Core Scientific's AI Pivot: Progress in High-Performance Computing Shift

===========================CRITICAL HARD RULES (QWEN-SAFE)===========================

1. You MUST output the FULL ORIGINAL ARTICLE.
2. The ONLY allowed change is inserting ONE tag.
3. You MUST NOT add any comments, explanation, or any text outside the article.
4. You MUST NOT rewrite, reorder, delete, normalize, or alter ANY character.
5. You MUST NOT output the tag alone.
6. You MUST NOT place the tag at the beginning or at the end of the article.
7. The tag MUST be inserted inside a MIDDLE PARAGRAPH.

A “paragraph” = a block of text separated by one or more blank lines.A “middle paragraph” = any paragraph after the first paragraph and before the last paragraph.

If you break ANY rule → your output is INVALID.

===========================THE TAG (INSERT EXACTLY ONCE)===========================

Insert this exact structure:

Notes:• Replace MODEL_GENERATED_VALUE with the generated ticker.• Replace CATALOG_ID_VALUE with the matched catalog id.• Replace TIME_RANGE_VALUE with exactly one of: 0, 1, 2, 3, or 4.• DO NOT output any parentheses or explanations inside the tag.

The tag MUST be inserted at a natural boundary:• end of a sentence• or end of a line

Do NOT change any surrounding characters.

===========================HOW TO GENERATE THE ATTRIBUTES===========================

1. code (MODEL_GENERATED_VALUE)

Pick the single most relevant ticker found in the article.
If no ticker exists, choose the most relevant sector ETF based on context.If still unclear, default to SPY.

1.1 Crypto Rule

If the most relevant asset is a cryptocurrency:• Convert it to its USDT trading pair (e.g. BTC → BTCUSDT).• If the article uses the full name (e.g., “Bitcoin”), map it to the standard ticker first (Bitcoin → BTC → BTCUSDT).

This rule applies ONLY to crypto assets.Non-crypto tickers MUST stay unchanged.

2. id (CATALOG_ID_VALUE)

NEWS_BACKTEST may be:• a Python dict• a JSON string

Parse it if needed.

Choose ONE id from:data.newsBacktest[0].items[*].id

Selection MUST be based on semantic matching between:• ARTICLE text• items[*].details

If no strong match:• choose the item describing trend/momentum
If still unclear:• choose the FIRST item in the catalog

3. range (TIME_RANGE_VALUE)

Use a 5-year backtest window (timeRangeId="3") as the default.
Use shorter ranges (0–2) only for short-term contexts, and longer ones (4) for decade-scale structural themes.

===========================MANDATORY OUTPUT FORMAT===========================

You MUST output:✔ the original ✔ with the inserted tag inside a middle paragraph
✘ no explanation
✘ no extra text

===========================INPUTS===========================

Core Scientific's shift from Bitcoin mining to AI and high-performance computing (HPC) services is showing initial traction through rising revenue streams. The company's AI compute revenue surged 45% year-over-year to $15 million in Q3 2025, providing a key counterweight to persistent declines in its core mining business. This growth helped push HPC hosting revenue up from $10.3 million in the same quarter of 2024 to $15.0 million in Q3 2025, reflecting a clear strategic repurposing of its infrastructure.

The pivot is further evidenced by a significant commitment to partner with AI cloud provider CoreWeave. Core Scientific plans to dedicate 250 megawatts (MW) of its power capacity to CoreWeave's colocation services by the end of 2025, signaling a concrete reallocation of resources away from Bitcoin mining. This partnership is a critical element of the company's new business model.

While the revenue gains in AI and HPC services represent a positive step, the company's transition remains in early stages. Notably, detailed metrics specific to AI compute revenue or cost ratios for 2024 are not provided in the latest report, limiting deeper performance analysis for the prior year. Investor sentiment is also clouded by concerns over the valuation in CoreWeave's proposed $9 billion all-stock acquisition, which could impact the sustainability and timing of Core Scientific's growth trajectory. The overall financial impact hinges on successfully executing these large-scale partnerships and demonstrating consistent revenue growth beyond the initial gains.

Funding Strategy & Growth Enablers

Core Scientific's pivot hinges on a significant funding boost and asset redeployment. The company secured $625 million in senior convertible notes, lifting its cash pile to $836.2 million. This liquidity fortifies its balance sheet for the strategic shift, though the notes carry inherent interest and conversion risks typical of such instruments. Infrastructure expansion is central to the plan, with the company reaching 1,300MW of powered capacity by year-end 2024. A key allocation involves dedicating 250MW to CoreWeave's AI colocation by the end of 2025, reflecting a major reallocation of resources away from Bitcoin mining.

This capital deployment directly addresses the revenue shift. Bitcoin mining income declined sharply, falling 22% year-over-year, creating a funding gap. AI compute services partially offset this, with revenue growing 45% YoY to $15 million in Q3 2025. HPC hosting also provided a recent cash flow boost, generating $8.5 million in Q4 2024 with a 13% non-GAAP gross margin. However, total capital expenditure remains substantial at $244.5 million, with partial funding sourced from the CoreWeave partnership.

Despite the funding and growth initiatives, significant execution risks persist. No detailed AI-specific cost metrics or competitive positioning data are publicly available, making profitability assessments difficult. Investor skepticism remains high regarding the valuation in CoreWeave's proposed $9 billion all-stock acquisition of Core Scientific, casting doubt on the long-term value creation potential of this key partnership. The success of the HPC/AI transition remains contingent on overcoming these operational and valuation hurdles.

Execution Risks & Competitive Pressure

Core Scientific's pivot to AI/HPC services faces significant operational and valuation hurdles despite recent revenue growth. While AI compute revenue rose 45% YoY to $15 million in Q3 2025, investor concerns about CoreWeave's $9 billion all-stock acquisition persist, raising questions about whether Core Scientific's assets are fully valued in the deal structure. This uncertainty complicates fundraising and strategic clarity, especially as the company plans to dedicate 250 MW to CoreWeave's colocation by year-end.

The decline in Bitcoin mining revenue (-22% YoY) further strains the transition. Though AI revenue offsets some losses, Bitcoin still represents a core revenue stream, creating a dual-focus operational challenge as facilities repurpose toward AI workloads. Capital expenditures reached $244.5 million, partially financed by CoreWeave, but the sustainability of this funding model remains unclear amid Bitcoin's volatility.

Critical gaps in transparency compound these risks. No AI-specific cost metrics or competitive benchmarks (e.g., cost-per-compute or utilization rates) are disclosed, making it difficult to assess scalability or efficiency. Without granular data, investors cannot determine if rising revenue margins are sustainable or if competitive pressures will erode profitability as AI demand intensifies.

While facility repurposing and CoreWeave partnerships signal progress, the absence of clear cost controls, coupled with valuation disputes and lingering reliance on declining Bitcoin revenue, creates execution risks that could delay the AI transition if not addressed.

Valuation Catalysts & Market Sentiment

The $5.2 trillion projected global AI capital expenditure market by 2030 represents a massive potential tailwind for infrastructure providers like Core Scientific, offering a clear long-term growth thesis for the company's strategic pivot. This enormous market opportunity, driven by surging generative AI adoption and enterprise integration, provides a compelling backdrop for the firm's shift towards high-performance computing (HPC) services, suggesting significant upside potential if execution succeeds. Core Scientific is actively positioning itself within this market, expanding its powered infrastructure to 1,300 MW by year-end 2024 and securing substantial capital, including $625 million in senior convertible notes and bolstering its $836.2 million cash reserves to fund this transition.

Recent financial results provide tangible evidence of this strategic shift gaining traction. AI compute revenue surged 45% year-over-year to $15 million in Q3 2025, marking a significant increase from $10.3 million in the same period last year and demonstrating concrete progress in monetizing its HPC capabilities. Management's plan to allocate 250 MW specifically to CoreWeave's AI colocation further underscores the focus on capturing value within this expanding ecosystem. This growth, coupled with a strong cash position, supports the argument for increased investment weight based on achieving key milestones in AI revenue penetration.

However, significant valuation uncertainty and execution risks temper the excitement surrounding these growth signals. Investor concerns persist over the perceived undervaluation in CoreWeave's proposed $9 billion all-stock acquisition of Core Scientific, a major partnership, casting a shadow over the current share price and suggesting potential friction in realizing full market value. Furthermore, the company's heavy reliance on Bitcoin mining revenue, which declined 22% year-over-year, introduces cyclicality and market sensitivity that could complicate the transition if AI demand doesn't ramp as quickly as planned. While the macro AI capex market is vast, the path to capturing a meaningful share remains unproven for Core Scientific, requiring successful navigation of competitive pressures and project execution over the coming quarters. The increasing penetration rate in AI/HPC services is promising, but its sustainability and profitability at scale are still untested in the market.