{ "features_resolved": { "snapshot_any": false, "snapshot_mem": false, "fork": false, "pause_resume": false, "persistent_disk": false, "whole_disk": false, "persist_scope": null, "volumes": true, "auto_stop_idle": true, "long_sessions": false, "custom_image": true, "image_snapshot": false, "image_oci": true, "vm_isolation": true, "docker_inside": false, "nested_virt": false, "root": false, "systemd": false, "browser": false, "desktop": false, "computer_use": false, "code_interpreter": false, "browser_control": null, "desktop_control": null, "browser_and_desktop": false, "anti_bot": false, "captcha_solving": false, "residential_ip": false, "agent_harnesses": false, "ssh": false, "public_ipv4": false, "inbound_https": true, "custom_domain": false, "raw_tcp_inbound": false, "egress_allowlist": true, "open_internet": true, "static_egress_ip": false, "private_network": true, "self_host": false, "byoc": false, "open_source": false, "soc2": true, "hipaa": true, "sso": true, "gpu": false, "arm64": true, "windows": false, "macos": false, "lang_python": true, "lang_node": true, "lang_go": null, "lang_rust": null, "lang_java": true, "gpu_desktop": false, "credential_injection": null, "egress_http_rules": false, "wake_on_request": true, "live_resize": false, "fork_running": false, "memory_snapshot_fork": false, "webhooks": null, "scheduled_wakeups": true, "audit_logs": true, "eu_data_residency": true, "zero_data_retention": null, "scoped_api_keys": null, "spend_limits": null, "mcp_server": true, "usage_api": true, "snapshot_auto": false, "snapshot_on_demand": false, "harness_api": false, "own_agent_api": false, "fast_boot": null, "strong_isolation": true, "ingress_rules": true, "guest_firewall": false, "secret_proxy": false, "secret_proxy_any": false, "volume_attach": true, "volume_shared": true, "sidecar": null, "fuse": null, "kernel_features": true, "burst": null }, "company": { "id": "aws-lambda", "checked": "2026-09-30", "company": "Amazon.com, Inc. (AWS Lambda is a service of Amazon Web Services)", "ownership": { "value": "subsidiary", "parent": "Amazon.com, Inc. (via Amazon Web Services, Inc.)", "source": "https://en.wikipedia.org/wiki/Amazon_Web_Services", "quote": "AWS Lambda (a serverless function that can perform arbitrary code written in any language that can be configured to be triggered by hundreds of events, including HTTP calls)" }, "stage": { "value": "public", "source": "https://en.wikipedia.org/wiki/Amazon_Web_Services", "quote": "Amazon.com Announces Fourth Quarter Results", "confidence": "high", "note": "AWS Lambda is a product line inside Amazon Web Services, a segment of Amazon.com, Inc. (NASDAQ: AMZN), a public company. No independent funding history applies." }, "accelerators": [], "total_funding_usd": { "value": null, "status": "unknown" }, "last_round": { "type": "n/a", "amount_usd": null, "date": null, "leads": [], "status": "unknown", "source": "", "quote": "" }, "investors": [], "portfolio_listings": [], "revenue": { "value_usd": 128700000000, "kind": "annual revenue", "as_of": "2025", "status": "confirmed", "basis": "AWS segment net sales for full-year 2025, per Amazon.com's Q4 2025 earnings press release, as cited by Wikipedia's AWS infobox. Figure is for the entire AWS segment (all services, not AWS Lambda specifically); Lambda's own revenue is not separately disclosed.", "source": "https://ir.aboutamazon.com/news-release/news-release-details/2026/Amazon-com-Announces-Fourth-Quarter-Results/", "quote": "US$128.7 billion (2025)" }, "employees": { "value": null, "source": "" }, "note": "AWS Lambda is not an independent company; it is a serverless-compute product within Amazon Web Services, which is a business segment of publicly traded Amazon.com, Inc. Funding/accelerator/ownership fields reflect this: no seed/Series rounds apply, and revenue given is for the whole AWS segment (launched 2006), not Lambda in isolation (Lambda itself launched 2014)." }, "id": "aws-lambda", "name": "AWS Lambda", "url": "https://aws.amazon.com/lambda/pricing/", "category": "hyperscaler", "isolation": "firecracker", "modes": [ { "key": "on-demand", "label": "x86 functions, duration tier 1", "flags": [], "os": [ "linux" ], "regions": [ "us", "eu", "asia" ], "gpu_only": false, "features": {}, "pricing": "resource", "vcpu_h": 0, "ram_gib_h": 0.06, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ 1.769, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 6, "max_ram_gib": 10, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": 0.001, "granularity_s": 0.001, "start_fee": 2e-7, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "Lambda-GB-Second tier 1 $0.0000166667 (0-6B GB-s/month) = $0.06/GB-h; memory-only pricing, CPU is proportional (~1 vCPU per 1,769 MB, 6 vCPU at 10,240 MB). Billed per 1 ms. Requests $0.20/M (start_fee = 1 invocation per session). Same rate in us-east-1, eu-west-1, eu-central-1, eu-north-1, ap-south-1, ap-northeast-1, ap-southeast-1 (Price List API).", "cpu_class": "shared", "plans": [ "Pay-as-you-go" ] }, { "key": "on-demand-arm64", "label": "Arm64 (Graviton) functions, duration tier 1", "flags": [], "os": [ "linux" ], "regions": [ "us", "eu", "asia" ], "gpu_only": false, "features": { "arm64": true }, "pricing": "resource", "vcpu_h": 0, "ram_gib_h": 0.048, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ 1.769, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 6, "max_ram_gib": 10, "sizes": null, "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": 0.001, "granularity_s": 0.001, "start_fee": 2e-7, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "Lambda-GB-Second-ARM tier 1 $0.0000133334 (0-7.5B GB-s) = $0.048/GB-h, 20% below x86.", "cpu_class": "shared", "plans": [ "Pay-as-you-go" ] }, { "key": "savings-plan-1y", "label": "Compute Savings Plan 1y/3y no-upfront, x86", "flags": [ "commit" ], "os": [ "linux" ], "regions": [ "us" ], "gpu_only": false, "features": {}, "pricing": "resource", "vcpu_h": 0, "ram_gib_h": 0.05292, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ 1.769, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 6, "max_ram_gib": 10, "sizes": null, "pool_tiers": null, "requires_always_on": true, "always_on_month_per_instance": null, "commit_note": "Compute SP $/hour commitment; for Lambda the 1y and 3y no-upfront rates are identical (12% off). All-upfront $0.0000138/GB-s = 17% off.", "min_billed_seconds": 0.001, "granularity_s": 0.001, "start_fee": 2e-7, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "AWSComputeSavingsPlan us-east-1: Lambda-GB-Second $0.0000147 (1y and 3y no-upfront).", "cpu_class": "shared", "plans": [ "Pay-as-you-go" ] }, { "key": "savings-plan-1y-arm64", "label": "Compute Savings Plan no-upfront, Arm64", "flags": [ "commit" ], "os": [ "linux" ], "regions": [ "us" ], "gpu_only": false, "features": { "arm64": true }, "pricing": "resource", "vcpu_h": 0, "ram_gib_h": 0.04212, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ 1.769, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 6, "max_ram_gib": 10, "sizes": null, "pool_tiers": null, "requires_always_on": true, "always_on_month_per_instance": null, "commit_note": "As above; all-upfront Arm $0.0000111/GB-s.", "min_billed_seconds": 0.001, "granularity_s": 0.001, "start_fee": 2e-7, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "AWSComputeSavingsPlan us-east-1: Lambda-GB-Second-ARM $0.0000117.", "cpu_class": "shared", "plans": [ "Pay-as-you-go" ] }, { "key": "provisioned-concurrency", "label": "Provisioned concurrency x86 (kept warm, fully busy)", "flags": [], "os": [ "linux" ], "regions": [ "us" ], "gpu_only": false, "features": {}, "pricing": "resource", "vcpu_h": 0, "ram_gib_h": 0.05, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ 1.769, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 6, "max_ram_gib": 10, "sizes": null, "pool_tiers": null, "requires_always_on": true, "always_on_month_per_instance": null, "commit_note": "Provisioned concurrency is billed for every second it is configured (min 5 min), whether or not invocations arrive; free tier does not apply.", "min_billed_seconds": 0.001, "granularity_s": 0.001, "start_fee": 2e-7, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "Lambda-Provisioned-Concurrency $0.0000041667/GB-s ($0.015/GB-h) for the configured window + Lambda-Provisioned-GB-Second $0.0000097222 ($0.035/GB-h) while executing. Rate shown assumes 100% busy; idle provisioned capacity costs $0.015/GB-h.", "cpu_class": "shared", "plans": [ "Pay-as-you-go" ] }, { "key": "provisioned-concurrency-arm64", "label": "Provisioned concurrency Arm64 (fully busy)", "flags": [], "os": [ "linux" ], "regions": [ "us" ], "gpu_only": false, "features": { "arm64": true }, "pricing": "resource", "vcpu_h": 0, "ram_gib_h": 0.04, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ 1.769, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 6, "max_ram_gib": 10, "sizes": null, "pool_tiers": null, "requires_always_on": true, "always_on_month_per_instance": null, "commit_note": "Provisioned concurrency is billed for every second it is configured (min 5 min), whether or not invocations arrive; free tier does not apply.", "min_billed_seconds": 0.001, "granularity_s": 0.001, "start_fee": 2e-7, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "PC-ARM $0.0000033334 + Provisioned-GB-Second-ARM $0.0000077778 per GB-s = $0.012 + $0.028 per GB-h.", "cpu_class": "shared", "plans": [ "Pay-as-you-go" ] }, { "key": "managed-instances", "label": "Lambda Managed Instances (EC2 + 15% fee)", "flags": [ "alt" ], "os": [ "linux" ], "regions": [ "us", "eu", "asia", "other" ], "gpu_only": false, "features": {}, "pricing": "sizes", "vcpu_h": null, "ram_gib_h": null, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": null, "max_ram_gib": null, "sizes": [ { "name": "c7g.xlarge", "vcpu": 4, "ram_gib": 8, "disk_gib": 0, "hour": 0.16675, "month_cap": null }, { "name": "c7i.xlarge", "vcpu": 4, "ram_gib": 8, "disk_gib": 0, "hour": 0.205275, "month_cap": null }, { "name": "m7i.xlarge", "vcpu": 4, "ram_gib": 16, "disk_gib": 0, "hour": 0.23184, "month_cap": null } ], "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": "Always-on floor: publishing a function version launches 3 instances by default (default MinExecutionEnvironments 3 across AZs); an active function cannot scale to zero (min 0 only valid with max 0 = deactivated, and a deactivated function does not scale back up on traffic). Scales on CPU only, doubling traffic within ~5 min. 15% management fee is always on the EC2 on-demand price (not discounted by SPs/RIs).", "min_billed_seconds": 60, "granularity_s": 1, "start_fee": 2e-7, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "EC2 on-demand + Lambda-Managed-Instances management fee (Price List API: c7g.xlarge $0.02175, c7i.xlarge $0.026775, m7i.xlarge $0.03024 per h = 15%) + $0.20/M requests. Instances run in your account; EC2 SPs/RIs apply to the EC2 part only. Isolation is containers on EC2 Nitro instances (capacity provider = security boundary), NOT Firecracker. Multi-concurrent (up to 64 invokes per vCPU). Function size 2 GB/1 vCPU min, up to 32 GB/16 vCPU (since 2026-03-27); capacity provider MaxVCpuCount default 400. Timeout up to 90 min for async/ESM invocations (15 min for sync). 500+ instance types incl. Graviton4/Graviton5.", "cpu_class": "dedicated", "plans": [ "Pay-as-you-go" ] }, { "key": "microvms", "label": "Lambda MicroVMs (Arm64 sandboxes, suspend/resume)", "flags": [], "os": [ "linux" ], "regions": [ "us", "eu", "asia" ], "gpu_only": false, "features": { "arm64": true, "isolation": "firecracker", "max_session_h": 8, "pause_resume": true, "snapshot": "mem", "snapshot_auto": true, "snapshot_on_demand": true, "auto_stop_idle": true, "wake_on_request": true, "ssh": true, "root": true, "docker_inside": null, "nested_virt": null, "systemd": null, "guest_firewall": null, "inbound_https": true, "raw_tcp_inbound": false, "public_ipv4": false, "ingress_rules": true, "egress_allowlist": true, "egress_allowlist_level": "ip", "private_network": true, "open_internet": true, "custom_image": true, "image_oci": true, "volumes": null, "volume_attach": null, "volume_shared": null, "code_interpreter": false }, "pricing": "sizes", "vcpu_h": 0.09969984, "ram_gib_h": 0.01320012, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ 2, 2 ], "ram_fixed_per_vcpu": 2, "vcpu_options": [ 0.25, 0.5, 1, 2, 4 ], "min_vcpu": 0.25, "max_vcpu": 4, "max_ram_gib": 8, "sizes": [ { "name": "0.5 GB baseline", "vcpu": 0.25, "ram_gib": 0.5, "disk_gib": 8, "hour": 0.03152502, "month_cap": null }, { "name": "1 GB baseline", "vcpu": 0.5, "ram_gib": 1, "disk_gib": 8, "hour": 0.06305004, "month_cap": null }, { "name": "2 GB baseline (default)", "vcpu": 1, "ram_gib": 2, "disk_gib": 8, "hour": 0.12610008, "month_cap": null }, { "name": "4 GB baseline", "vcpu": 2, "ram_gib": 4, "disk_gib": 16, "hour": 0.25220016, "month_cap": null }, { "name": "8 GB baseline", "vcpu": 4, "ram_gib": 8, "disk_gib": 32, "hour": 0.50440032, "month_cap": null } ], "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": 1, "granularity_s": 1, "start_fee": 0.0031, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "Graviton (Arm64) only. us-east-1 $0.0000276944/vCPU-s + $0.0000036667/GB-s on the configured baseline while running ($0.0997/vCPU-h + $0.0132/GB-h; 1 vCPU / 2 GB = $0.126/h, 4 vCPU / 8 GB = $0.504/h). Memory sets the baseline, vCPU fixed at 2 GB per vCPU (0.5 GB / 0.25 vCPU up to 8 GB / 4 vCPU); each size bursts automatically to 4x baseline (up to 32 GB / 16 vCPU), and only the active seconds above baseline are billed extra (not modelled). Max disk 8 GB (sizes up to 2 GB), 16 GB (4 GB), 32 GB (8 GB). Max lifespan 8 h (28,800 s) including suspended time. Suspended = no compute, only snapshot storage $0.08/GB-month; each suspend writes the state ($0.0038/GB) and each resume or launch reads it ($0.00155/GB). start_fee $0.0031 = one launch reading a 2 GB image (AWS pricing example image size). Prices in other regions differ (Price List publication 2026-10-12: eu-west-1 $0.0000291572/vCPU-s, eu-central-1 $0.000034996, ap-northeast-1 $0.0000322421, ap-southeast-1 $0.00003467).", "cpu_class": "shared" } ], "storage": { "included_disk_gib": 0.5, "disk_gib_month": 0.0812, "disk_billed_when_stopped": false, "snapshot_gib_month": 0.08, "paused_ram_billed": false }, "network": { "egress_free_gib": 100, "egress_gib": 0.09, "ipv4_month": null, "ipv4_available": false, "internet": { "default": "full" } }, "plans": [ { "name": "Pay-as-you-go", "fee": 0, "flags": [], "fee_is_credit": false, "included_usd": 0, "per_seat": 0, "concurrency": 1000, "max_session_h": 0.25, "max_vcpu": 6, "max_ram_gib": 10, "sizes_allowed": null, "note": "15-minute max invocation (900 s; LMI async/ESM 90 min). Default regional concurrency quota 1,000 (adjustable to 'tens of thousands'), BUT new accounts get reduced concurrency and memory quotas (Functions and MicroVMs; Web Functions get reduced vCPU and request-rate quotas), raised automatically with usage. Scaling rate: 1,000 execution environments per 10 s per function (not accrued). Function URLs give a free HTTPS endpoint. Network bandwidth 625 Mbps per environment (up to 3,000 Mbps on request outside a VPC)." }, { "name": "MicroVMs pay-as-you-go", "fee": 0, "flags": [], "fee_is_credit": false, "included_usd": 0, "per_seat": 0, "concurrency": null, "max_session_h": 8, "max_vcpu": 4, "max_ram_gib": 8, "sizes_allowed": null, "max_starts_per_min": 300, "note": "Same AWS account, no fee; only the microvms mode is sold on it. Max lifespan 8 h per MicroVM (not adjustable). No machine-count quota: memory allocated across all MicroVMs is 400 GB per account per Region by default (200 MicroVMs at 2 GB), 1,024 GB in us-east-1, us-west-2, us-east-2 and ap-northeast-1, burstable up to 4x, adjustable; new accounts get less. RunMicrovm 5 TPS (300 starts/min), ResumeMicrovm 5 TPS, SuspendMicrovm 2 TPS, all adjustable. Concurrent connections per MicroVM 8 (1 vCPU) to 128 (16 vCPU); endpoint bandwidth 1 MB/s (0.5 GB) to 16 MB/s (8 GB)." } ], "free": { "monthly_credit": 6.87, "one_time_credit": 200 }, "features": { "linux": true, "windows": false, "macos": false, "arm64": true, "gpu": false, "gpu_types": [], "snapshot": "none", "pause_resume": false, "fork": false, "persistent_disk": false, "volumes": true, "auto_stop_idle": true, "max_session_h": 0.25, "custom_image": true, "docker_inside": false, "nested_virt": false, "root": false, "systemd": false, "browser": false, "desktop": false, "computer_use": false, "code_interpreter": false, "agent_harnesses": false, "ssh": false, "public_ipv4": false, "inbound_https": true, "custom_domain": false, "raw_tcp_inbound": false, "egress_allowlist": true, "static_egress_ip": false, "private_network": true, "self_host": false, "byoc": false, "open_source": false, "regions": [ "us", "eu", "asia", "other" ], "soc2": true, "hipaa": true, "sso": true, "lang_python": true, "lang_node": true, "lang_go": null, "lang_rust": null, "lang_java": true, "gpu_desktop": false, "credential_injection": null, "egress_http_rules": false, "wake_on_request": true, "live_resize": false, "fork_running": false, "memory_snapshot_fork": false, "webhooks": null, "scheduled_wakeups": true, "audit_logs": true, "eu_data_residency": true, "zero_data_retention": null, "scoped_api_keys": null, "spend_limits": null, "mcp_server": true, "snapshot_auto": false, "snapshot_on_demand": false, "harness_api": false, "egress_allowlist_level": "ip", "ingress_rules": true, "guest_firewall": false, "secret_proxy": false, "volume_attach": true, "volume_shared": true, "isolation": "firecracker", "image_oci": true, "anti_bot": false, "captcha_solving": false, "egress_ip_type": "datacenter", "own_agent_api": false, "open_internet": true, "usage_api": true }, "perf": { "cpu_runs_s": null, "cold_start_ms": null, "burst_tti_ms": null, "dax_s": null }, "knobs": [], "caveats": [ "Rates from the AWS Price List API (AWSLambda, publication 2026-09-19) and AWSComputeSavingsPlan (2026-09-25).", "Volume tiers (automatic, per account per region): x86 $0.0000166667 (0-6B GB-s), $0.000015 (6-15B), $0.0000133334 (>15B); Arm $0.0000133334 (0-7.5B), $0.0000120001 (7.5-18.75B), $0.0000106667 (>18.75B). 6B GB-s = 1.67M GB-hours; not modelled. The previous provider record had tier 2 wrong ($0.0000133334).", "Regional variance: the major us/eu/asia regions all charge tier-1 $0.0000166667; af-south-1 $0.0000221, ap-east-1 $0.00002292, eu-south-1 $0.0000195172, me-south-1 $0.0000206667; requests $0.20/M in most regions, up to $0.28/M (ap-east-1).", "free.monthly_credit 6.87 = always-free 400,000 GB-s x $0.0000166667 ($6.67) + 1M requests x $0.20/M ($0.20), derived; it applies to x86+Arm combined and not to provisioned concurrency.", "Memory-only pricing: vcpu_h is 0 and ram_per_vcpu [1.769, null] makes the engine buy ~1.77 GB per requested vCPU (4 vCPU needs >= 7.1 GB). 1,769 MB per vCPU is AWS docs knowledge, not on the pricing page.", "Lambda Functions are not a sandbox: 15-minute cap, stateless, no inbound shell, /tmp 512 MB free then $0.0000000309/GB-s ($0.0812/GB-month-equivalent while executing), max 10,240 MB. The sandbox product is the separate microvms mode (Lambda MicroVMs).", "Other per-request fees: $0.20/M requests; streaming responses $0.008/GB beyond 100 GiB free; SnapStart cache $0.0000015046/GB-s + restore $0.0001397998/GB; durable functions $0.000008/operation + $0.25/GB written + $0.15/GB-month; tenant-isolated environments $0.000166667/GB created.", "Lambda in a VPC needs a NAT gateway for internet access: NAT gateway gotcha: tasks/functions in private subnets reach the internet through a NAT gateway at $0.045/h per gateway (~$32.85/month per AZ) plus $0.045/GB processed (us-east-1; eu-north-1 $0.046, ap-south-1 $0.056), charged ON TOP of the $0.09/GB egress. Not modelled; public IPv4 ($0.005/h) is the usual alternative.", "Internet egress (AWS Price List API, AWSDataTransfer us-east-1, 2026-09): first 100 GB/month free aggregated across all services and regions, then $0.09/GB first 10 TB, $0.085 next 40 TB, $0.07 next 100 TB, $0.05 over 150 TB. eu-west-1 is also $0.09; ap-south-1 $0.1093; ap-northeast-1 $0.114; ap-southeast-1 $0.12 (first-10-TB tier). The engine uses the flat us-east-1 first tier.", "Regional on-demand variance (Price List 2026-09-19, x86 tier 1 per GB-s): CHEAPER than us-east-1 in ap-east-2 Taipei, ap-southeast-5 Malaysia, ap-southeast-7 Thailand ($0.000015, requests $0.18/M, -10%); dearer in eu-central-2 Zurich $0.0000183334, il-central-1 / mx-central-1 / ap-southeast-6 $0.0000175, me-central-1 & me-south-1 $0.0000206667, eu-south-1 $0.0000195172, af-south-1 $0.0000221, ap-east-1 $0.00002292. Provisioned concurrency and ephemeral storage vary even in regions where on-demand is flat (eu-west-1 storage $0.000000034 vs $0.0000000309).", "Tenant isolation mode (per-tenant execution environments, the Lambda feature closest to 'one sandbox per user') adds a per-environment-creation fee of $0.000166667 per GB of configured memory (x86; Arm $0.000133333): an 8 GB tenant environment costs $0.00133 per cold start, on top of $0.20/M requests. Not modelled in start_fee.", "Durable functions let one logical execution span up to 1 year via checkpoint + replay; waits incur no duration charge on on-demand functions, but each sub-invocation still has the 15-min cap and replay re-bills handler time. Extra: $0.000008/durable operation, $0.25/GB written, $0.15/GB-month retained (1-90 days after completion, default 14).", "Lambda MicroVMs (microvms mode, checked 2026-10-14): purpose-built for \"sandboxes for AI\", interactive dev environments and multi-tenant CI/CD. Firecracker microVM running your Dockerfile as a container on an Amazon Linux 2023 base; Arm64 (Graviton) only. Sold on its own plan in the card (\"MicroVMs pay-as-you-go\") so the 15-minute function cap does not apply to it; the card-level features describe Lambda Functions, the microvms mode overrides pause_resume, snapshot, ssh, root and max_session_h.", "MicroVMs billing not modelled: (1) peak burst up to 4x baseline is billed only for the active seconds above baseline at the same per-vCPU-s / per-GB-s rates, so a size is priced at its baseline and a workload asking for more than 4 vCPU / 8 GB gets no MicroVM row; (2) each suspend writes the state at $0.0038/GB and each resume reads it at $0.00155/GB (2 GB state, 6 cycles a day = $0.065/day); (3) suspended time bills snapshot storage at $0.08/GB-month (storage.snapshot_gib_month, which the engine uses only for retained state); (4) MicroVM images bill $0.08/GB-month with a 1-week minimum retention; (5) launch reads cost image GB x $0.00155 (start_fee assumes 2 GB).", "MicroVMs disk is included per size and capped (8 GB for baselines up to 2 GB, 16 GB at 4 GB, 32 GB at 8 GB); storage.disk_gib_month ($0.0812, the function /tmp rate) does not apply to MicroVMs, and no larger disk can be bought.", "MicroVMs networking: inbound only through a per-MicroVM public HTTPS endpoint (HTTP/1.1, HTTP/2, WebSockets, gRPC, SSE; port chosen with X-aws-proxy-port), every request needs a port-scoped JWE token, no raw TCP/UDP inbound and no public IPv4; egress is public internet by default or a customer VPC through a VPC egress connector (security groups / NACLs). Endpoint bandwidth 1 MB/s (0.5 GB) to 16 MB/s (8 GB). Shell access is a separate token API on MicroVMs started with the SHELL_INGRESS connector, not sshd on port 22.", "MicroVMs regions (AWS Price List, publication 2026-10-12): us-east-1, us-east-2, us-west-2, eu-west-1, eu-central-1, eu-north-1, ap-northeast-1, ap-southeast-1, ap-southeast-2, ap-south-1. us-east-1/us-east-2/us-west-2 share the rates in the mode; eu-central-1 is +26% per vCPU-s, ap-southeast-1 +25%.", "The Lambda always-free tier (400,000 GB-s + 1M requests a month, free.monthly_credit) covers Lambda Functions only; the engine applies free credits to whichever mode wins, so a MicroVMs row is understated by up to $6.87/month when credits are on.", "Lambda Web Functions (new, checked 2026-10-14): serverless HTTP servers priced per request plus active vCPU time plus memory for as long as the environment runs, billed per 1 ms; 2 vCPU / 2 GB per execution environment, 512 MB /tmp, request timeout 30 s default (3-900 s), up to 64 concurrent requests per environment (1-128), Arm64 vCPU quota 2,000 per account per Region. An app platform, not a sandbox; not modelled." ], "sources": [ "https://aws.amazon.com/lambda/pricing/", "https://pricing.us-east-1.amazonaws.com/offers/v1.0/aws/index.json", "https://pricing.us-east-1.amazonaws.com/offers/v1.0/aws/AWSLambda/current/region_index.json", "https://pricing.us-east-1.amazonaws.com/savingsPlan/v1.0/aws/AWSComputeSavingsPlan/current/region_index.json", "https://docs.aws.amazon.com/lambda/latest/dg/lambda-microvms-guide.html", "https://docs.aws.amazon.com/lambda/latest/dg/microvms-launching.html", "https://docs.aws.amazon.com/lambda/latest/dg/microvms-images.html", "https://docs.aws.amazon.com/lambda/latest/dg/microvms-networking.html", "https://docs.aws.amazon.com/lambda/latest/dg/microvms-security.html", "https://docs.aws.amazon.com/lambda/latest/dg/gettingstarted-limits.html" ], "fleet": { "default_max_instances": null, "default_max_vcpu": null, "default_max_gpus": null, "raise": "automatic", "raise_time_days": null, "verification": "Personal details, email and phone; payment card may be required.", "tiers": "Reduced new-account profile grows with usage; concurrency increases can also be requested.", "notes": "Do not treat published 1,000 regional concurrent executions as a new-account entitlement. AWS explicitly assigns reduced concurrency/memory to new accounts without one numeric initial profile. Invocation concurrency is not a persistent VM fleet. The reduced new-account profile also covers Lambda MicroVMs (memory quota) and Lambda Web Functions (vCPU and request-rate quotas). MicroVMs default to 400 GB of configured memory across all MicroVMs per Region (1,024 GB in us-east-1, us-west-2, us-east-2, ap-northeast-1), adjustable through Service Quotas.", "sources": [ "https://docs.aws.amazon.com/lambda/latest/dg/gettingstarted-limits.html", "https://aws.amazon.com/free/complete-signup/" ], "researched": "2026-10-14" }, "audit_2026_09_29": { "applied": 15, "source": "checked against the provider's live pricing and docs pages", "reviewed": 2 }, "feature_evidence": { "snapshot_auto": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtime-environment.html", "quote": "Lambda freezes the execution environment when the runtime and each extension have completed and there are no pending events.", "note": "Idle environments are only frozen in place and later shut down (state lost; /tmp may survive only on warm reuse). SnapStart snapshots are taken at version publish, not on stop/idle. Durable functions checkpoint workflow steps and recover by replay (application-level, not machine state). Lambda MicroVMs (microvms mode, checked 2026-10-14) differ; see the mode override.", "checked": "2026-09-29", "mode_microvms": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/microvms-launching.html", "quote": "If no traffic arrives for the configured idle duration, the MicroVM is treated as idle and suspended.", "note": "Idle policy (opt-in per run-microvm); memory and disk preserved.", "checked": "2026-10-14" } }, "snapshot_on_demand": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/snapstart.html", "quote": "With SnapStart, Lambda initializes your function when you publish a function version. Lambda takes a Firecracker MicroVM snapshot of the memory and disk state of the initialized execution environment", "note": "SnapStart is the closest thing: memory+disk Firecracker snapshot of the post-init state, taken when you publish a version (PublishVersion API), and every new execution environment resumes from it. But you cannot checkpoint an arbitrary running environment and resume/fork it; only Java 11+, Python 3.12+, .NET 8+; not with EFS, S3 Files, provisioned concurrency or >512 MB /tmp. Lambda MicroVMs (microvms mode, checked 2026-10-14) differ; see the mode override.", "checked": "2026-09-29", "mode_microvms": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/microvms-launching.html", "quote": "API call (explicit) – Call suspend-microvm to suspend immediately", "note": "Checkpoints memory + disk of that MicroVM for a later resume of the same MicroVM only; it cannot be branched into new MicroVMs.", "checked": "2026-10-14" } }, "harness_api": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/agent-setup-guide.html", "quote": "By configuring your agent with the right combination of plugins, skills, and MCP Servers, you equip it with expertise required to build and manage serverless applications.", "note": "Lambda's only agent material is for local coding agents (Claude Code, Codex, Kiro, OpenCode...) building Lambda apps via MCP/skills; no API to run a harness inside a function. AWS's agent-runtime product is Bedrock AgentCore (separate product, not on this card).", "checked": "2026-09-29" }, "egress_allowlist": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/configuration-vpc.html", "quote": "Security groups control inbound and outbound traffic for your resources, and network ACLs control inbound and outbound traffic for your subnets.", "note": "Only for VPC-attached functions (non-VPC functions get unrestricted internet). A VPC function in a private subnet with no NAT has no internet at all. Domain-level filtering is possible only via separate VPC services (AWS Network Firewall / Route 53 Resolver DNS Firewall), not Lambda itself.", "checked": "2026-09-29", "level": "ip" }, "ingress_rules": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/urls-auth.html", "quote": "AWS_IAM – Lambda uses AWS Identity and Access Management (IAM) to authenticate and authorize requests based on the IAM principal's identity policy and the function's resource-based policy.", "note": "No inbound ports at all; the only ingress is invocation (Function URL with AWS_IAM or NONE auth, API Gateway, resource-based policies).", "checked": "2026-09-29" }, "guest_firewall": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/images-create.html", "quote": "Lambda follows security best practices by defining a default Linux user with least-privileged permissions.", "note": "Firecracker microVM, but code runs as an unprivileged user with no root / CAP_NET_ADMIN, so iptables/nftables cannot be used.", "checked": "2026-09-29", "derived": true }, "secret_proxy": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/with-secrets-manager.html", "quote": "AWS parameters and secrets Lambda extension - A runtime-agnostic solution that provides a simple HTTP interface for retrieving secrets", "note": "Secrets are env vars (KMS-encrypted) or fetched into the function via Secrets Manager/Parameter Store extension; the real secret enters the environment.", "checked": "2026-09-29" }, "volume_attach": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/configuration-filesystem.html", "quote": "You can configure a Lambda function to mount a file system to a local directory. Lambda supports the following file system types: Amazon Elastic File System (Amazon EFS) – Serverless file system that scales automatically with your workloads. Amazon S3 Files – Serverless file system for mounting your Amazon S3 bucket.", "note": "Network file systems only (no block volumes); function must be VPC-attached; EFS or S3 Files, not both. Ephemeral /tmp configurable 512 MB-10 GB.", "checked": "2026-09-29" }, "volume_shared": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/configuration-filesystem-efs.html", "quote": "With Amazon EFS, your function code can access and modify shared resources safely and at high concurrency.", "note": "Read-write (or read-only connections without ClientWrite); NFS on port 2049; shared across all concurrent environments and other compute.", "checked": "2026-09-29" }, "isolation": { "value": "firecracker", "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtime-environment.html", "quote": "AWS Lambda uses Firecracker microVM technology to provide a highly secure, isolated execution environment for each function invocation.", "note": "Confirmed consistent with card top-level isolation field; matches SnapStart docs' explicit 'Firecracker MicroVM snapshot' language too.", "checked": "2026-09-30" }, "image_oci": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/images-create.html", "quote": "Lambda supports two types of deployment packages: container images and .zip file archives. There are three ways to build a container image for a Lambda function: Using an AWS base image for Lambda... Using an AWS OS-only base image... Using a non-AWS base image.", "note": "Container images are pushed to Amazon ECR as standard OCI/Docker images (up to 10 GB) and Lambda runs the function from that image.", "checked": "2026-09-30" }, "anti_bot": { "value": false, "source": "https://aws.amazon.com/lambda/pricing/", "quote": "AWS Lambda is a serverless, event-driven compute service that lets you run code for virtually any type of application or backend service without provisioning or managing servers.", "note": "not offered: outside the product's scope", "checked": "2026-09-30" }, "captcha_solving": { "value": false, "source": "https://aws.amazon.com/lambda/pricing/", "quote": "AWS Lambda is a serverless, event-driven compute service that lets you run code for virtually any type of application or backend service without provisioning or managing servers.", "note": "not offered: outside the product's scope", "checked": "2026-09-30" }, "egress_ip_type": { "value": "datacenter", "source": "https://docs.aws.amazon.com/lambda/latest/dg/configuration-vpc.html", "quote": "Security groups control inbound and outbound traffic for your resources, and network ACLs control inbound and outbound traffic for your subnets.", "note": "Non-VPC functions egress via AWS-managed shared IP ranges; VPC-attached functions egress through a NAT gateway (Elastic IP) or public IPv4 on ENI. Both are standard AWS datacenter IP ranges, not residential/mobile.", "checked": "2026-09-30" }, "wake_on_request": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtime-environment.html", "quote": "Lambda freezes the execution environment when the runtime and each extension have completed and there are no pending events.", "note": "Frozen/idle execution environments (and cold environments with none warm) are transparently created/thawed the moment an invocation (request) arrives; this is the core Lambda invoke-on-demand model. Lambda MicroVMs: \"When true , the MicroVM automatically resumes when traffic arrives at its endpoint while suspended.\" (autoResumeEnabled, microvms-launching.html, checked 2026-10-14).", "checked": "2026-09-30" }, "live_resize": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/configuration-function-common.html", "quote": "You can configure a range of settings for your function, including memory and timeout allocations.", "note": "Memory/CPU allocation is changed via UpdateFunctionConfiguration, which applies to new execution environments going forward; a currently running/warm environment cannot have its CPU/RAM changed in place.", "checked": "2026-09-30" }, "own_agent_api": { "value": false, "source": "https://aws.amazon.com/lambda/pricing/", "quote": "AWS Lambda is a serverless, event-driven compute service that lets you run code for virtually any type of application or backend service without provisioning or managing servers.", "note": "not offered: outside the product's scope. AWS's own browser/computer-use/research task agent product is Bedrock AgentCore, a separate product not modeled by this card.", "checked": "2026-09-30" }, "mcp_server": { "value": true, "source": "https://raw.githubusercontent.com/awslabs/mcp/main/src/aws-serverless-mcp-server/README.md", "quote": "Comprehensive tooling: Offers tools for initialization, deployment, monitoring, and troubleshooting of serverless applications... Initialize, build, and deploy Serverless Application Model (SAM) applications with SAM CLI. Test Lambda functions locally and remotely.", "note": "AWS ships the open-source AWS Serverless MCP Server (awslabs.aws-serverless-mcp-server), referenced from Lambda's own Agent setup guide, which deploys, tests/invokes, and monitors Lambda functions via MCP tools.", "checked": "2026-09-30" }, "egress_http_rules": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/configuration-vpc.html", "quote": "Security groups control inbound and outbound traffic for your resources, and network ACLs control inbound and outbound traffic for your subnets.", "note": "Egress control is IP/CIDR/port-level (security groups, NACLs) or domain-level via a separate service (Network Firewall / Route 53 Resolver DNS Firewall); nothing at HTTP method/path granularity.", "checked": "2026-09-30" }, "eu_data_residency": { "value": true, "source": "https://aws.amazon.com/compliance/gdpr-center/", "quote": "Choose the secured state of their customer data, including where it will be stored, including the type of storage and geographic region of that storage.", "note": "Lambda is generally available in multiple EU regions (eu-west-1 Ireland, eu-central-1 Frankfurt, eu-central-2 Zurich, eu-north-1 Stockholm, eu-south-1 Milan, eu-west-2/3 London/Paris); customers select the region and AWS does not move data out of it by default.", "checked": "2026-09-30" }, "persistent_disk": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/configuration-ephemeral-storage.html", "quote": "Lambda provides ephemeral storage for functions in the /tmp directory. This storage is temporary and unique to each execution environment.", "note": "No contradiction: 'volumes':true refers only to opt-in EFS/S3-Files network mounts a customer must attach explicitly; the default /tmp disk is ephemeral per execution environment and Lambda has no user-triggered pause/resume that preserves it (freeze/thaw is automatic and the environment can be discarded/reset at any time, e.g. 'the Lambda reset does not clear the /tmp directory content before the next init phase' is best-effort reuse, not a guarantee). Lambda MicroVMs keep their disk across suspend/resume, but a MicroVM lives at most 8 h (running + suspended), so files are not kept beyond that; persistent_disk stays false in the microvms mode.", "checked": "2026-09-30" }, "snapshot": { "value": "none", "source": "https://docs.aws.amazon.com/lambda/latest/dg/snapstart.html", "quote": "With SnapStart, Lambda initializes your function when you publish a function version. Lambda takes a Firecracker MicroVM snapshot of the memory and disk state of the initialized execution environment", "note": "SnapStart only snapshots the post-init state at version-publish time for cold-start acceleration (Java/Python/.NET only, restricted); there is no API to snapshot an arbitrary running execution environment on demand or automatically on idle/stop, so 'none' is correct for the general-purpose snapshot feature this field tracks. Lambda MicroVMs (microvms mode, checked 2026-10-14) differ; see the mode override.", "checked": "2026-09-30", "mode_microvms": { "value": "mem", "source": "https://docs.aws.amazon.com/lambda/latest/dg/microvms-launching.html", "quote": "When a MicroVM resumes (through an API call or auto-resume), Lambda restores memory and disk state from the suspend checkpoint.", "note": "Memory + disk checkpoint of the same MicroVM on suspend; MicroVM images are snapshots of the initialized app that many MicroVMs launch from. No fork of a running MicroVM is documented.", "checked": "2026-10-14" } }, "pause_resume": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtime-environment.html", "quote": "Lambda freezes the execution environment when the runtime and each extension have completed and there are no pending events.", "note": "This freeze/thaw is fully automatic reuse-or-discard behavior managed by the service between invocations, not a user-invokable pause/resume API; Lambda gives no way to explicitly pause a running environment and resume it later. Not true of Lambda MicroVMs: the microvms mode sets pause_resume true.", "checked": "2026-09-30", "mode_microvms": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-microvms-guide.html", "quote": "MicroVMs can be suspended when idle, preserving memory and disk state while reducing costs. When traffic arrives, they are resumed.", "note": "suspend-microvm / resume-microvm APIs or an idle policy (maxIdleDurationSeconds up to 8 h, autoResumeEnabled). Suspended time still counts toward the 8 h lifespan.", "checked": "2026-10-14" } }, "fork": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtime-environment.html", "quote": "Lambda invokes your function in an execution environment... If Lambda needs to scale up, it creates new execution environments.", "note": "New execution environments are always created fresh (or resumed from a SnapStart version snapshot); there is no API to fork a live running environment into a second independent copy. No AWS documentation describes a fork capability.", "checked": "2026-09-30" }, "fork_running": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtime-environment.html", "quote": "Lambda invokes your function in an execution environment, which provides a secure and isolated runtime environment.", "note": "Same basis as fork: no documented mechanism to fork a currently-running execution environment; each environment handles invocations serially and new ones are created independently, not branched from a live one.", "checked": "2026-09-30" }, "memory_snapshot_fork": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/snapstart.html", "quote": "With SnapStart, Lambda initializes your function when you publish a function version. Lambda takes a Firecracker MicroVM snapshot of the memory and disk state of the initialized execution environment", "note": "SnapStart resumes many new environments from one shared version-publish-time snapshot, but this is not a user-controlled 'fork a running instance's memory' primitive: you cannot pick an arbitrary in-flight environment and branch it.", "checked": "2026-09-30" }, "docker_inside": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/images-create.html", "quote": "The container image must be able to run on a read-only file system... The default Lambda user must be able to read all the files required to run your function code. Lambda follows security best practices by defining a default Linux user with least-privileged permissions.", "note": "The container image is only Lambda's deployment packaging format; the running function executes as an unprivileged, non-root user on a read-only filesystem, so it cannot start a Docker daemon or run nested containers inside the execution environment.", "checked": "2026-09-30" }, "nested_virt": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/images-create.html", "quote": "Lambda follows security best practices by defining a default Linux user with least-privileged permissions. This means that you don't need to specify a USER in your Dockerfile.", "note": "Code runs unprivileged inside a Firecracker microVM with no elevated capabilities (no CAP_SYS_ADMIN/KVM access), so customer code cannot start its own nested VM/hypervisor.", "checked": "2026-09-30" }, "root": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/images-create.html", "quote": "Lambda follows security best practices by defining a default Linux user with least-privileged permissions. This means that you don't need to specify a USER in your Dockerfile.", "note": "Function code always runs as a non-root, least-privileged default Linux user; there is no way to obtain root inside the execution environment. Lambda MicroVMs (microvms mode, checked 2026-10-14) differ; see the mode override.", "checked": "2026-09-30", "mode_microvms": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/microvms-images.html", "quote": "You can grant elevated Linux capabilities using the additionalOsCapabilities field when creating or updating a MicroVM image. The only supported value is [\"ALL\"] . Elevated capabilities enable operations such as mounting filesystems, creating network namespaces, or running eBPF programs.", "note": "Opt-in per image; by default MicroVMs run \"with a standard set of Linux capabilities\". The app runs as a container inside the MicroVM OS. Docker-in-MicroVM, systemd and nested virtualization are not documented (left unknown in the mode).", "checked": "2026-10-14" } }, "ssh": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtime-environment.html", "quote": "Lambda invokes your function in an execution environment, which provides a secure and isolated runtime environment that manages the resources required to run your function.", "note": "Lambda has no shell/SSH access product surface at all -- interaction is strictly through the Invoke API and Runtime/Extensions/Telemetry APIs; there is no listed inbound SSH/console feature anywhere in the docs, matching the card's separate ingress_rules note that the only ingress is invocation. Lambda MicroVMs (microvms mode, checked 2026-10-14) differ; see the mode override.", "checked": "2026-09-30", "mode_microvms": { "value": true, "source": "https://docs.aws.amazon.com/lambda/latest/dg/microvms-security.html", "quote": "Shell access uses a separate token API. The target MicroVM must have been run with the SHELL_INGRESS network connector", "note": "Interactive shell through create-microvm-shell-auth-token (5 TPS), not sshd on a public port.", "checked": "2026-10-14" } }, "desktop": { "value": false, "source": "https://aws.amazon.com/lambda/", "quote": "AWS Lambda is a serverless compute service that lets you run code without provisioning or managing servers.", "note": "Lambda is a stateless function-execution service with no display/GUI surface of any kind; a remote desktop feature is outside the product's scope.", "checked": "2026-09-30" }, "browser": { "value": false, "source": "https://aws.amazon.com/lambda/", "quote": "AWS Lambda is a serverless compute service that lets you run code without provisioning or managing servers.", "note": "No managed/embedded browser is offered by Lambda; it is a headless function runtime (customers may run headless Chromium themselves inside a function, but that is not a first-party browser feature).", "checked": "2026-09-30" }, "computer_use": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/agent-setup-guide.html", "quote": "By configuring your agent with the right combination of plugins, skills, and MCP Servers, you equip it with expertise required to build and manage serverless applications.", "note": "Lambda's agent material targets local coding agents that build/manage Lambda apps via MCP/skills; it provides no computer-use (screen/mouse/keyboard control) capability. AWS's actual computer-use product is Bedrock AgentCore, a separate service not represented by this card.", "checked": "2026-09-30" }, "public_ipv4": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/configuration-vpc.html", "quote": "Connecting a function to a public subnet doesn't give it internet access or a public IP address.", "note": "Functions run inside an AWS-managed Lambda VPC that customers cannot see; even VPC-attached functions in a public subnet get no public IPv4. There is no way to assign a public IPv4 address to a Lambda function.", "checked": "2026-09-30" }, "static_egress_ip": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/configuration-vpc.html", "quote": "By default, Lambda functions have access to the public internet. When you attach your function to a VPC, it can only access resources available within that VPC. To give your function access to the internet, you also need to configure the VPC to have internet access.", "note": "Lambda itself assigns no fixed egress IP; a static egress IP is only achievable by the customer separately provisioning a VPC NAT Gateway + Elastic IP (per the card's own NAT gateway caveat) -- not a built-in Lambda feature.", "checked": "2026-09-30" }, "agent_harnesses": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/agent-setup-guide.html", "quote": "By configuring your agent with the right combination of plugins, skills, and MCP Servers, you equip it with expertise required to build and manage serverless applications.", "note": "Same evidence already recorded under feature_evidence.harness_api: this guide is for external coding agents (Claude Code, Codex, Kiro...) that build Lambda apps; there is no hosted agent-harness runtime feature inside Lambda itself.", "checked": "2026-09-30" }, "gpu": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/gettingstarted-limits.html", "quote": "Function memory allocation 128 MB to 10,240 MB, in 1-MB increments. Note: Lambda allocates CPU power in proportion to the amount of memory configured.", "note": "Lambda's compute/quota documentation only ever describes memory-proportional vCPU allocation; there is no GPU option, GPU quota, or GPU pricing dimension anywhere in Lambda's docs or pricing pages.", "checked": "2026-09-30" }, "windows": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtimes.html", "quote": "The runtime is paired with one of the Amazon Linux distributions... Lambda supports only Linux-based container images.", "note": "All Lambda runtimes and container base images run on Amazon Linux (2 or 2023); there is no Windows runtime or Windows container support.", "checked": "2026-09-30" }, "macos": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtimes.html", "quote": "The runtime is paired with one of the Amazon Linux distributions. The underlying execution environment provides additional libraries and environment variables that you can access from your function code.", "note": "Lambda functions only run on Amazon Linux; there is no macOS runtime (AWS's macOS offering is the unrelated EC2 Mac dedicated-host product, not Lambda).", "checked": "2026-09-30" }, "self_host": { "value": false, "source": "https://aws.amazon.com/lambda/", "quote": "AWS Lambda is a serverless compute service that lets you run code without provisioning or managing servers.", "note": "Lambda is a fully managed, AWS-operated multi-tenant service; there is no distributable Lambda runtime/control-plane a customer can install and operate on their own infrastructure.", "checked": "2026-09-30" }, "byoc": { "value": false, "source": "https://aws.amazon.com/lambda/", "quote": "AWS Lambda is a serverless compute service that lets you run code without provisioning or managing servers.", "note": "Lambda only runs inside AWS's own account/region infrastructure (execution environments live in an AWS-managed VPC per configuration-vpc.html); there is no bring-your-own-cloud deployment mode. (Lambda Managed Instances still runs on the customer's own EC2 instances within their AWS account, but that is BYO-EC2-within-AWS, not BYOC to another cloud/on-prem.)", "checked": "2026-09-30" }, "open_source": { "value": false, "source": "https://docs.aws.amazon.com/lambda/latest/dg/images-create.html", "quote": "You must include a runtime interface client for your language in the image... open-source runtime interface clients implement the API.", "note": "Only the small Runtime Interface Client/Emulator glue libraries are open source; the Lambda execution service itself (scheduler, Firecracker fleet management, APIs) is closed-source, proprietary AWS infrastructure -- there is no way to run 'Lambda' outside AWS.", "checked": "2026-09-30" }, "cpu_class": { "value": { "on-demand": "shared", "on-demand-arm64": "shared", "savings-plan-1y": "shared", "savings-plan-1y-arm64": "shared", "provisioned-concurrency": "shared", "provisioned-concurrency-arm64": "shared", "managed-instances": "dedicated", "microvms": "shared" }, "source": "https://www.usenix.org/system/files/nsdi20-paper-agache.pdf", "quote": "It must be possible to over commit CPU, memory and other resources, with each function consuming only the resources it needs, not the resources it is entitled to... In Lambda, CPU, network, and storage throughput is allocated to functions proportionally to their configured memory limit... Oversubscription is fundamentally a statistical bet: the platform must ensure that resources are kept as busy as possible.", "note": "AWS's own Firecracker paper (co-authored by the Lambda team, cited from Lambda's docs) states Lambda uses 'soft allocation' -- CPU is proportionally entitled per configured memory but oversubscribed/over-committed across MicroVMs on the same host, i.e. no guaranteed dedicated core, matching the 'shared' definition. AWS's public docs (gettingstarted-limits.html, lambda-runtime-environment.html) only ever describe memory-proportional CPU allocation, never a dedicated-core guarantee or a published burst baseline %, so 'burstable:NN' does not apply. This applies identically to on-demand, on-demand-arm64, both Savings Plan modes (same underlying Firecracker execution environment, just a different discount), and provisioned concurrency (kept warm/initialized on the same oversubscribed Firecracker fleet -- provisioning concurrency changes billing/availability, not the CPU allocation model). Lambda Managed Instances (mode 'managed-instances') is different: it runs customer functions as containers directly on customer-owned EC2 Nitro instances (c7g.xlarge, c7i.xlarge, m7i.xlarge per the card) -- none of these are in EC2's T-family burstable-instance line (confirmed via AWS's EC2 burstable-performance-instances doc, which names only T8i/T4g/T3a/T3/T2 as burstable), so their vCPUs are standard non-burstable EC2 vCPUs, i.e. dedicated to the instance (shared only in the ordinary EC2 sense of no CPU-credit throttling, not oversubscribed the way Lambda's Firecracker fleet is). microvms (added 2026-10-14): same Firecracker fleet; the baseline is the billed allocation and the 4x peak is opportunistic (\"can vertically scale up to 4x\"), no dedicated-core guarantee published, so shared like the function modes.", "checked": "2026-09-30" }, "usage_api": { "value": true, "source": "https://docs.aws.amazon.com/AmazonCloudWatch/latest/APIReference/API_GetMetricStatistics.html", "quote": "GetMetricStatistics API returns CloudWatch metrics (Duration, Invocations) per Lambda function; Cost Explorer API breaks costs down by resource tag/label enabling per-function cost tracking", "note": "granularity: per-sandbox; cost: yes; CloudWatch metrics API (GetMetricStatistics, GetMetricData) reports billable usage (Duration in ms, Invocations count) per individual Lambda function. AWS Cost Explorer API and Cost and Usage Report API break costs down by Lambda function tags/dimensions. Together these provide complete per-sandbox usage and cost visibility.", "confidence": "high", "checked": "2026-10-01" }, "max_session_h": { "value": 0.25, "source": "https://docs.aws.amazon.com/lambda/latest/dg/gettingstarted-limits.html", "quote": "Maximum execution duration per MicroVM 8 hours (28,800 seconds)", "note": "Card value 0.25 h is the Lambda Functions invocation timeout (900 s). The microvms mode sets 8 h: lifespan including suspended time, not adjustable (\"--maximum-duration-in-seconds ... Range: 1–28,800 seconds (8 hours)\", microvms-launching.html).", "checked": "2026-10-14" } }, "locations": [ { "country": "US", "state": "VA", "city": "Ashburn", "provider_region": "us-east-1", "self_serve": true, "choosable": true }, { "country": "IE", "state": null, "city": "Dublin", "provider_region": "eu-west-1", "self_serve": true, "choosable": true }, { "country": "DE", "state": null, "city": "Frankfurt", "provider_region": "eu-central-1", "self_serve": true, "choosable": true }, { "country": "SE", "state": null, "city": "Stockholm", "provider_region": "eu-north-1", "self_serve": true, "choosable": true }, { "country": "GB", "state": null, "city": "London", "provider_region": "eu-west-2", "self_serve": true, "choosable": true }, { "country": "FR", "state": null, "city": "Paris", "provider_region": "eu-west-3", "self_serve": true, "choosable": true }, { "country": "IT", "state": null, "city": "Milan", "provider_region": "eu-south-1", "self_serve": true, "choosable": true }, { "country": "CH", "state": null, "city": "Zurich", "provider_region": "eu-central-2", "self_serve": true, "choosable": true }, { "country": "AU", "state": "NSW", "city": "Sydney", "provider_region": "ap-southeast-2", "self_serve": true, "choosable": true }, { "country": "IN", "state": null, "city": "Mumbai", "provider_region": "ap-south-1", "self_serve": true, "choosable": true }, { "country": "JP", "state": null, "city": "Tokyo", "provider_region": "ap-northeast-1", "self_serve": true, "choosable": true }, { "country": "SG", "state": null, "city": "Singapore", "provider_region": "ap-southeast-1", "self_serve": true, "choosable": true }, { "country": "MY", "state": null, "city": "Kuala Lumpur", "provider_region": "ap-southeast-5", "self_serve": true, "choosable": true }, { "country": "TH", "state": null, "city": "Bangkok", "provider_region": "ap-southeast-7", "self_serve": true, "choosable": true }, { "country": "TW", "state": null, "city": "Taipei", "provider_region": "ap-east-2", "self_serve": true, "choosable": true }, { "country": "ID", "state": null, "city": "Jakarta", "provider_region": "ap-southeast-6", "self_serve": true, "choosable": true }, { "country": "HK", "state": null, "city": "Hong Kong", "provider_region": "ap-east-1", "self_serve": true, "choosable": true }, { "country": "ZA", "state": null, "city": "Cape Town", "provider_region": "af-south-1", "self_serve": true, "choosable": true }, { "country": "OM", "state": null, "city": "Muscat", "provider_region": "me-south-1", "self_serve": true, "choosable": true }, { "country": "AE", "state": null, "city": "Dubai", "provider_region": "me-central-1", "self_serve": true, "choosable": true }, { "country": "IL", "state": null, "city": "Tel Aviv", "provider_region": "il-central-1", "self_serve": true, "choosable": true }, { "country": "MX", "state": null, "city": "Mexico City", "provider_region": "mx-central-1", "self_serve": true, "choosable": true } ], "locations_source": "https://aws.amazon.com/lambda/pricing/", "locations_quote": "Regional variance: the major us/eu/asia regions all charge tier-1 $0.0000166667; af-south-1 $0.0000221, ap-east-1 $0.00002292, eu-south-1 $0.0000195172, me-south-1 $0.0000206667; requests $0.20/M in most regions, up to $0.28/M (ap-east-1). Regional on-demand variance includes ap-east-2 Taipei, ap-so", "locations_checked": "2026-10-01", "audit_2026_10_14": { "applied": 3, "source": "checked against the provider's live pricing and docs pages", "reviewed": 3 } }