AI Wiki

NVIDIA B100

AI HardwareNVIDIA
14 min read
Updated
Suggest editHistoryTalk
RawGraph

Last edited

Fact-checked

In review queue

Sources

31 citations

Revision

v5 · 2,857 words

Fact-checks are independent of edits: a reviewer re-verifies the article against its sources and stamps the date. How we verify

The NVIDIA B100 is a data center graphics processing unit (GPU) based on the Blackwell architecture, announced by NVIDIA CEO Jensen Huang at the GTC 2024 keynote on March 18, 2024.[1] It is the lower-power, drop-in-compatible Blackwell SKU: rated at 700 W thermal design power (TDP), it carries 192 GB of HBM3e memory, 8 TB/s of memory bandwidth, and roughly 7 petaFLOPS of dense (14 petaFLOPS sparse) FP4 tensor performance, packing two reticle-limited dies and 208 billion transistors into a single package.[2][3][10] The B100 was designed to slot into existing 8-GPU H100 HGX server platforms at the same power envelope, sitting one tier below the 1,000 W B200 in the Blackwell lineup.[2][16] Despite its initial billing as the "first to ship" Blackwell variant,[4] the B100 was substantially deprioritized over the course of 2024 amid manufacturing issues with NVIDIA's CoWoS-L packaging, and KeyBanc Capital Markets analyst John Vinh reported in late August 2024 that NVIDIA had "effectively canceled" the B100 in favor of the B200 (for hyperscalers) and a then-forthcoming B200A variant (for enterprise customers).[5][6] By the time Blackwell-based servers began shipping to large customers in volume in December 2024, NVIDIA's allocations were concentrated almost entirely on the GB200 NVL72 rack-scale system rather than HGX B100 boards.[7][8]

What is the NVIDIA B100?

NVIDIA's Blackwell architecture, named after American statistician David Blackwell, is the successor to the Hopper architecture used in the H100 and H200 data center accelerators.[9] Blackwell was first confirmed during an NVIDIA investor presentation in October 2023, and its data center accelerators, publicly named B100 and B200, were formally introduced at GTC 2024 on March 18, 2024.[9][1] Introducing the lineup, Huang framed the generational jump in blunt terms: "Hopper is fantastic, but we need bigger GPUs," he told the GTC audience in San Jose, presenting Blackwell as up to five times more powerful than Hopper.[30]

The architecture marked the first time NVIDIA used a multi-die (chiplet) design in a flagship data center GPU: each Blackwell package consists of two reticle-limited dies fabricated on TSMC's custom 4NP process and connected by a 10 TB/s NV-High Bandwidth Interface (NV-HBI), yielding a total transistor count of 208 billion (104 billion per die).[10][9] The B100 was the lower-power member of this family, engineered so that a Blackwell GPU could replace a Hopper GPU one-for-one inside HGX systems without redesigning chassis power and cooling.[2][3]

Three primary data center products were unveiled at GTC 2024: the air-cooled B100, the higher-power B200, and the GB200 Grace Blackwell Superchip, which combines two B200 GPUs with one Grace CPU on a single board.[1][11] All three share the same Blackwell silicon and use TSMC's CoWoS-L 2.5D packaging, which embeds passive silicon bridges within an organic redistribution layer (RDL) interposer.[9][12]

What are the B100 specifications?

The B100 was specified as a Blackwell GPU configured for the same 700 W per-GPU power envelope as the H100 and H200 it was intended to replace, allowing HGX B100 baseboards to be drop-in compatible with existing HGX H100 server infrastructure.[3][2] Per most published spec sheets:[2][13][14][15]

AttributeNVIDIA B100 (SXM)
ArchitectureBlackwell (dual GB100 dies via NV-HBI)
Process nodeTSMC 4NP (custom variant of N4P)
Transistors208 billion (104 B per die)
Memory192 GB HBM3e (8 stacks x 24 GB)
Memory bandwidth8 TB/s
Memory bus8192-bit (8 x 1024-bit HBM3e interfaces)
NVLink (5th gen)1.8 TB/s bidirectional per GPU
Chip-to-chip (NV-HBI)10 TB/s between dies
Tensor FP4 (dense / sparse)7 / 14 PFLOPS
Tensor FP8/FP6 (dense / sparse)3.5 / 7 PFLOPS
Tensor INT8 (dense / sparse)3.5 / 7 POPS
Tensor FP16/BF16 (dense / sparse)1.75-1.8 / 3.5 PFLOPS
FP64 (dense)~30 TFLOPS
TDP700 W (air-cooled)
SocketSXM6
PCIe interfacePCIe Gen 6
Form factor8-GPU HGX B100 baseboard

The B100's distinguishing trait within the Blackwell lineup was its thermal envelope: at 700 W, it delivered roughly the same per-GPU power draw as the H100, but with 2.4x the HBM memory (192 GB vs. 80 GB), 2.5x the memory bandwidth (8 TB/s vs. 3.2 TB/s), and substantially higher AI throughput.[2][14] Air-cooled HGX B100 systems were quoted as delivering 14 PFLOPS of FP4 per GPU at the same 700 W envelope as Hopper.[11]

By comparison, the B200 used the same silicon but raised TDP to 1,000 W, increasing dense FP4 throughput to 9 PFLOPS (18 PFLOPS sparse) and dense FP16/BF16 to 2.25 PFLOPS.[16][17] The liquid-cooled GB200, which pairs two B200 GPUs with a Grace CPU, ran each Blackwell GPU at up to ~1,200 W and delivered up to 20 PFLOPS FP4 per GPU.[11]

How does the dual-die chiplet design and NV-HBI work?

Unlike the monolithic H100, both the B100 and B200 are physically composed of two GB100 dies sitting side-by-side on a CoWoS-L interposer.[10][9] The two dies are joined by NV-HBI, an internal 10 TB/s interconnect that NVIDIA describes as based on the NVLink 7 protocol and that allows the software stack to see the package as a single logical GPU.[9] This was the first time NVIDIA shipped a multi-die flagship data center GPU; AMD had previously taken a similar approach with the MI250X and MI300 accelerators.

How much memory does the B100 have?

Each die in a Blackwell package interfaces with four HBM3e stacks of 24 GB and 1 TB/s of bandwidth on a 1024-bit bus, yielding the package totals of 192 GB of HBM3e and 8 TB/s of bandwidth.[16] Memory capacity per GPU was a major leap over Hopper, where the H100 carried 80 GB HBM3 and the H200 carried 141 GB HBM3e.[14]

How fast is the B100 interconnect?

The B100 uses NVIDIA's 5th-generation NVLink, providing 1.8 TB/s of bidirectional bandwidth per GPU, double Hopper's NVLink 4 bandwidth, for high-speed GPU-to-GPU communication within an HGX system.[16][14] An 8-GPU HGX B100 baseboard could therefore aggregate 14.4 TB/s of NVLink bandwidth, matched by an NVSwitch fabric on the baseboard.

What form factors and systems used the B100?

Two server-level form factors were specified for the B100:

  • HGX B100 baseboard: 8 SXM6-socketed B100 GPUs on a single board, paired over fifth-generation NVLink and NVSwitch. HGX B100 was explicitly engineered to be a drop-in upgrade for HGX H100/H200 baseboards so that OEM systems could be updated to Blackwell without redesigning chassis power and cooling.[2][3]
  • Original equipment manufacturer (OEM) servers: At launch, Dell announced PowerEdge XE9680 configurations with the air-cooled HGX B100 option, Supermicro committed to making its existing HGX H100/H200 8-GPU systems drop-in ready for HGX B100 to shorten time-to-delivery, and Hewlett Packard Enterprise (HPE) and Lenovo similarly announced HGX B100 server plans.[18][3]

Cloud providers including Amazon Web Services, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure, and Lambda Labs all announced plans on March 18, 2024 to host Blackwell-based instances, with AWS specifically naming "EC2 instances featuring the new B100 GPUs deployed in EC2 UltraClusters" in its joint announcement with NVIDIA.[19] However, several of those announcements, Lambda Labs in particular, concretely listed only B200 and GB200 instances in their actual on-demand and reserved offerings, with B100 mentioned only descriptively rather than as a planned SKU.[20]

Why was the B100 effectively canceled?

Initial ramp plan (March-July 2024)

NVIDIA's initial March 2024 messaging cast the air-cooled 700 W B100 as the first Blackwell variant to ship.[4] SemiAnalysis, writing in early April 2024, described the B100 as targeted at ramping first, with the B200 "very soon after" and the liquid-cooled GB200 NVL72 rack-scale system later in the year.[4] TrendForce on August 7, 2024 still expected B100 and B200 shipments to "commence after 3Q24" for cloud-service-provider customers, with broader OEM availability in 2025.[21]

Design flaw and CoWoS-L yield problems (August 2024)

In early August 2024, multiple outlets reported a Blackwell design flaw that required NVIDIA to perform a mask respin on the package, pushing volume shipments by at least a quarter.[22][23] SemiAnalysis attributed the root cause to NVIDIA's first high-volume use of TSMC's CoWoS-L packaging: embedding multiple silicon bridges in the RDL interposer caused a coefficient-of-thermal-expansion (CTE) mismatch between the GPU dies, LSI bridges, RDL interposer, and motherboard substrate, leading to substrate warpage and broken connections.[7][24] The top global routing metal layers and the Blackwell die's bump-out had to be redesigned.[7][24]

At a public appearance in October 2024, CEO Jensen Huang accepted responsibility: "We had a design flaw in Blackwell. It was functional, but the design flaw caused the yield to be low. It was 100% NVIDIA's fault."[25][31] He credited TSMC with helping NVIDIA recover yields and resume manufacturing, dismissing reports of friction between the two companies as "fake news."[25][31]

"Effective cancellation" of B100 (August 2024)

On August 22, 2024, KeyBanc Capital Markets analyst John Vinh issued a client note stating that "given the Blackwell delay, we believe NVIDIA will prioritize the ramp of B200 for hyperscalers and has effectively canceled B100, which will be replaced with a lower cost/performance GPU (B200A) targeted at enterprise customers."[5][6] Vinh's note was widely reported by financial and trade press over the following week.[6]

Concurrent supply-chain reporting from SemiAnalysis was even more explicit at the product-line level: "HGX form-factors with the B100 and B200 are effectively now being cancelled outside of some initial lower volumes," with NVIDIA instead focusing what limited CoWoS-L capacity it had "almost entirely on GB200 NVL 36x2 and NVL72 rack-scale systems."[7] The B200A, based on a single-die "B102" SKU using CoWoS-S packaging with 144 GB of HBM3e and the same 700 W envelope as the B100, was positioned as the de-facto replacement for both HGX B100 and HGX B200 in 8-GPU server form factors, with availability targeted at the second half of 2025.[7][6]

Volume shipments concentrated on GB200 (Q4 2024-2025)

By late 2024, NVIDIA's recovered Blackwell capacity flowed predominantly into GB200 NVL36 and NVL72 rack-scale systems. Tom's Hardware reported that the first Blackwell server deliveries began the first week of December 2024, with Microsoft receiving one of the largest initial allocations, followed by Oracle, AWS, and Meta.[8] Foxconn and Quanta were the primary system integrators for GB200 racks.[8] Industry estimates compiled in late 2024 projected 150,000-200,000 Blackwell GB200 servers shipped in Q4 2024 alone, with 500,000-550,000 units in Q1 2025, essentially none of that volume being HGX B100 SKUs.[26]

By the time Blackwell Ultra (B300) and the GB300 NVL72 became NVIDIA's leading data-center platform in 2025, Supermicro confirmed it was shipping HGX B300 systems and GB300 NVL72 racks in volume, while the HGX B100 and HGX B200 8-GPU products had been quietly de-emphasized.[27]

Customer-visible availability

As of early 2026, public cloud-GPU rental aggregators listed only a single provider claiming on-demand access to B100 instances, and no major public on-demand pricing for B100 was advertised by the largest cloud service providers.[28] In practice, the B100 became a niche enterprise SKU that some specialist providers continued to list but that never anchored a major hyperscaler deployment.

How does the B100 compare to the B200 and other Blackwell SKUs?

The B100 and B200 share identical silicon, transistor count, and memory configuration; the practical differences are power, clock-driven throughput, and which form factor each anchored. The B100 runs at 700 W for drop-in Hopper replacement, while the B200 runs at 1,000 W for higher performance, and the GB200 pushes each Blackwell GPU to roughly 1,200 W in a liquid-cooled rack.[2][16][11]

GPUTDPMemoryFP4 denseFP4 sparseForm factorVolume trajectory
B100700 W192 GB HBM3e7 PFLOPS14 PFLOPSHGX B100 (SXM6)Effectively de-prioritized after Aug 2024[5][7]
B2001,000 W192 GB HBM3e9 PFLOPS18 PFLOPSHGX B200 (SXM6)Limited HGX volume; volume concentrated in GB200[7][8]
GB200 (per GPU)up to 1,200 W192 GB HBM3e10 PFLOPS20 PFLOPSGB200 NVL36 / NVL72 rackPrimary Blackwell shipping product Q4 2024-2025[8][11]
B200A (planned)700 W / 1,000 W variants144 GB HBM3eN/AN/AHGX 8-GPU using CoWoS-SReplacement for HGX B100/B200 in 2H 2025[7][6]

Sources for the B100 / B200 / GB200 comparison are NVIDIA's GTC 2024 materials and follow-on independent reporting; the GB300 (Blackwell Ultra) succeeded both at the top of the lineup in 2025.[27][29]

How was the B100 received by the industry?

Initial reception of the Blackwell announcement on March 18, 2024 was overwhelmingly positive for the architecture as a whole: press coverage emphasized 5x H100 AI performance per GPU, 192 GB of HBM3e, and a new FP4 data type for trillion-parameter inference.[1][11] The Register noted that with Blackwell, "Nvidia is turning up the AI heat," characterizing the lineup, including the 700 W B100, as a clear generational leap.[11]

Within roughly five months, however, the narrative around the specific B100 SKU had shifted from "the first Blackwell to ship" to "effectively canceled." Trade press, financial analysts, and supply-chain researchers converged on the assessment that the combination of CoWoS-L yield issues, finite TSMC packaging capacity, and overwhelming hyperscaler demand for rack-scale GB200 had collapsed the original three-tier product strategy (B100 / B200 HGX / GB200 rack) down to a one-tier rollout focused on GB200 rack-scale systems.[7][5][6] The B200A and the subsequent Blackwell Ultra refresh (B300/GB300) absorbed the demand the B100 had originally been designed to serve.[27][7]

Independent benchmarking sites and cloud-GPU comparison guides published throughout 2025 and into 2026 routinely observed that "the industry as a whole skipped over B100s to go straight to their more powerful siblings" and that "B100s are difficult to find" in commercial deployments.[14]

See also

References

  1. NVIDIA Newsroom, "NVIDIA Blackwell Platform Arrives to Power a New Era of Computing," March 18, 2024. https://nvidianews.nvidia.com/news/nvidia-blackwell-platform-arrives-to-power-a-new-era-of-computing ↩
  2. CUDO Compute, "NVIDIA's Blackwell Architecture: Breaking Down the B100, B200, and GB200." https://www.cudocompute.com/blog/nvidias-blackwell-architecture-breaking-down-the-b100-b200-and-gb200 ↩
  3. Server-Parts.eu, "NVIDIA Blackwell B100 vs B200 Comparison: What is the difference?" https://www.server-parts.eu/post/nvidia-blackwell-b100-b200-comparison-differences ↩
  4. SemiAnalysis, "NVIDIA Blackwell Perf TCO Analysis - B100 vs B200 vs GB200 NVL72," April 10, 2024. https://newsletter.semianalysis.com/p/nvidia-blackwell-perf-tco-analysis ↩
  5. Wccftech, "KeyBanc Believes That NVIDIA Has 'Effectively Canceled' The B100 Chip." https://wccftech.com/keybanc-believes-that-nvidia-has-effectively-canceled-the-b100-chip/ ↩
  6. TweakTown, "Analyst: NVIDIA has 'effectively canceled' B100 AI GPU over design flaw: B200A to replace it." https://www.tweaktown.com/news/100083/analyst-nvidia-has-effectively-canceled-b100-ai-gpu-over-design-flaw-b200a-to-replace-it/index.html ↩
  7. SemiAnalysis, "Nvidia's Blackwell Reworked - Shipment Delays & GB200A Reworked Platforms," August 4, 2024. https://newsletter.semianalysis.com/p/nvidias-blackwell-reworked-shipment ↩
  8. Tom's Hardware, "Nvidia will reportedly start Blackwell server deliveries in early December. The first servers will go to Microsoft." https://www.tomshardware.com/pc-components/gpus/nvidia-will-reportedly-start-blackwell-server-deliveries-in-early-december-the-first-servers-will-go-to-microsoft ↩
  9. Wikipedia, "Blackwell (microarchitecture)." https://en.wikipedia.org/wiki/Blackwell_(microarchitecture) ↩
  10. Wccftech, "NVIDIA Blackwell GPU Architecture Official: 208 Billion Transistors, 5x AI Performance, 192 GB HBM3e Memory, 8 TB/s Bandwidth." https://wccftech.com/nvidia-blackwell-gpu-architecture-official-208-billion-transistors-5x-ai-performance-192-gb-hbm3e-memory/ ↩
  11. The Register, "Nvidia turns up the AI heat with 1,200W Blackwell GPUs," March 18, 2024. https://www.theregister.com/2024/03/18/nvidia_turns_up_the_ai/ ↩
  12. 3D InCites, "Why Nvidia's Blackwell is Having Issues with TSMC CoWoS-L Technology," October 2024. https://www.3dincites.com/2024/10/iftle-607-why-nvidias-blackwell-is-having-issues-with-tsmc-cowos-l-technology/ ↩
  13. Hyperstack, "Everything You Need to Know About the NVIDIA Blackwell GPUs." https://www.hyperstack.cloud/blog/thought-leadership/everything-you-need-to-know-about-the-nvidia-blackwell-gpus ↩
  14. Modal, "NVIDIA Showdown: H100s and H200s vs B100s and B200s." https://modal.com/blog/h100-and-h200-vs-b100-and-b200 ↩
  15. Flopper.io, "NVIDIA B100 SXM 192GB Spec Sheet." https://flopper.io/gpu/nvidia-b100-sxm-192gb/spec-sheet.pdf ↩
  16. Exxact, "NVIDIA GTC 2024 Recap | NVIDIA Blackwell, NIMs, Robotics." https://www.exxactcorp.com/blog/news/nvidia-gtc-2024-recap ↩
  17. Civo, "Comparing NVIDIA's B200 and H100: A deep dive into next-gen AI performance." https://www.civo.com/blog/comparing-nvidia-b200-and-h100 ↩
  18. Constellation Research, "Dell Technologies preps new AI servers with Nvidia's B100, B200, GB200 SuperChip." https://www.constellationr.com/blog-news/insights/dell-technologies-preps-new-ai-servers-nvidia-s-b100-b200-gb200-superchip ↩
  19. NVIDIA Newsroom, "AWS and NVIDIA Extend Collaboration to Advance Generative AI Innovation," March 18, 2024. https://nvidianews.nvidia.com/news/aws-nvidia-generative-ai-innovation ↩
  20. Lambda, "Lambda among first NVIDIA Cloud Partners to deploy NVIDIA Blackwell-based GPUs," March 18, 2024. https://lambda.ai/blog/lambda-among-first-nvidia-cloud-partners-to-deploy-nvidia-blackwell-based-gpus ↩
  21. TrendForce, "NVIDIA's New B200A Targets OEM Customers; High-End GPU Shipments Expected to Grow 55% in 2025," August 7, 2024. https://www.trendforce.com/presscenter/news/20240807-12244.html ↩
  22. The Register, "Nvidia delays Blackwell GPUs until 2025, say reports," August 5, 2024. https://www.theregister.com/2024/08/05/nvidia_delays_blackwell_gpus_until/ ↩
  23. Tom's Hardware, "Nvidia Blackwell GPUs allegedly delayed due to design flaws; launch expected to be pushed back by three months or more." https://www.tomshardware.com/pc-components/gpus/nvidia-blackwell-gpus-allegedly-delayed-due-to-design-flaws ↩
  24. EDN, "Nvidia, TSMC, and advanced packaging realignment in 2025." https://www.edn.com/nvidia-tsmc-and-advanced-packaging-realignment-in-2025/ ↩
  25. TweakTown, "NVIDIA CEO: 'we had a design flaw in Blackwell, it was 100% NVIDIA's fault' not TSMC's fault," October 24, 2024. https://www.tweaktown.com/news/101277/nvidia-ceo-we-had-design-flaw-in-blackwell-it-was-100-nvidias-fault-not-tsmcs/index.html ↩
  26. Wccftech, "NVIDIA To Ship 150K-200K Blackwell GB200 AI Servers In Q4 2024 Alone & 500-550K Units In Q1 2025, Microsoft Being The Leading Buyer." https://wccftech.com/nvidia-ship-150k-200k-blackwell-gb200-ai-servers-q4-2024-500-550k-units-q1-2025/ ↩
  27. Supermicro, "Supermicro Begins Volume Shipments of NVIDIA Blackwell Ultra Systems and Rack Plug-and-Play Data Center-Scale Solutions." https://www.supermicro.com/en/pressreleases/supermicro-begins-volume-shipments-nvidia-blackwell-ultra-systems-and-rack-plug-and ↩
  28. GetDeploying, "B100 Cloud Pricing: Compare 1+ Providers (2026)." https://getdeploying.com/gpus/nvidia-b100 ↩
  29. NVIDIA, "Blackwell Architecture: The Engine Behind AI Factories." https://www.nvidia.com/en-us/data-center/technologies/blackwell-architecture/ ↩
  30. The National, "Nvidia GTC: Jensen Huang unveils new Blackwell B200 AI superchip," March 19, 2024. https://www.thenationalnews.com/future/technology/2024/03/19/nvidia-gtc-jensen-huang-blackwell-ai-superchip/ ↩
  31. Tom's Hardware, "Nvidia's Jensen Huang admits AI chip design flaw was '100% Nvidia's fault' - TSMC not to blame, now-fixed Blackwell chips are in production," October 2024. https://www.tomshardware.com/tech-industry/artificial-intelligence/nvidias-jensen-huang-admits-ai-chip-design-flaw-was-100-percent-nvidias-fault-tsmc-not-to-blame-now-fixed-blackwell-chips-are-in-production ↩

Improve this article

Add missing citations, update stale details, or suggest a clearer explanation. Every suggestion is reviewed for sourcing before it goes live.

4 revisions by 1 contributor · full history

Suggest edit