Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsJPMorgan Chase is using AWS services including Amazon Bedrock and Amazon SageMaker to move selected artificial-intelligence workloads from experimentation into production. But AWS is not the bank’s entire AI platform. JPMorgan describes a broader, multi-cloud technology strategy that combines public and private cloud, governed enterprise data, multiple AI models, internal tools such as LLM Suite, and strict security and model-risk controls.
What JPMorgan described at AWS re:Invent
At AWS re:Invent in December 2024, JPMorgan Global CIO Lori Beer described an effort to take Amazon SageMaker and Amazon Bedrock beyond proofs of concept and into production applications. The emphasis was not on moving every banking workload to one cloud at once. Instead, JPMorgan is modernizing selectively, prioritizing use cases where AI can create meaningful operational or business value.
As an Amazon Associate I earn from qualifying purchases.
That approach reflects the constraints of a global bank. AI systems must operate within requirements for data protection, resilience, auditability, access control, regulatory compliance, and model oversight. JPMorgan’s AWS relationship therefore appears to be an important part of a controlled modernization program—not evidence of an AWS-only architecture or a newly announced exclusive contract. CIO’s coverage of the re:Invent discussion provides the AWS-specific context.
What JPMorgan means by an AI foundation
JPMorgan’s AI foundation is best understood as a stack and operating model rather than a single product:
#1 Best Overall
- 𝗢𝗻𝗲 𝗦𝘄𝗶𝘁𝗰𝗵 𝗠𝗮𝗱𝗲 𝘁𝗼 𝗘𝘅𝗽𝗮𝗻𝗱 𝗡𝗲𝘁𝘄𝗼𝗿𝗸: 5× 10/100/1000Mbps RJ45 Ports supporting Auto Negotiation and Auto MDI/MDIX.
- 𝗚𝗶𝗴𝗮𝗯𝗶𝘁 𝘁𝗵𝗮𝘁 𝗦𝗮𝘃𝗲𝘀 𝗘𝗻𝗲𝗿𝗴𝘆: Latest innovative energy-efficient technology greatly expands your network capacity with much less power consumption and helps save money.
- 𝗥𝗲𝗹𝗶𝗮𝗯𝗹𝗲 𝗮𝗻𝗱 𝗤𝘂𝗶𝗲𝘁: IEEE 802.3X flow control provides reliable data transfer and Fanless design ensures quiet operation.
- 𝗣𝗹𝘂𝗴 𝗮𝗻𝗱 𝗣𝗹𝗮𝘆: Easy setup with no software installation or configuration needed.
- 𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗱 𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀: Prioritize your traffic and guarantee high quality of video or voice data transmission with Port-based 802.1p/DSCP QoS and IGMP Snooping.
- Infrastructure: Data centers, public and private cloud, compute, storage, networking, security, and resilience.
- Data: Connected enterprise information with controls for quality, access, privacy, lineage, and use.
- Model services: Access to multiple machine-learning and foundation models rather than dependence on one provider.
- Shared platforms: Reusable services for model access, retrieval, evaluation, monitoring, and application development.
- Business applications: Tools for employees, call centers, fraud prevention, research, coding, operations, credit, trading, and customer service.
- Governance: Model validation, human review, audit logs, identity controls, records retention, incident response, and regulatory oversight.
This platform approach matters because separate business units do not need to independently select models, build data controls, and recreate the same security architecture for every AI experiment.
JPMorgan’s 2024 annual report describes a multi-cloud strategy spanning more than 6,000 applications and nearly an exabyte of data. Those figures describe the firm’s overall technology estate, not an AWS environment.
Where AWS fits
Amazon Bedrock: managed access to foundation models
Amazon Bedrock is the higher-level generative-AI service in this story. It gives organizations managed access to foundation models from multiple providers and tools for building applications around them. AWS materials describe capabilities including model selection, retrieval-augmented generation through Knowledge Bases, guardrails, evaluation, prompt optimization, and agents.
Those capabilities can help an enterprise build a governed application without operating every model server itself. A bank could, for example, connect approved internal information to a retrieval system, restrict inappropriate outputs, evaluate responses, and route requests among models suited to different tasks.
However, AWS’s capabilities should not automatically be read as a list of JPMorgan deployments. Public reporting establishes that JPMorgan is using or advancing Bedrock in its AI effort, but it does not show that the bank uses every Bedrock feature. AWS’s Bedrock product page and pricing page describe the available service and its usage-based economics.
Rank #2
- GIGABIT ETHERNET PORTS: Features 5 x 1.0Gbps Ethernet ports for high-speed connectivity. Auto-negotiating ports detect the optimal speed for connected devices and work with existing Cat5e or Cat6 Ethernet cables.
- PLUG-AND-PLAY UNMANAGED NETWORK SWITCH: Simple plug-and-play setup with no software to install or configuration required.
- FLEXIBLE MOUNTING OPTIONS: Compact metal design supports desktop or wall-mount placement for versatile installation.
- SILENT & ENERGY-EFFICIENT OPERATION: Fanless design ensures silent performance, while IEEE 802.3az Energy Efficient Ethernet reduces power consumption without compromising high-speed network performance.
- REGIONAL COMPATIBILITY: Made for use in U.S. & CA only
Amazon SageMaker AI: the broader machine-learning environment
SageMaker is more closely associated with the machine-learning lifecycle: data science, model development, training, deployment, hosting, and MLOps. It can provide greater control for teams developing or operating custom predictive models.
The practical distinction is:
| Need | Likely fit | Important qualification |
|---|---|---|
| Use a managed foundation model through an API | Bedrock | Model, region, modality, tier, and token usage affect cost and availability. |
| Build a retrieval-augmented generative-AI application | Bedrock and related data services | Embedding, storage, retrieval, evaluation, and monitoring costs also matter. |
| Train or tune specialized models | SageMaker or related ML services | Compute, storage, data movement, and operational resources must be budgeted. |
| Deploy conventional predictive machine learning | SageMaker | The platform requires more ML engineering and MLOps expertise than a simple hosted API. |
| Handle predictable, high-volume inference | Provisioned or reserved capacity | Commitments can reduce unit costs but may create waste when demand fluctuates. |
AWS’s Bedrock-versus-SageMaker decision guide distinguishes Bedrock’s model and API-oriented economics from SageMaker’s compute, storage, training, hosting, and operational-resource charges.
Recommended Free Tools
LLM Suite is JPMorgan’s product, not an AWS product
LLM Suite is JPMorgan’s internal generative-AI platform and employee-facing front door. JPMorgan launched it to eligible employees in 2024, giving them controlled access to large language models while protecting company and customer data. The platform is also intended to give business units shared capabilities for creating additional workflows and applications.
JPMorgan reported more than 200,000 LLM Suite users or employees with access in its 2024 reporting and 2025 Investor Day materials. That is a significant distribution and adoption signal, but it is not the same as 200,000 daily active users, external customers, or a measured productivity gain for every employee.
LLM Suite is described as model-agnostic. AWS may provide infrastructure or model-delivery services for portions of JPMorgan’s AI environment, but public evidence does not establish that LLM Suite runs wholly on AWS. Calling it “LLM Suite on AWS” would therefore overstate what is documented. JPMorgan’s own technology blog explains the platform’s internal role.
Rank #3
- 𝗘𝗶𝗴𝗵𝘁 𝟮.𝟱 𝗚𝗯𝗽𝘀 𝗣𝗼𝗿𝘁𝘀 𝗳𝗼𝗿 𝗦𝘂𝗽𝗲𝗿-𝗙𝗮𝘀𝘁 𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝗼𝗻𝘀: 8× 2.5-Gigabit ports unlock the highest performance of your Multi-Gig bandwidth and devices, and provide up to 40 Gbps of switching capacity.
- 𝗔𝘂𝘁𝗼-𝗡𝗲𝗴𝗼𝘁𝗶𝗮𝘁𝗶𝗼𝗻: Auto-negotiation intelligently senses the link speeds and adjusts between 3-speeds (100Mb/1G/2.5G) for compatibility and optimal performance for all your devices, including 2.5G WiFi 6 AP, 2.5G NAS, 2.5G PCIe Adapter, 2.5G Server, gaming computer, 4K video, and more.
- 𝗜𝗱𝗲𝗮𝗹 𝗳𝗼𝗿 𝗩𝗮𝗿𝗶𝗼𝘂𝘀 𝗦𝗰𝗲𝗻𝗮𝗿𝗶𝗼𝘀: Built for LAN parties, home entertainment, small and home offices, and instant transfer for workstations.
- 𝗛𝗮𝘀𝘀𝗹𝗲-𝗙𝗿𝗲𝗲 𝗖𝗮𝗯𝗹𝗶𝗻𝗴: Instantly upgrade to 2.5 Gbps without the need to upgrade to Cat6 wiring, reducing wiring costs and hassle. *
- 𝗦𝗶𝗹𝗲𝗻𝘁 𝗢𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻: Industry-leading fanless design ensures silent operation, ideal for any home or business.
From pilots to production
JPMorgan’s reported numbers show that AI is moving beyond isolated demonstrations, although the figures measure different things and should not be added together.
- Approximately 100 generative-AI solutions were reported in production in JPMorgan’s 2025 Investor Day materials.
- More than 40,000 engineers were reported as using AI coding assistants. Usage does not prove a uniform productivity improvement.
- More than 175 AI use cases were reported by JPMorgan’s Commercial & Investment Bank in its 2024 annual-report materials. That business-unit count may include traditional machine learning as well as generative AI.
- Roughly 80% of applications were reported as processing primarily in public or private cloud. This firmwide statistic does not mean that 80% run on AWS.
Reported use cases span employee productivity, software engineering, call-center assistance, fraud detection, credit decisioning, pricing and hedging, trading, operations, personalization, research, know-your-customer processes, sanctions screening, cash-flow prediction, and capital optimization.
JPMorgan also reported a nearly 40% reduction in KYC unit costs associated with AI and machine learning, alongside improvements in sanctions screening and fraud protection. That is a company-reported business-unit result. It should not be attributed solely to generative AI or AWS.
Why the banking context changes the design
A consumer chatbot can be judged mainly on convenience. A bank’s AI system must also answer harder questions: Which data was used? Which model processed it? Can the output be audited? Was a human involved? What happens when the system is wrong? Can the firm demonstrate that the process treats customers fairly and complies with applicable rules?
A JPMorgan-like architecture therefore needs controls such as:
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #4
- GIGABIT ETHERNET PORTS: Features 8 x 1.0Gbps Ethernet ports for high-speed connectivity. Auto-negotiating ports detect the optimal speed for connected devices and work with existing Cat5e or Cat6 Ethernet cables.
- PLUG-AND-PLAY UNMANAGED NETWORK SWITCH: Simple plug-and-play setup with no software to install or configuration required.
- FLEXIBLE MOUNTING OPTIONS: Compact metal design supports desktop or wall-mount placement for versatile installation.
- SILENT & ENERGY-EFFICIENT OPERATION: Fanless design ensures silent performance, while IEEE 802.3az Energy Efficient Ethernet reduces power consumption without compromising high-speed network performance.
- REGIONAL COMPATIBILITY: Made for use in U.S. & CA only
- Data segregation and least-privilege access.
- Identity, secrets, and network controls.
- Prompt, response, model, and system monitoring.
- Testing for accuracy, bias, security, and unexpected behavior.
- Human review for consequential recommendations or decisions.
- Model validation and documented approval processes.
- Retention and audit mechanisms.
- Vendor, concentration, resilience, and regional-deployment controls.
- Incident escalation and recovery procedures.
Bedrock guardrails, identity services, logging, evaluation, and monitoring can support this framework, but no cloud feature by itself satisfies every banking, privacy, model-risk, or consumer-protection obligation. JPMorgan’s internal controls and governance remain decisive.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.The economics: cloud AI is not automatically cheaper
A shared AI platform can lower duplicated engineering work and reduce the cost of individual processes. Potential benefits include faster software delivery, more efficient employee workflows, improved fraud and compliance controls, and lower unit costs in areas such as KYC.
But AI also introduces new expenses. A production system may pay for model tokens, embeddings, retrieval, vector or database storage, GPUs or other accelerators, training, fine-tuning, data transfer, logging, evaluation, monitoring, redundancy, and application integration.
Bedrock pricing varies by model, provider, modality, region, and service tier. AWS lists Standard, Priority, Flex, and Reserved options, with different economics for different workload patterns. Some services are billed by input and output tokens; other capabilities can involve images, model units, knowledge-base queries, or related usage.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →SageMaker pricing is more infrastructure-oriented. Charges can involve compute instances, storage, data processing, training, hosting, and MLOps capabilities. A serious buyer should model costs using actual request volumes, context lengths, latency targets, retention periods, peak demand, and failover requirements rather than relying on a single headline price. AWS provides a pricing calculator for that purpose.
Best Value
- PLUG-AND-PLAY GIGABIT MANAGED SWITCH: 8 x 1Gbps auto-negotiating ports work the moment you plug in — full-gigabit speed over Cat5e/Cat6 cabling.
- MANAGED, WITHOUT THE COMPLEXITY: Easy Smart web GUI on Windows, Mac or Linux — no app or Windows-only utility, unlike many competing switches.
- SEGMENT & PRIORITIZE TRAFFIC: Up to 64 VLANs, QoS, IGMP snooping and port mirroring keep voice, video and data fast, secure and organized.
- BUILT-IN PROTECTION: Auto DoS prevention, loop detection, broadcast storm control and cable test keep your network stable and easy to troubleshoot.
- RELIABLE 24/7 BACKBONE: Rugged fanless metal housing runs cool and silent at 0 dBA — the managed switch trusted in homes, offices and small business.
Why JPMorgan may not put everything on AWS
Multi-cloud architecture can improve bargaining power and reduce dependence on one provider. It can also let JPMorgan match workloads to different clouds, data centers, models, regions, latency requirements, and regulatory constraints.
The trade-off is complexity. Operating across providers can mean duplicated controls, harder observability, inconsistent APIs, more complicated data movement, and additional resilience testing. Moving an existing application may also be more expensive or risky than leaving it in place. Model availability and performance can vary by provider and region.
For JPMorgan, the strategic question is not simply whether AWS is capable. It is where AWS creates the best combination of model choice, security integration, operational control, resilience, cost, and negotiating flexibility.
What this means for enterprise AI
JPMorgan’s strategy points to a pattern likely to matter beyond banking: regulated enterprises are building internal AI platforms instead of sending sensitive work directly to unmanaged public chatbots. The platform typically combines multiple models, shared data services, reusable application components, and centralized governance.
AWS benefits when it becomes part of that foundation, supplying infrastructure and managed services that can support both conventional machine learning and generative-AI applications. JPMorgan benefits when those services help business units experiment within a common control framework and move successful use cases into production.
The important distinction is between cloud provider, AI platform, and business application. AWS supplies part of the technical foundation. JPMorgan owns the operating model, internal controls, and products such as LLM Suite. The resulting architecture is ambitious precisely because it is selective, multi-cloud, and governed—not because the entire bank has moved to AWS.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
