has also received warrants to purchase up to 30 million Core Scientific shares at $23.47 U.S. each. CORZ stock is currently trading at $20.74 U.S. per share.
The agreement lifts Core Scientific's leased customer capacity to roughly 1.1 gigawatts of compute capacity, representing more than $24 billion U.S. in potential revenue.
The partnership with AMD accelerates Core Scientific's move away from Bitcoin mining. Core Scientific currently owns 848 Bitcoin worth $49.7 million U.S.
CORZ stock has risen 50% in the last 12 months to trade at $20.74 U.S. per share. AMD stock has gained 185% in the past year to change hands at $494.95 U.S. a share.
Four leading AI models discuss this article
"CORZ’s multi-billion revenue backlog is real but heavily back-loaded, making current valuation a bet on 2027+ execution rather than immediate inflection."
The 15-year, 529 MW lease with AMD could generate >$14B in revenue for CORZ and lifts total contracted capacity to 1.1 GW / $24B potential, a credible acceleration of its pivot from BTC mining. At 11.6x forward P/E against expected 19%+ EPS CAGR through 2029, the stock looks undervalued if utilization ramps on schedule. However, the 2027 start date means material cash flows are still 2+ years away, capital intensity for the build-out is high, and AMD’s warrant position at $23.47 (only 13% above current $20.74) caps upside near-term while giving AMD an exit ramp if AI demand softens.
The $14B and $24B figures are undiscounted, non-guaranteed maximums spread over 15 years; actual revenue depends on AMD filling the capacity, power costs, and successful execution of yet-to-be-built data centers in a market that could see hyperscaler capex pullbacks by 2027.
"The partnership transforms CORZ from a commodity-exposed Bitcoin miner into a stable, long-duration infrastructure play, justifying a valuation re-rating despite the substantial execution risks involved in scaling to 1.1 gigawatts."
This deal is a massive pivot for Core Scientific (CORZ), essentially validating their transition from a volatile Bitcoin miner to a high-margin infrastructure utility. Securing 529MW of long-term capacity with a Tier-1 hyperscaler like AMD provides the predictable, multi-year cash flow profile that institutional investors demand. However, the market is pricing this as a 'done deal' while ignoring the massive capital expenditure (CapEx) required to build out these sites to AI-grade specifications. If CORZ cannot secure cheap, reliable power at scale or faces construction delays in Texas and Oklahoma, the $14 billion revenue figure is purely speculative. The warrants at $23.47 suggest AMD is hedging against their own success, but the execution risk remains significant.
The 15-year lease is a long-term liability for CORZ if AI hardware cycles accelerate faster than their facility designs, potentially leaving them with obsolete, power-hungry infrastructure that cannot be easily retrofitted.
"AMD's 15-year commitment validates CORZ's data centre model at scale, but the deal's profitability hinges entirely on power cost stability and full utilization—neither guaranteed."
This deal is structurally bullish for CORZ but the headline revenue figure ($14B over 15 years = ~$933M annually) needs scrutiny against execution risk. AMD's warrant strike at $23.47 suggests confidence, but also that CORZ negotiated hard—AMD wouldn't overpay. The real signal: 529MW of committed capacity at scale validates CORZ's pivot from Bitcoin mining. However, the article conflates 'potential revenue' with actual cash. Deployment starts 2027, power costs are rising, and AI capex cycles are volatile. CORZ's current $1.1GW pipeline ($24B potential) is impressive on paper but assumes 100% utilization and no customer churn over 15 years—unrealistic.
The $14B figure is undiscounted nominal revenue over 15 years with zero margin detail; if power costs spike or AMD reduces utilization post-2027, CORZ's margin compression could be severe, and the warrants dilute existing shareholders while giving AMD an exit ramp if the deal underperforms.
"Core Scientific's revenue promises depend on highly uncertain utilization, energy costs, and deal terms, making the upside far from guaranteed."
This looks like a bullish headline on AI infrastructure, but the math is fragile. A 15-year lease for 529 MW and a claimed $14B in revenue assumes near-perfect utilization, stable energy pricing, and favorable deal terms that aren’t disclosed (escaliers, maintenance, capex). The real test is whether Core Scientific can monetize 1.1 GW of capacity at meaningful margins, given power costs, cooling, and competition from hyperscalers and other data-center players. AMD’s 30M warrants introduce equity dilution if exercised, and the deal hinges on AMD’s hardware supply and chip pricing—not a guaranteed tailwind for CORZ. Missing debt load, capex cadence, and regulatory risks leave substantial downside risk.
The strongest counter is that the $14B revenue and $24B 'potential' metrics look promotional unless utilization and pricing are locked in; even with AMD, capacity expansion is capital-intensive and energy-cost-sensitive, and the 30M warrants create meaningful dilution if the stock rallies.
"Near-term debt wall precedes any material AMD cash flows, amplifying execution risk."
Claude's $933M annual revenue glosses over the fact that the 529 MW lease is back-loaded to 2027-2042 while CORZ must fund capex years earlier. Nobody has flagged the refinancing risk: CORZ carries $600M+ debt with maturities clustered 2025-2027; if AI capex slows and power prices rise, debt covenants could tighten before first dollar of AMD revenue arrives.
"Regulatory and interconnection delays in Texas are a more immediate threat to the 2027 timeline than the refinancing risk Grok highlighted."
Grok correctly identifies the liquidity trap, but misses the regulatory tail risk. The 529MW allocation isn't just a capital hurdle; it’s a political one. ERCOT and regional grids are increasingly restricting data center permits due to load volatility. If CORZ faces interconnection delays or 'social license' pushback, that 2027 start date is a fantasy. The market is ignoring that power scarcity is a zero-sum game, and hyperscalers like AMD may prioritize their own sovereign cloud build-outs over third-party providers.
"CORZ's existing operational footprint reduces greenfield regulatory risk, but debt maturity clustering remains the binding constraint before 2027 revenue arrives."
Gemini's ERCOT constraint is real, but underweights CORZ's existing footprint advantage. They already operate 600+ MW in Texas/Oklahoma with grid interconnects live. AMD's 529 MW likely layers onto existing infrastructure, not greenfield sites. Regulatory delays are a risk, but retrofitting existing capacity is materially faster than Gemini implies. The refinancing cliff Grok flagged is the sharper near-term threat—debt maturities 2025-2027 hit before revenue ramps.
"The 529MW deal hinges on a fragile set of assumptions—real interconnection and cooling/capex constraints, rising power costs, and a 2027 start—so the undiscounted $14B over 15 years is not a reliable value signal; AMD warrants simply add dilution risk and optionality."
Claude overweights the speed of retrofitting around existing assets; the real bottlenecks are interconnection delays, grid upgrades, and cooling Opex that could push the 2027 start, not just CapEx cadence. Even with AMD as an anchor, a back-loaded 529MW lease subject to power price swings and utilization risk makes the undiscounted $14B over 15 years a poor value proxy. Warrants add dilution tail risk if AMD is later rewarded for underutilization.
The panel generally views the 15-year, 529 MW lease with AMD as a significant pivot for Core Scientific (CORZ), potentially generating $14B in revenue. However, they express concerns about the deal's reliance on future cash flows, high capital intensity, and potential execution risks.
Securing 529MW of long-term capacity with a Tier-1 hyperscaler like AMD provides a predictable, multi-year cash flow profile.
Refinancing risk due to debt maturities clustered 2025-2027 before the first dollar of AMD revenue arrives.