Researchers use large-scale AI-driven molecular modeling to investigate thousands of immune-recognition interactions, offering a promising new computational approach while revealing the limits of current prediction methods.

A longstanding challenge in immunology involves identifying the specific molecular targets recognized by T cells, a pursuit with significant implications for our understanding of immune responses, infectious diseases, and oncology. Predicting these interactions computationally remains complex, as researchers must account for the intricate three-dimensional relationships among T-cell receptors, peptides, and the major histocompatibility complex (MHC) molecules that present them.

A recent study published in Science Advances https://www.science.org/doi/10.1126/sciadv.aef5674 evaluates the utility of artificial intelligence in addressing this problem. Specifically, the research uses AlphaFold 3 to predict the structures of these molecular complexes, aiming to assess the likelihood of a particular T-cell receptor recognizing a given peptide-MHC target.

Titled “General prediction of T cell receptor antigen specificity from sequence using AlphaFold 3,” this research transcends the conventional application of AI for generating plausible molecular structures. Instead, it investigates whether such structures can provide actionable predictive insights into biological recognition, including interactions involving antigens that lack prior structural characterization. For the supercomputing community, this work serves as a compelling demonstration of how AI-driven molecular modeling is becoming an integral component of the computational infrastructure supporting modern biological research.

Turning molecular structure prediction into a recognition problem

T cells play a central role in adaptive immunity. Their receptors recognize specific molecular targets presented by MHC proteins on cell surfaces. This recognition process is highly selective, but the number of possible receptor-antigen combinations is enormous.

Identifying the targets of individual T-cell receptors experimentally can be labor-intensive. Computational prediction offers a potential way to narrow the search, helping researchers prioritize candidates for further laboratory investigation.

The study examines whether AlphaFold 3 can contribute to that process by modeling the molecular triad formed by the T-cell receptor, peptide, and MHC molecule.

The researchers assembled a substantial collection of known cognate and noncognate interactions. The training data included 7,702 cognate and 157,025 noncognate class I triads, along with 1,625 cognate and 33,628 noncognate class II triads. Together, the dataset contained 9,327 known cognate interactions and 190,653 noncognate examples.

The distinction matters. A model that can generate a plausible structure is not necessarily capable of determining whether two biological components actually interact. The researchers therefore used the predicted structural information to develop a framework for distinguishing recognized targets from nonrecognized combinations.

A substantial computational workload

The methodology also illustrates the computational demands of applying modern structure-prediction systems across large biological datasets.

The researchers ran each molecular triad through AlphaFold 3 using five random seeds, with one diffusion sample per seed and ten recycling steps. Multiple sequence alignments and templates were enabled.

Applying that configuration to the full training dataset implies approximately one million seed-level prediction runs. This is an estimate based on the reported dataset size and five-seed protocol, not a reported count of independent jobs or a measurement of accelerator-hours.

The distinction is important for HPC readers. The paper describes the computational procedure, but does not provide a detailed accounting of the hardware used, GPU configuration, total runtime, memory consumption, energy use, or cluster architecture. It therefore cannot be treated as a benchmark of AlphaFold 3 performance on a particular supercomputer.

Nevertheless, the scale of the experiment demonstrates why high-performance computing, accelerated AI infrastructure and efficient molecular modeling workflows are increasingly important to computational biology. Applying structure-prediction methods to hundreds of thousands of molecular combinations creates a workload involving repeated inference, substantial data handling, and downstream statistical analysis.

As researchers expand these approaches to larger receptor and antigen collections, the ability to manage and execute such workloads efficiently will become an important part of the scientific process.

Promising results, but important limitations

The study reports encouraging results on held-out data, although performance differed between MHC class I and class II predictions.

For class I, the reported median area under the receiver operating characteristic curve (AUC) was 0.92, with results ranging from 0.50 to 1.00. For class II, the median AUC was 0.81, with results ranging from 0.71 to 0.91.

The validation included 14 class I antigens and eight class II antigens. The researchers also examined previously structurally uncharacterized class II antigens, an important test of whether the approach could extend beyond familiar structural examples.

But the headline performance figures do not tell the entire story.

The researchers found that sensitivity could be limited when the model was required to maintain very high specificity. For class II predictions, median sensitivity was 13% at 100% specificity, 26% at 99% specificity, and 52% at 95% specificity.

In practical terms, the model could identify a subset of likely interactions with high confidence, but it could also miss a substantial proportion of genuine targets. That is a significant limitation for applications where failing to identify an important antigen could matter as much as incorrectly predicting one.

Performance also depended strongly on the experimental assay used to establish the underlying interactions. Binding assays produced median AUC values of 0.93 for class I and 0.94 for class II. Other assay categories produced substantially weaker results, with several class II assay types showing median AUC values below 0.6.

That variability underscores a broader challenge in computational biology: AI models are only as reliable as the biological evidence used to train and evaluate them. Different experimental methods measure different aspects of molecular recognition, and those differences can influence the apparent predictive performance of a computational system.

The search space remains enormous

The researchers also acknowledge a major obstacle to applying the approach at unrestricted scale.

Even if a model can distinguish likely interactions among a carefully assembled collection of candidates, searching across billions of possible antigen combinations is another matter entirely. The paper identifies unconstrained mapping across such a vast antigen space as beyond the reach of the current approach.

This is where the HPC implications become particularly relevant.

The challenge is not simply generating more molecular structures. It is determining which combinations are worth modeling, how to prioritize computationally expensive predictions, how to manage uncertainty, and how to integrate structural results with experimental evidence.

Future systems may need more sophisticated screening pipelines that combine sequence-based filtering, structural prediction, statistical ranking, and laboratory validation. Such a workflow could reserve the most computationally expensive modeling for candidates that survive earlier stages of analysis.

That kind of tiered approach is familiar to HPC researchers working with large simulation and scientific AI workloads: use less expensive methods to reduce the search space, then apply more demanding computation to the most promising candidates.

Implications for cancer research and infectious disease

The potential applications are significant, particularly in areas where understanding T-cell recognition could help researchers identify relevant immune targets.

Cancer immunotherapy research, infectious disease studies, and personalized approaches to immune profiling could all benefit from improved methods for predicting which antigens are recognized by particular T-cell receptors.

However, this study does not establish a clinical diagnostic or treatment capability. Its results represent progress in computational prediction, not proof that the approach can independently identify clinically actionable targets across real-world patient repertoires.

The limitations in sensitivity, assay-dependent performance, and the size of the antigen search space remain substantial barriers.

The value of the work is that it provides evidence that AI-generated structural information can contribute to a difficult biological classification problem, while also making clear where the current methodology falls short.

Scientific AI needs more than impressive predictions

This research highlights the growing synergy between AI, molecular simulation, and HPC. While AlphaFold 3 improves how we model biological structures, translating these predictions into scalable, reliable discoveries requires addressing immense search spaces and verifying genuine biological interactions. This study serves as a critical blueprint for managing complex AI-driven workloads in computational biology.

The AI infrastructure giant is expanding at extraordinary speed, but its latest financial results expose a difficult question: Can the economics of renting advanced computing capacity keep pace with the cost of building it?

Three years ago, CoreWeave operated as an ambitious AI cloud startup centered on the acquisition of NVIDIA H100 GPU servers via Dell Technologies. Its business model was conceptually straightforward: procure high-performance computing infrastructure, establish the necessary facilities and networking, and provide accelerated computing capacity to organizations driving artificial intelligence development.

Today, the company has emerged as a preeminent force within the rapidly expanding AI infrastructure trade. Its third-quarter 2026 financial results, released on November 10, demonstrate revenue of $3.661 billion, a 168% increase year-over-year, with a reported revenue backlog reaching approximately $132.5 billion.

However, beneath this significant growth lies a complex financial structure that warrants rigorous analysis. CoreWeave reported a $709 million quarterly net loss and $846 million in net interest expense, with full-year 2026 capital expenditure projections ranging between $37 billion and $39 billion. Consequently, the firm's expansion is predicated on substantial borrowed capital, long-term customer commitments, and sustained access to external financing.

The central challenge currently facing the AI infrastructure industry has evolved; the primary concern is no longer merely the volume of GPU deployments or the speed of data center operationalization. Rather, the critical question is whether the operational economics of these high-cost computing systems can viably support the financial obligations necessary for their development.

From a $2 billion startup to an AI infrastructure powerhouse

The CoreWeave story began to attract considerable attention in 2023, when the company was valued at approximately $2 billion and was rapidly expanding its NVIDIA GPU infrastructure.

In December of that year, Dell announced that CoreWeave had purchased thousands of PowerEdge XE9680 servers equipped with NVIDIA H100 Tensor Core GPUs.

The systems were designed to support artificial intelligence, machine learning, visual effects rendering and large-scale simulations.

The original transaction was never assigned a publicly disclosed dollar value or an exact server count. Nevertheless, it represented an important moment in the evolution of the AI computing market.

Rather than building traditional enterprise data centers and waiting for customers to purchase hardware, CoreWeave was establishing a specialized cloud infrastructure business centered on accelerated computing.

The model subsequently attracted enormous investment and customer commitments.

CoreWeave went public in March 2025, and its infrastructure footprint continued expanding through successive GPU generations. Its relationships with Microsoft, OpenAI, and other major AI developers helped establish it as a significant supplier of specialized computing capacity.

The scale of the opportunity is now reflected in its financial statements.

However, the same strategy that accelerated CoreWeave's growth has created substantial capital requirements.

Revenue is exploding, but profitability remains elusive

CoreWeave's third-quarter results present two very different pictures of the business.

Revenue increased from $1.365 billion in Q3 2025 to $3.661 billion in Q3 2026. Operating income improved from $52 million to $207 million.

That is meaningful progress.

Yet net interest expense increased from $310 million to $846 million over the same period.

The company's quarterly operating income was therefore substantially smaller than the amount it paid in net interest expense.

The result was a $709 million net loss, compared with a $110 million loss in the third quarter of 2025.

The comparison raises an important question about the company's ability to convert rapid infrastructure expansion into sustainable earnings.

CoreWeave reported adjusted EBITDA of $2.256 billion, representing a 62% margin. That is an impressive figure, but it requires context.

Adjusted EBITDA excludes interest, taxes, depreciation and amortization, among other adjustments. Those exclusions are particularly important for a business whose primary assets are expensive computing systems and the facilities required to operate them.

Depreciation and amortization reached $1.889 billion during the quarter.

These are not incidental accounting expenses. They reflect the enormous amount of infrastructure CoreWeave must continually acquire, deploy, and eventually replace.

For a company whose competitive advantage depends on providing access to the latest GPU generations, hardware depreciation is central to the economics of the business.

A 62% adjusted EBITDA margin does not mean that 62% of revenue is available to repay debt, purchase replacement GPUs, or finance new data centers.

The distinction between adjusted profitability and the actual cost of maintaining a competitive computing fleet is one of the most important issues investors and infrastructure customers should understand.

The capital expenditure challenge: spending faster than cash generation

CoreWeave's infrastructure investment requirements are becoming increasingly difficult to ignore.

The company generated approximately $1.975 billion in operating cash flow during Q3 2026. During the same quarter, it spent approximately $9.441 billion purchasing property, equipment, and capitalized internal-use software.

That produces an approximate quarterly free cash flow figure of negative $7.466 billion, calculated using operating cash flow less those purchases.

For the first nine months of 2026, operating cash flow totaled approximately $5.638 billion, while property and equipment purchases reached $23.558 billion.

The resulting approximate free cash flow deficit was $17.92 billion.

These figures illustrate the capital intensity of the AI infrastructure business.

CoreWeave is not simply purchasing servers. It is financing an enormous ecosystem of computing infrastructure involving GPU accelerators, high-speed networking, storage, power distribution, cooling systems, and data center facilities.

The company's full-year 2026 capital expenditure forecast now stands at $37 billion to $39 billion, compared with expected revenue of $13 billion to $13.3 billion.

In other words, planned infrastructure spending is approaching three times projected annual revenue.

Management argues that accelerated spending reflects strong customer demand and the opportunity to bring additional infrastructure online sooner.

That may be true. But the financial implications remain substantial.

Every new deployment requires capital before the infrastructure can generate its full expected revenue. The equipment must be installed, powered, networked, and integrated into production environments before customers can consume the computing capacity.

The faster CoreWeave expands, the more capital it must secure to support that expansion.

The $46.9 billion debt burden

The balance sheet provides another indication of the financial risks associated with the company's growth strategy.

At September 30, 2026, CoreWeave reported approximately $46.9 billion in combined recourse and non-recourse debt.

That included approximately $41.6 billion in recourse debt and $5.3 billion in non-recourse debt.

The distinction matters because different financing arrangements carry different claims against the company and its assets.

CoreWeave also reported approximately $4.1 billion in cash and cash equivalents.

Its total assets reached $91.5 billion, compared with total liabilities of approximately $87.5 billion. Stockholders' equity stood at approximately $4 billion.

The balance sheet also contained approximately $18.7 billion in operating lease liabilities and another $649 million in finance lease liabilities.

These obligations matter because the company's infrastructure strategy depends on access to physical facilities, not simply ownership of GPUs.

The debt burden is particularly important when considering how quickly AI computing hardware evolves.

A system purchased today must generate sufficient revenue throughout its useful economic life to justify its acquisition and financing costs. If newer GPU generations offer significantly better performance, efficiency, or economics, older systems may face pressure to reduce their rental prices.

That creates a potentially difficult situation for an infrastructure provider carrying substantial debt secured against expensive equipment.

The challenge is not necessarily that the hardware becomes obsolete overnight. It is that the economic value of a previous generation can decline faster than the financing obligations associated with acquiring it.

The backlog looks impressive, but delivery is the real test

CoreWeave's reported $132.5 billion revenue backlog is arguably the most impressive number in its latest earnings release.

The company has secured substantial customer commitments for future computing capacity.

However, backlog should not be confused with cash already received, revenue already earned or guaranteed profit.

CoreWeave defines its revenue backlog as remaining performance obligations combined with other estimated amounts expected to be recognized under committed customer contracts. Recognition remains subject to delivery and service availability requirements.

That distinction is critical.

The company must build and deliver the infrastructure necessary to satisfy those commitments.

It must secure GPUs, networking equipment, storage, electricity, data center capacity, and financing. It must also ensure that its systems remain competitive throughout the periods covered by customer agreements.

A large backlog can provide visibility into future demand, but it also creates enormous execution obligations.

CoreWeave's ability to convert those commitments into delivered computing services and positive cash generation will be a much more meaningful measure of its long-term performance.

The customer concentration problem

Another issue is the concentration of revenue among a relatively small number of enormous customers.

CoreWeave's 2025 annual disclosures identified Microsoft as representing approximately 67% of its revenue.

The customer mix has continued evolving, with OpenAI, Meta, and other AI developers becoming increasingly important.

The company's expanding relationships with major AI customers demonstrate the demand for specialized accelerated computing.

But concentration creates exposure.

If one major customer delays deployments, renegotiates capacity requirements or changes its infrastructure strategy, the consequences could be significant.

The economics of large AI infrastructure agreements also depend on the financial strength of the customers making those commitments.

A $132.5 billion backlog is valuable only to the extent that customers can fulfill their obligations and CoreWeave can deliver the contracted infrastructure.

The industry must also distinguish between independent end-user demand and commitments that exist within an increasingly interconnected ecosystem of strategic investments, financing arrangements and supplier relationships.

NVIDIA: supplier, investor and capacity backstop

The relationship between CoreWeave and NVIDIA deserves particular attention.

NVIDIA is not simply supplying the GPUs that power CoreWeave's infrastructure.

It has also become an important financial participant in the company's expansion.

NVIDIA invested in CoreWeave, participated in its IPO, and subsequently announced a $2 billion equity investment in January 2026. It also entered into a $6.3 billion capacity backstop agreement under which it would purchase qualifying unsold CoreWeave capacity through April 2032.

That arrangement provides an important layer of commercial support.

But it also creates a question about the structure of the AI infrastructure market.

NVIDIA is simultaneously a major hardware supplier, an investor in the infrastructure provider and a potential buyer of capacity that CoreWeave cannot otherwise sell.

This is not evidence of wrongdoing or an indication that the underlying demand is artificial.

It does, however, illustrate how interconnected the AI infrastructure economy has become.

The financial health of a major computing provider increasingly depends on relationships involving the companies supplying its most important hardware, financing its expansion, and purchasing its services.

Understanding those relationships is essential to evaluating the sustainability of the business.

Dell benefited from the same infrastructure boom

The original 2023 transaction also provides a useful perspective on Dell Technologies.

Dell's PowerEdge XE9680 systems helped CoreWeave establish its early H100 infrastructure footprint.

Since then, the hardware market has moved through successive GPU generations, including H200, Blackwell, and the emerging Vera Rubin architecture.

Dell's AI server business has grown dramatically alongside that transition.

The company has reported enormous AI server orders and a substantial backlog, positioning it as one of the major hardware beneficiaries of the AI infrastructure expansion.

But the financial exposure differs.

Dell is primarily selling hardware and associated infrastructure services, while CoreWeave assumes the responsibility of financing, deploying, and monetizing computing capacity over time.

Dell's accounts receivable and financing receivables also deserve attention as customer financing becomes a more important part of the infrastructure purchasing ecosystem.

There is no established evidence that Dell is a direct lender to CoreWeave. The more significant industry question is whether equipment manufacturers, infrastructure operators, financiers, and AI customers are becoming increasingly exposed to the same concentration of capital spending.

The AI infrastructure boom is creating enormous opportunities for hardware manufacturers.

But the financial sustainability of the companies purchasing that hardware remains a separate question.

Supercomputing is becoming an infrastructure business measured in gigawatts

There is another reason CoreWeave deserves attention from the supercomputing community.

In December 2025, an 8,192-NVIDIA-H100 GPU cluster operating in a CoreWeave Dallas data center topped the Graph500 breadth-first search benchmark with approximately 410 trillion traversed edges per second.

The result demonstrated that commercial cloud infrastructure can compete at the highest levels of large-scale computing.

It also illustrated how modern supercomputing is increasingly defined by distributed GPU systems, high-speed interconnects, massive memory bandwidth and sophisticated software orchestration.

CoreWeave's expansion is now measured not only in GPUs and customer contracts but also in electrical capacity.

The company reported more than 1.7 gigawatts of active power at the end of Q3 and expects to exceed 1.85 gigawatts by year-end.

That is an extraordinary infrastructure footprint.

Yet power availability is only one part of the equation. The company must also ensure that its computing systems are utilized sufficiently to justify their capital costs.

A gigawatt of available electrical capacity does not automatically translate into a profitable supercomputing business.

The economics depend on utilization, workload demand, hardware efficiency, electricity prices, cooling requirements, networking costs, and the revenue generated per unit of deployed capacity.

These are precisely the metrics that will determine which AI infrastructure companies become sustainable long-term computing providers.

The question is no longer how fast CoreWeave can grow

CoreWeave’s leadership seems pretty bullish, bumping their revenue guidance up to $13 billion to $13.3 billion and eyeing a $20 billion to $21 billion run rate by the end of 2026. They’re also expecting about $9 billion in prepayments and deposits soon, which helps with cash flow but definitely adds pressure to actually deliver the goods. On top of that, they’ve locked in $4.2 billion via convertible notes, set up a new equity program, and grabbed a $2.6 billion loan. It’s a clear sign that while they’ve got plenty of ways to fund this massive growth, they’re still leaning heavily on outside cash and customer money to keep the lights on.