{"success":true,"data":{"pressRelease":{"id":"175485","rtpr_id":"nBw454ZBha-20261021","ticker":"HYFT","exchange":"NASDAQ","all_tickers":["HYFT"],"title":"MindWalk Screens 890,634 Antibody Sequences for Immune Risk in Under Five Hours","author":"Business Wire","published_at":"2026-10-21T12:01:00.179Z","article_body":"MindWalk Screens 890,634 Antibody Sequences for Immune Risk in Under Five\nHours\n\nNew white paper shows LensAI™ immunogenicity screening running at up to 43%\nlower cost per sequence on Amazon EC2 Hpc8a instances powered by AMD EPYC\nprocessors, making a critical drug-safety check practical at the scale\nAI-driven discovery now demands\n\nMindWalk Holdings Corp. (\"MindWalk\") (NASDAQ: HYFT), a Bio-Native AI company,\ntoday published a white paper showing that its LensAI™ immunogenicity\nscreening ran across an entire public antibody database of 890,634 sequences\nin 4 hours and 41 minutes on Amazon EC2 Hpc8a instances powered by 5th Gen AMD\nEPYC™ processors, delivering up to 43% lower cost per sequence than a\ncomparable Intel configuration running identical code.\n\nThis press release features multimedia. View the full release here:\nhttps://www.businesswire.com/news/home/20261021750576/en/\n(https://www.businesswire.com/news/home/20261021750576/en/)\n\n“Every antibody drug carries a question: will the patient’s own immune\nsystem attack it,” said Dr. Jennifer Bath, President and Chief Executive\nOfficer of MindWalk. “Answering that question late is one of the most\nexpensive failures in this industry. Answering it early has been impractical,\nbecause AI now designs candidates far faster than anyone could screen them.\nThis work shows LensAI screening all of them. Candidate selection and safety\nassessment are no longer the bottleneck. It also lowers what it costs us to\ndeliver at this scale, so we can take on more volume without our costs rising\nin step with it.”\n\nThe problem the paper addresses\n\nWhen the body treats a therapeutic antibody as foreign, it produces anti-drug\nantibodies. These can blunt the drug's effect, change how it moves through the\nbody, and raise safety concerns. Discovering that risk in late-stage\ndevelopment can delay or end a program after years of investment. Regulators\nincreasingly expect the risk to be assessed at the start of discovery instead.\n\nThe obstacle has been volume. Modern AI design methods generate thousands of\ncandidate molecules, and a single antibody heavy chain expands into 2,420\nvariants under a routine single-point mutation scan. Checking all of them has\nbeen too slow and too expensive to be part of everyday work.\n\nWhat the benchmark measured\n\nThe paper reports LensAI immunogenicity screening run on three cloud\nconfigurations with no code changes between them, repeated five times and\naveraged.\n\nOn the identical workload, Amazon EC2 Hpc8a instances delivered 37% faster\nperformance and approximately 43% lower cost per sequence compared with an\nAmazon EC2 C8i instance powered by Intel(®) Xeon(®) processors. A\ncompute-optimized Amazon EC2 C8a configuration delivered 35% faster\nperformance and approximately 29% lower cost per sequence.\n\nAt full scale on Amazon EC2 Hpc8a, LensAI screened the Fitness Landscape for\nAntibodies (FLAb) database of 890,634 heavy and light chain sequences in 4\nhours and 41 minutes, and the Patent and Literature Antibody Database\n(PLAbDab) of 348,550 sequences in 2 hours and 12 minutes.\n\nResults reflect the specific configurations evaluated and are not\nrepresentative of all workloads.\n\nWhere the advantage comes from\n\nLensAI produces a composite immune-risk score from three inputs: HYFT-based\nhumanness scoring, MHC II binding prediction, and population allele weighting.\n\nThe humanness step is where MindWalk's data layer does the work. Conventional\napproaches compare every new sequence against a full reference library, so the\ncost of the comparison grows with the size of both the candidate set and the\nreference. LensAI instead matches sequences against HYFT(®) patterns that are\nindexed once and reused, which turns the comparison into a lookup, and\nassesses every antibody against shared biological context rather than treating\nit as an isolated file. In the published work, that step completed in roughly\none minute for 6,220 sequences on a single 64-core machine.\n\nThat design has a second consequence. Because the remaining work is divided\ninto independent units that share no state, the pipeline scales to use every\ncore made available to it, which is what turns high core counts directly into\nfinished results. The Amazon EC2 Hpc8a instances used in the benchmark provide\nup to 192 physical cores across two sockets, with simultaneous multithreading\ndisabled so every virtual core maps to a physical one, alongside the AVX-512,\nVNNI, and bfloat16 instructions that accelerate the MHC II binding models\ndominating runtime. Each successive generation of AMD EPYC processors has\ndelivered double-digit performance gains running the same LensAI code.\n\nA deepening engineering relationship with AMD\n\nThe work published today builds on MindWalk’s previous work using AMD\ntechnologies, spanning AMD Instinct™ GPUs and AMD EPYC server processors.\n\nIn July 2026, ReefIQ™ made its public debut at AMD Advancing AI 2026. AMD\nsubsequently published a case study on MindWalk's use of AMD Instinct MI300X\nGPUs for large-scale antibody screening. In September 2026, MindWalk brought\nthe open-source OpenFold3 model into production using the AMD Enterprise AI\nReference stack, running on Vultr Kubernetes Engine with AMD Instinct MI325X\nGPUs. Across those engagements, MindWalk returned findings on multi-GPU\nexecution, inference control, and deployment workflow to AMD and its cloud\npartners. The benchmark published today extends MindWalk’s use of AMD\ntechnologies to AMD EPYC server processors.\n\n\"Our work with AMD now spans their accelerators and server CPUs,\" said Dirk\nVan Hyfte, MD, PhD, Chief Technology Officer of MindWalk. “That gives us a\ncompute foundation that can scale with the complexity and volume of modern\ndrug discovery. Just as importantly, we can continue taking advantage of\nadvances in AMD EPYC processors without having to rewrite the underlying\nLensAI code.”\n\nWhy the result extends beyond one workflow\n\nImmunogenicity screening is one of several LensAI workflows that run on\nReefIQ™, MindWalk's biological context layer. The characteristics that\nproduced these results, indexed HYFT patterns and a pipeline that can be split\nacross any number of machines, belong to that shared layer rather than to this\none workflow, and are designed to carry to others built on it. Enrichment\nadded to ReefIQ for one program also reaches programs already on the platform.\n\nFor MindWalk, the commercial consequence sits in the cost of delivery. As\nenterprise platform agreements move the company toward contracted, recurring\nrevenue, the expense of running the underlying workflows becomes a direct\ninput to margin. Reducing that expense by up to 43% on one of the heaviest\nworkloads in the catalog, without changing the software, is intended to let\nMindWalk absorb additional volume without a proportional increase in compute\ncost.\n\nWhat it changes for discovery teams\n\nScreening at this speed and cost lets teams flag high-risk sequences before\nsynthesis, rank and triage thousands of candidates by population-level immune\nrisk in hours rather than weeks, and run exhaustive in silico engineering such\nas scanning every single amino acid substitution across an antibody's heavy\nchain. Work that was previously reserved for a short list of finalists becomes\navailable across a full design space.\n\nIn a study covering 217 therapeutic antibodies cited in the paper, LensAI\nclassified 197 of the 217, approximately 91%, as having an anti-drug antibody\nincidence risk greater than 30%.\n\nAvailability\n\nLensAI™ immunogenicity screening is commercially available from MindWalk.\nAmazon EC2 C8a and Hpc8a instances, powered by AMD EPYC™ processors, are\navailable through Amazon Web Services.\n\nThe white paper, “Built to Inherit the Curve\n(https://cts.businesswire.com/ct/CT?id=smartlink&url=https%3A%2F%2Fwww.mindwalkai.com%2Fhubfs%2FMindWalk%2FWhite-Papers%2FMindWalk-LensAI-Immunogenicity-Screening-AMD-EPYC-White-Paper-2026.pdf&esheet=54626287&newsitemid=20261021750576&lan=en-US&anchor=Built+to+Inherit+the+Curve&index=1&md5=214bc17458580939c014a3973e3cfee5)\n,” is available at www.mindwalkai.com\n(https://cts.businesswire.com/ct/CT?id=smartlink&url=http%3A%2F%2Fwww.mindwalkai.com&esheet=54626287&newsitemid=20261021750576&lan=en-US&anchor=www.mindwalkai.com&index=2&md5=d3f0f923284a2c01914309291080ff42)\n.\n\nAbout MindWalk Holdings Corp.\n\nMindWalk Holdings Corp. (NASDAQ: HYFT) is a Bio-Native AI company building the\nBioIntelligence infrastructure that life sciences AI and agentic AI require,\nintegrating AI, data, and advanced wet lab capabilities into one connected\ndiscovery ecosystem. At its core is HYFT(®) Technology, a proprietary,\nfunction-aware representation of biology. Its HYFT fingerprints span sequence\nand structural biology and, refined over 20 years of curation, form a\ncontinuously evolving biological representation of 660 million patterns and 25\nbillion relationships. This enriched biological representation is the\narchitecture behind ReefIQ™, the biological data substrate that provides\ncontext for life sciences, enriching data at ingestion and growing more\nvaluable with each program run on it, and LensAI™, the reasoning and\napplication layer for target discovery, candidate diligence, portfolio\ndecision support, and the agentic AI workflows pharma is racing to deploy. By\ndesign, value compounds in this HYFT representation layer, not in any\nindividual AI model that runs on top of it.\n\nForward-Looking Statements\n\nThis press release contains forward-looking statements within the meaning of\napplicable United States and Canadian securities laws. Forward-looking\nstatements are based on management's current expectations and assumptions and\nare subject to known and unknown risks, uncertainties, and other factors that\nmay cause actual results to differ materially, including: that the benchmark\nresults described are specific to the workloads, datasets, processor\nconfigurations, cloud regions, and pricing tested in July 2026, and may not be\nrepresentative of other workloads, configurations, pricing, or future hardware\ngenerations; that performance and cost figures may change; MindWalk's reliance\non third-party compute and cloud providers, including AMD, Amazon Web\nServices, and Vultr, and their respective product roadmaps, availability, and\npricing; that the engineering collaborations with AMD described in this\nrelease are project-based and non-exclusive, may not continue or expand, and\nare not expected to generate revenue; the expectation that future processor\ngenerations will deliver performance gains to the same software without\nmodification, which may not occur; the expectation that architectural\ncharacteristics demonstrated in this workflow will extend to other LensAI™\nworkflows or reduce the cost of delivery at enterprise scale, which may not\noccur; the pace and scale of commercial adoption of LensAI™ and ReefIQ™\nand the ability to convert technical results into contracted, recurring\narrangements; the fact that immunogenicity predictions are statistical in\nnature and do not guarantee clinical outcomes, regulatory acceptance, or\nproduct success; competition; intellectual property risks; regulatory\ndeterminations; and capital markets conditions. Additional information is\navailable in MindWalk's Annual Report on Form 20-F and other filings on SEDAR+\n(sedarplus.ca) and EDGAR (sec.gov/edgar). Except as required by law, MindWalk\nundertakes no obligation to update any forward-looking statement. Nothing in\nthis press release constitutes investment advice. This press release does not\nconstitute an offer to sell or the solicitation of an offer to buy any\nsecurities.\n\nTrademarks\n\nHYFT(®) is a registered trademark of MindWalk Holdings Corp. LensAI™ and\nReefIQ™ are trademarks of MindWalk Holdings Corp. or its subsidiaries;\nReefIQ™ registration is pending. All other trademarks are the property of\ntheir respective owners.\n\nAMD, the AMD Arrow logo, AMD EPYC, the EPYC logo, AMD Instinct, and\ncombinations thereof are trademarks of Advanced Micro Devices, Inc.\n\nAmazon Web Services, AWS, Amazon EC2, and related marks are trademarks of\nAmazon.com, Inc. or its affiliates. Intel and Xeon are trademarks of Intel\nCorporation. References to Amazon Web Services and Intel describe the compute\nenvironment used to run these benchmarks and do not imply any partnership,\nsponsorship, or endorsement.\n\n\n\nView source version on businesswire.com:\nhttps://www.businesswire.com/news/home/20261021750576/en/\n(https://www.businesswire.com/news/home/20261021750576/en/)\n\nInvestor Relations Contact \n\nLouie Toma, CPA, CFA\n\nManaging Director, CoreIR\n\ninvestors@mindwalkai.com (mailto:investors@mindwalkai.com)\n\n\nCopyright Business Wire 2026","article_body_html":"","raw_payload":{"data":{"id":"nBw454ZBha-20261021","title":"MindWalk Screens 890,634 Antibody Sequences for Immune Risk in Under Five Hours","author":"Business Wire","ticker":"HYFT","created":"2026-10-21T12:01:00.179Z","tickers":["HYFT"],"exchange":"NASDAQ","article_body":"MindWalk Screens 890,634 Antibody Sequences for Immune Risk in Under Five\nHours\n\nNew white paper shows LensAI™ immunogenicity screening running at up to 43%\nlower cost per sequence on Amazon EC2 Hpc8a instances powered by AMD EPYC\nprocessors, making a critical drug-safety check practical at the scale\nAI-driven discovery now demands\n\nMindWalk Holdings Corp. (\"MindWalk\") (NASDAQ: HYFT), a Bio-Native AI company,\ntoday published a white paper showing that its LensAI™ immunogenicity\nscreening ran across an entire public antibody database of 890,634 sequences\nin 4 hours and 41 minutes on Amazon EC2 Hpc8a instances powered by 5th Gen AMD\nEPYC™ processors, delivering up to 43% lower cost per sequence than a\ncomparable Intel configuration running identical code.\n\nThis press release features multimedia. View the full release here:\nhttps://www.businesswire.com/news/home/20261021750576/en/\n(https://www.businesswire.com/news/home/20261021750576/en/)\n\n“Every antibody drug carries a question: will the patient’s own immune\nsystem attack it,” said Dr. Jennifer Bath, President and Chief Executive\nOfficer of MindWalk. “Answering that question late is one of the most\nexpensive failures in this industry. Answering it early has been impractical,\nbecause AI now designs candidates far faster than anyone could screen them.\nThis work shows LensAI screening all of them. Candidate selection and safety\nassessment are no longer the bottleneck. It also lowers what it costs us to\ndeliver at this scale, so we can take on more volume without our costs rising\nin step with it.”\n\nThe problem the paper addresses\n\nWhen the body treats a therapeutic antibody as foreign, it produces anti-drug\nantibodies. These can blunt the drug's effect, change how it moves through the\nbody, and raise safety concerns. Discovering that risk in late-stage\ndevelopment can delay or end a program after years of investment. Regulators\nincreasingly expect the risk to be assessed at the start of discovery instead.\n\nThe obstacle has been volume. Modern AI design methods generate thousands of\ncandidate molecules, and a single antibody heavy chain expands into 2,420\nvariants under a routine single-point mutation scan. Checking all of them has\nbeen too slow and too expensive to be part of everyday work.\n\nWhat the benchmark measured\n\nThe paper reports LensAI immunogenicity screening run on three cloud\nconfigurations with no code changes between them, repeated five times and\naveraged.\n\nOn the identical workload, Amazon EC2 Hpc8a instances delivered 37% faster\nperformance and approximately 43% lower cost per sequence compared with an\nAmazon EC2 C8i instance powered by Intel(®) Xeon(®) processors. A\ncompute-optimized Amazon EC2 C8a configuration delivered 35% faster\nperformance and approximately 29% lower cost per sequence.\n\nAt full scale on Amazon EC2 Hpc8a, LensAI screened the Fitness Landscape for\nAntibodies (FLAb) database of 890,634 heavy and light chain sequences in 4\nhours and 41 minutes, and the Patent and Literature Antibody Database\n(PLAbDab) of 348,550 sequences in 2 hours and 12 minutes.\n\nResults reflect the specific configurations evaluated and are not\nrepresentative of all workloads.\n\nWhere the advantage comes from\n\nLensAI produces a composite immune-risk score from three inputs: HYFT-based\nhumanness scoring, MHC II binding prediction, and population allele weighting.\n\nThe humanness step is where MindWalk's data layer does the work. Conventional\napproaches compare every new sequence against a full reference library, so the\ncost of the comparison grows with the size of both the candidate set and the\nreference. LensAI instead matches sequences against HYFT(®) patterns that are\nindexed once and reused, which turns the comparison into a lookup, and\nassesses every antibody against shared biological context rather than treating\nit as an isolated file. In the published work, that step completed in roughly\none minute for 6,220 sequences on a single 64-core machine.\n\nThat design has a second consequence. Because the remaining work is divided\ninto independent units that share no state, the pipeline scales to use every\ncore made available to it, which is what turns high core counts directly into\nfinished results. The Amazon EC2 Hpc8a instances used in the benchmark provide\nup to 192 physical cores across two sockets, with simultaneous multithreading\ndisabled so every virtual core maps to a physical one, alongside the AVX-512,\nVNNI, and bfloat16 instructions that accelerate the MHC II binding models\ndominating runtime. Each successive generation of AMD EPYC processors has\ndelivered double-digit performance gains running the same LensAI code.\n\nA deepening engineering relationship with AMD\n\nThe work published today builds on MindWalk’s previous work using AMD\ntechnologies, spanning AMD Instinct™ GPUs and AMD EPYC server processors.\n\nIn July 2026, ReefIQ™ made its public debut at AMD Advancing AI 2026. AMD\nsubsequently published a case study on MindWalk's use of AMD Instinct MI300X\nGPUs for large-scale antibody screening. In September 2026, MindWalk brought\nthe open-source OpenFold3 model into production using the AMD Enterprise AI\nReference stack, running on Vultr Kubernetes Engine with AMD Instinct MI325X\nGPUs. Across those engagements, MindWalk returned findings on multi-GPU\nexecution, inference control, and deployment workflow to AMD and its cloud\npartners. The benchmark published today extends MindWalk’s use of AMD\ntechnologies to AMD EPYC server processors.\n\n\"Our work with AMD now spans their accelerators and server CPUs,\" said Dirk\nVan Hyfte, MD, PhD, Chief Technology Officer of MindWalk. “That gives us a\ncompute foundation that can scale with the complexity and volume of modern\ndrug discovery. Just as importantly, we can continue taking advantage of\nadvances in AMD EPYC processors without having to rewrite the underlying\nLensAI code.”\n\nWhy the result extends beyond one workflow\n\nImmunogenicity screening is one of several LensAI workflows that run on\nReefIQ™, MindWalk's biological context layer. The characteristics that\nproduced these results, indexed HYFT patterns and a pipeline that can be split\nacross any number of machines, belong to that shared layer rather than to this\none workflow, and are designed to carry to others built on it. Enrichment\nadded to ReefIQ for one program also reaches programs already on the platform.\n\nFor MindWalk, the commercial consequence sits in the cost of delivery. As\nenterprise platform agreements move the company toward contracted, recurring\nrevenue, the expense of running the underlying workflows becomes a direct\ninput to margin. Reducing that expense by up to 43% on one of the heaviest\nworkloads in the catalog, without changing the software, is intended to let\nMindWalk absorb additional volume without a proportional increase in compute\ncost.\n\nWhat it changes for discovery teams\n\nScreening at this speed and cost lets teams flag high-risk sequences before\nsynthesis, rank and triage thousands of candidates by population-level immune\nrisk in hours rather than weeks, and run exhaustive in silico engineering such\nas scanning every single amino acid substitution across an antibody's heavy\nchain. Work that was previously reserved for a short list of finalists becomes\navailable across a full design space.\n\nIn a study covering 217 therapeutic antibodies cited in the paper, LensAI\nclassified 197 of the 217, approximately 91%, as having an anti-drug antibody\nincidence risk greater than 30%.\n\nAvailability\n\nLensAI™ immunogenicity screening is commercially available from MindWalk.\nAmazon EC2 C8a and Hpc8a instances, powered by AMD EPYC™ processors, are\navailable through Amazon Web Services.\n\nThe white paper, “Built to Inherit the Curve\n(https://cts.businesswire.com/ct/CT?id=smartlink&url=https%3A%2F%2Fwww.mindwalkai.com%2Fhubfs%2FMindWalk%2FWhite-Papers%2FMindWalk-LensAI-Immunogenicity-Screening-AMD-EPYC-White-Paper-2026.pdf&esheet=54626287&newsitemid=20261021750576&lan=en-US&anchor=Built+to+Inherit+the+Curve&index=1&md5=214bc17458580939c014a3973e3cfee5)\n,” is available at www.mindwalkai.com\n(https://cts.businesswire.com/ct/CT?id=smartlink&url=http%3A%2F%2Fwww.mindwalkai.com&esheet=54626287&newsitemid=20261021750576&lan=en-US&anchor=www.mindwalkai.com&index=2&md5=d3f0f923284a2c01914309291080ff42)\n.\n\nAbout MindWalk Holdings Corp.\n\nMindWalk Holdings Corp. (NASDAQ: HYFT) is a Bio-Native AI company building the\nBioIntelligence infrastructure that life sciences AI and agentic AI require,\nintegrating AI, data, and advanced wet lab capabilities into one connected\ndiscovery ecosystem. At its core is HYFT(®) Technology, a proprietary,\nfunction-aware representation of biology. Its HYFT fingerprints span sequence\nand structural biology and, refined over 20 years of curation, form a\ncontinuously evolving biological representation of 660 million patterns and 25\nbillion relationships. This enriched biological representation is the\narchitecture behind ReefIQ™, the biological data substrate that provides\ncontext for life sciences, enriching data at ingestion and growing more\nvaluable with each program run on it, and LensAI™, the reasoning and\napplication layer for target discovery, candidate diligence, portfolio\ndecision support, and the agentic AI workflows pharma is racing to deploy. By\ndesign, value compounds in this HYFT representation layer, not in any\nindividual AI model that runs on top of it.\n\nForward-Looking Statements\n\nThis press release contains forward-looking statements within the meaning of\napplicable United States and Canadian securities laws. Forward-looking\nstatements are based on management's current expectations and assumptions and\nare subject to known and unknown risks, uncertainties, and other factors that\nmay cause actual results to differ materially, including: that the benchmark\nresults described are specific to the workloads, datasets, processor\nconfigurations, cloud regions, and pricing tested in July 2026, and may not be\nrepresentative of other workloads, configurations, pricing, or future hardware\ngenerations; that performance and cost figures may change; MindWalk's reliance\non third-party compute and cloud providers, including AMD, Amazon Web\nServices, and Vultr, and their respective product roadmaps, availability, and\npricing; that the engineering collaborations with AMD described in this\nrelease are project-based and non-exclusive, may not continue or expand, and\nare not expected to generate revenue; the expectation that future processor\ngenerations will deliver performance gains to the same software without\nmodification, which may not occur; the expectation that architectural\ncharacteristics demonstrated in this workflow will extend to other LensAI™\nworkflows or reduce the cost of delivery at enterprise scale, which may not\noccur; the pace and scale of commercial adoption of LensAI™ and ReefIQ™\nand the ability to convert technical results into contracted, recurring\narrangements; the fact that immunogenicity predictions are statistical in\nnature and do not guarantee clinical outcomes, regulatory acceptance, or\nproduct success; competition; intellectual property risks; regulatory\ndeterminations; and capital markets conditions. Additional information is\navailable in MindWalk's Annual Report on Form 20-F and other filings on SEDAR+\n(sedarplus.ca) and EDGAR (sec.gov/edgar). Except as required by law, MindWalk\nundertakes no obligation to update any forward-looking statement. Nothing in\nthis press release constitutes investment advice. This press release does not\nconstitute an offer to sell or the solicitation of an offer to buy any\nsecurities.\n\nTrademarks\n\nHYFT(®) is a registered trademark of MindWalk Holdings Corp. LensAI™ and\nReefIQ™ are trademarks of MindWalk Holdings Corp. or its subsidiaries;\nReefIQ™ registration is pending. All other trademarks are the property of\ntheir respective owners.\n\nAMD, the AMD Arrow logo, AMD EPYC, the EPYC logo, AMD Instinct, and\ncombinations thereof are trademarks of Advanced Micro Devices, Inc.\n\nAmazon Web Services, AWS, Amazon EC2, and related marks are trademarks of\nAmazon.com, Inc. or its affiliates. Intel and Xeon are trademarks of Intel\nCorporation. References to Amazon Web Services and Intel describe the compute\nenvironment used to run these benchmarks and do not imply any partnership,\nsponsorship, or endorsement.\n\n\n\nView source version on businesswire.com:\nhttps://www.businesswire.com/news/home/20261021750576/en/\n(https://www.businesswire.com/news/home/20261021750576/en/)\n\nInvestor Relations Contact \n\nLouie Toma, CPA, CFA\n\nManaging Director, CoreIR\n\ninvestors@mindwalkai.com (mailto:investors@mindwalkai.com)\n\n\nCopyright Business Wire 2026"},"type":"article","timestamp":"2026-10-21T12:01:00.259727864Z","server_sent_at_ms":1792584060259},"received_at":"2026-10-21T12:01:00.432Z","source_url":"https://www.businesswire.com/news/home/20261021750576/en/"},"analysis":{"id":"163499","press_release_id":"175485","analysis_json":{"industry":{"label":"Pharmaceuticals, Biotechnology & Life Sciences","sector":"Health Care"},"redFlags":["Benchmark results are configuration-specific and 'not representative of all workloads' per the company's own disclaimer","Forward-looking statements note AMD collaborations are non-exclusive and not expected to generate revenue","The cited study classifying ~91% of 217 therapeutic antibodies at >30% anti-drug antibody risk may raise questions about platform signal quality"],"eventType":"operations_update","narrative":"MindWalk's white paper shows its LensAI immunogenicity screening processed 890,634 antibody sequences in 4 hours 41 minutes on Amazon EC2 Hpc8a instances powered by AMD EPYC processors.\n\nThe benchmark reported up to 43% lower cost per sequence and 37% faster performance versus a comparable Intel-based configuration, with no code changes across configurations.\n\nThe company frames the compute savings as a direct margin lever as enterprise platform agreements move it toward contracted, recurring revenue, and the AMD relationship now spans accelerators and server CPUs.","sentiment":"bullish","agentHooks":{"shouldPost":false,"suggestedAngle":"LensAI cost-per-sequence cut up to 43% on AMD EPYC -- margin leverage story as MindWalk scales contracted enterprise volume."},"keyFigures":{"customDimensions":{"runtime_hpc8a":"4 hours 41 minutes","plabdab_runtime":"2 hours 12 minutes","plabdab_sequences":348550,"cost_reduction_c8a":"approximately 29%","sequences_screened":890634,"performance_gain_c8a":"35% faster","performance_gain_hpc8a":"37% faster","anti_drug_antibody_risk":"197 of 217 (~91%) therapeutic antibodies classified with ADAO incidence risk >30%","cost_per_sequence_reduction":"up to 43%"}},"quotedText":"It also lowers what it costs us to\ndeliver at this scale, so we can take on more volume without our costs rising\nin step with it.","namedEntities":{"people":[{"name":"Dr. Jennifer Bath","role":"President and CEO of MindWalk"},{"name":"Dirk Van Hyfte","role":"Chief Technology Officer of MindWalk"},{"name":"Louie Toma","role":"Managing Director, CoreIR (investor relations contact)"}],"products":["LensAI","ReefIQ","HYFT Technology","AMD EPYC","AMD Instinct MI300X","AMD Instinct MI325X","Amazon EC2 Hpc8a","Amazon EC2 C8a","Amazon EC2 C8i","OpenFold3","FLAb database","PLAbDab"],"companies":[{"name":"MindWalk Holdings Corp.","ticker":"HYFT","relationship":"filer"},{"name":"AMD","relationship":"technology partner"},{"name":"Amazon Web Services","relationship":"cloud infrastructure provider"},{"name":"Vultr","relationship":"cloud infrastructure provider"},{"name":"Intel","relationship":"benchmark comparison vendor"},{"name":"CoreIR","relationship":"investor relations firm"}],"dollarAmounts":[]},"materialImpact":{"score":2,"reasoning":"Vendor-style white paper demonstrating LensAI performance/cost improvements and a deepened AMD relationship. Positive for cost/margin narrative but a benchmark announcement, not a revenue or regulatory event."},"tickerRelevance":{"others":[{"ticker":"AMD","relevance":"technology partner (EPYC/Instinct hardware)"},{"ticker":"AMZN","relevance":"cloud infrastructure provider (AWS EC2 instances)"},{"ticker":"INTC","relevance":"mentioned (comparison benchmark baseline)"}],"primary":"HYFT"},"globalImportance":22,"audienceRelevance":18,"eventTypeSecondary":["product_launch"],"importanceComponents":{"tickerTier":"small-cap biotech AI","eventGravity":"vendor white paper / benchmark","revenueImpact":"indirect (cost/margin narrative)","issuerAuthored":true}},"event_type":"operations_update","event_type_secondary":["product_launch"],"sentiment":"bullish","material_impact_score":2,"narrative":"MindWalk's white paper shows its LensAI immunogenicity screening processed 890,634 antibody sequences in 4 hours 41 minutes on Amazon EC2 Hpc8a instances powered by AMD EPYC processors.\n\nThe benchmark reported up to 43% lower cost per sequence and 37% faster performance versus a comparable Intel-based configuration, with no code changes across configurations.\n\nThe company frames the compute savings as a direct margin lever as enterprise platform agreements move it toward contracted, recurring revenue, and the AMD relationship now spans accelerators and server CPUs.","key_figures":{"customDimensions":{"runtime_hpc8a":"4 hours 41 minutes","plabdab_runtime":"2 hours 12 minutes","plabdab_sequences":348550,"cost_reduction_c8a":"approximately 29%","sequences_screened":890634,"performance_gain_c8a":"35% faster","performance_gain_hpc8a":"37% faster","anti_drug_antibody_risk":"197 of 217 (~91%) therapeutic antibodies classified with ADAO incidence risk >30%","cost_per_sequence_reduction":"up to 43%"}},"named_entities":{"people":[{"name":"Dr. Jennifer Bath","role":"President and CEO of MindWalk"},{"name":"Dirk Van Hyfte","role":"Chief Technology Officer of MindWalk"},{"name":"Louie Toma","role":"Managing Director, CoreIR (investor relations contact)"}],"products":["LensAI","ReefIQ","HYFT Technology","AMD EPYC","AMD Instinct MI300X","AMD Instinct MI325X","Amazon EC2 Hpc8a","Amazon EC2 C8a","Amazon EC2 C8i","OpenFold3","FLAb database","PLAbDab"],"companies":[{"name":"MindWalk Holdings Corp.","ticker":"HYFT","relationship":"filer"},{"name":"AMD","relationship":"technology partner"},{"name":"Amazon Web Services","relationship":"cloud infrastructure provider"},{"name":"Vultr","relationship":"cloud infrastructure provider"},{"name":"Intel","relationship":"benchmark comparison vendor"},{"name":"CoreIR","relationship":"investor relations firm"}],"dollarAmounts":[]},"model_name":"glm-5.3-flashx","prompt_hash":"sha256:5c812bd9e28b810d","schema_hash":"sha256:05005c02d9cffac9","created_at":"2026-10-21T12:01:07.873Z","global_importance":22,"audience_relevance":18,"importance_components":{"tickerTier":"small-cap biotech AI","eventGravity":"vendor white paper / benchmark","revenueImpact":"indirect (cost/margin narrative)","issuerAuthored":true}},"durationMs":7421,"modelName":"glm-5.3-flashx"}}