AI09/25/2026@TechDesk
Qualcomm and Amazon Agree a Multi-Generation AI Data-Centre Chip Deal
QualcommQualcomm N, 5775, Morehouse Drive, Barnes Canyon, Sorrento Mesa, San Diego, San Diego County, California, 92121, United States
Starts 09/08/2026 1:00PMCONFIRMED 92% CONFIDENCE17 days agoThe Business Wire copy of the release was published at 13:00 UTC on 8 September 2026; Qualcomm's^[1] 8-K says the warrant was issued 'On September 3, 2026, in connection with a strategic collaboration' with Amazon.
Qualcomm will build customised AI inference silicon and up-to-1.6T optical connectivity for Amazon's data centres over several generations, and gave Amazon a warrant for up to 25 million Qualcomm shares vesting against up to $60 billion of purchases.
$60B
0.79
AI09/25/2026@TechDesk
HUMAIN Picks Qualcomm AI200 and AI250 Racks for 200 MW in Saudi Arabia
QualcommRiyadh, Riyadh governorate, Riyadh Region, 12643, Saudi Arabia
Starts 10/27/2025CONFIRMED 87% CONFIDENCE333 days agoThe release dated 27 October 2025 says HUMAIN and Qualcomm^[1][2][3][4] 'today announced' the collaboration 'ahead of the 9th edition of the Future Investment Initiative (FII) conference'.
Saudi AI company HUMAIN became the first customer for Qualcomm's data-centre chips, targeting 200 megawatts of AI200 and AI250 racks from 2026 to sell AI inference services in the Kingdom and worldwide.
0.78
Semiconductors09/25/2026@TechDesk
Qualcomm's Dragonfly AI250 With High Bandwidth Compute Due in 2027
Qualcomm N, 5775, Morehouse Drive, Barnes Canyon, Sorrento Mesa, San Diego, San Diego County, California, 92121, United StatesQualcomm
Starts 12/31/2027ESTIMATED 66% CONFIDENCEin 1.3 yearsThe October 2025 launch release says AI250 is 'expected to be commercially available in ... 2027', and the June 2026 roadmap expects 'Commercial sampling of HBC Gen 1 with AI250 ... in mid-2027'; a bare year is dated to 31 December.^[1]
Qualcomm's second-generation data-center AI accelerator, the Dragonfly AI250, is due in 2027 with High Bandwidth Compute near-memory stacking that gives 133 TB/s of effective bandwidth per card, 18 times the AI200.
0.77
Business09/25/2026@TechDesk
Qualcomm Agrees to Buy AI Software Startup Modular
Qualcomm N, 5775, Morehouse Drive, Barnes Canyon, Sorrento Mesa, San Diego, San Diego County, California, 92121, United StatesQualcomm
Starts 06/24/2026 12:00PMCONFIRMED 86% CONFIDENCE93 days agoThe Business Wire copy of 'Qualcomm^[1] to Acquire Modular^[5]' was published at 12:00 UTC on 24 June 2026; Qualcomm's 8-K says the definitive agreement was entered into on June 21, 2026.
Qualcomm agreed to buy Modular, Chris Lattner's startup whose AI software runs models across CPUs, GPUs, NPUs and custom chips without rewrites, in an all-stock deal of up to 19.2 million shares valued at about $3.9 billion.
$3.9B
0.76
Robotics09/25/2026@TechDesk
Arduino Opens Pre-Orders for the $299 VENTUNO Q Edge-AI Board
QualcommTurin, Piedmont, Italy
Starts 08/25/2026 2:00PMCONFIRMED 88% CONFIDENCE31 days agoThe release is datelined 'Aug 25, 2026 | TURIN' and says Arduino^[2] 'today opened pre-orders for Arduino VENTUNO Q'; Qualcomm's^[1][3] page stamps it 14:00 UTC.
Qualcomm's Arduino opened worldwide pre-orders for the VENTUNO Q, a $299 board pairing a 40 TOPS Dragonwing IQ-8275 processor and 16 GB of memory with an STM32H5 microcontroller so robots and machines can run AI models and act on them offline.
$299.00
0.73
Semiconductors09/25/2026@TechDesk
Qualcomm Launches Dragonfly Data Center Line: C1000 CPU and AI300 at Investor Day
QualcommManhattan, New York County, New York, United States
Starts 06/24/2026 7:30PMCONFIRMED 84% CONFIDENCE93 days agoThe release, dated Jun 24, 2026 and datelined NEW YORK, says Qualcomm^[1][2][3] 'today announced at its Investor Day' the new data center solutions; Qualcomm's page stamps it 19:30 UTC.
At its 2026 Investor Day in New York, Qualcomm branded its data-center products Dragonfly and unveiled the 250-plus-core Dragonfly C1000 CPU, the AI300 accelerator and High Bandwidth Compute memory, with Meta signed up for the CPU; the stock jumped 15%.
0.72
Business09/25/2026@TechDesk
Qualcomm Adds Carnegie Mellon AI Professor Zico Kolter to Its Board
QualcommQualcomm N, 5775, Morehouse Drive, Barnes Canyon, Sorrento Mesa, San Diego, San Diego County, California, 92121, United States
Starts 09/02/2025CONFIRMED 84% CONFIDENCE1.1 years agoThe 8-K says "On September 2, 2025, the Board of Directors ... elected Dr. Jeremy (Zico) Kolter".^[1]
Qualcomm elected Zico Kolter, head of Carnegie Mellon's Machine Learning Department, to its board as it pushes into AI chips for devices and data centers.
0.72
Semiconductors09/24/2026@TechDesk
OpenAI and Broadcom Unveil Jalapeño, OpenAI's First AI Chip
3421 Hillview Avenue, Palo Alto, California 94306, United StatesBroadcom
Starts 06/24/2026 1:00PMCONFIRMED 83% CONFIDENCE93 days agoThe release is dated "June 24, 2026" and says OpenAI and Broadcom^[1] "today unveiled" Jalapeño; Nasdaq's^[4] GlobeNewswire copy is stamped "Jun 24, 2026 9:00am EDT" (13:00 UTC).
OpenAI's first Intelligence Processor, an inference accelerator it designed and Broadcom implemented, went from design to tape-out in nine months; Hock Tan and Charlie Kawwas handed it to Sam Altman and Greg Brockman.
0.72
AI09/25/2026@TechDesk
Qualcomm and HUMAIN Sign an MoU to Build AI Data Centres in Saudi Arabia
King Abdul Aziz International Conference Center, Shaban Bin Abi Sheha, Al Hada District, Shmeisi Municipality, Riyadh, Riyadh governorate, Riyadh Region, 12916, Saudi ArabiaQualcomm
Starts 05/13/2025CONFIRMED 87% CONFIDENCE1.4 years agoThe release dated 13 May 2025 says the companies 'today signed a Memorandum of Understanding' during the Saudi-US Investment Forum in Riyadh; the White House dates that forum Tuesday, May 13, 2025.^[1]
At the Saudi-US Investment Forum in Riyadh, Qualcomm and Saudi Arabia's national AI company HUMAIN agreed to develop AI data centres and supply Qualcomm data-centre CPU and AI chips, with a Qualcomm chip design centre for the Kingdom.
0.72
Semiconductors07/23/2017@ThePinGang
Intel Movidius Neural Compute Stick available for US$79
Intel
Starts 07/21/2017 CONFIRMED 90% CONFIDENCE9.2 years agoHappened as dated, per hexus.net^[1]: "Intel^[2] Movidius Neural Compute Stick available for US$79 Tags: Intel (NASDAQ:INTC) Quick Link: HE…"
Offers more than 100 GFLOPS of inference performance within a 1W power envelope.
$79.00
0.71
AI09/23/2026@TechDesk
NVIDIA Vera Rubin NVL72 Debuts in MLPerf Inference v6.1
NVIDIANVIDIA Endeavor, 2788 San Tomas Expressway, Santa Clara, California 95051, United States
Starts 09/16/2026 3:00PMCONFIRMED 86% CONFIDENCE9 days agoMLCommons^[4] published the MLPerf Inference v6.1 results on 16 September 2026 - its GlobeNewswire release is datelined 'SAN FRANCISCO, Sept. 16, 2026' and stamped '11:00 AM EDT' - and NVIDIA's^[1][2] post of the same day says the results were 'released today'.
In its first peer-reviewed benchmark, NVIDIA's Vera Rubin NVL72 rack posted up to 3.7x the throughput of GB300 NVL72 on Qwen3-VL and 2.5x on DeepSeek-R1, submitted in the preview category of MLCommons' MLPerf Inference v6.1 round.
0.71
Semiconductors09/20/2026@TechDesk
AMD Launches Helios Rack and Instinct MI450 GPUs
AMDSanta Clara, California
Starts 07/23/2026CONFIRMED 88% CONFIDENCE64 days agoYahoo^[1] Finance reported the launch on July 23, 2026.
AMD launched its Helios rack-scale AI system, built on MI455X GPUs and Venice EPYC CPUs, with shipments to Microsoft starting in the second half of 2026 and a 2 GW Anthropic deployment from 2027.
0.70

CNBC^[4] reports 'Qualcomm^[1][2][3] announced Monday' (Mon, Oct 27 2025, published 10:00 AM EDT) that it will release the AI200 and AI250; Qualcomm's release carries the same announcement.

Start: Oct 27, 202590% CONFIDENCEfrom the source

Qualcomm Unveils AI200 and AI250 Rack-Scale AI Inference Chips

Qualcomm entered the data-center AI market with the AI200 and AI250 accelerator cards and liquid-cooled racks, with 768 GB of memory per card, due in 2026 and 2027; its stock jumped 11% that day.

Qualcomm announced the AI200 and AI250, chip-based accelerator cards and full racks for data-center AI inference, part of a data-center roadmap with an annual cadence^[1]
AI200 was expected to be commercially available in 2026 and AI250 in 2027^[1][4]
Qualcomm shares rose 11% on the day, their best day since April^[4]
Durga Malladi said the products let customers deploy AI 'at unprecedented TCO', with one-click deployment of Hugging Face models^[1]

Notable features

AI200: 768 GB of LPDDR memory per card for large language and multimodal model inference^[1]
AI250: a near-memory computing architecture with more than 10x higher effective memory bandwidth and much lower power^[1]
Both racks: direct liquid cooling, PCIe scale-up, Ethernet scale-out, confidential computing and 160 kW of rack power^[1]
Built on Qualcomm's Hexagon NPU technology^[1][4]

@TechDesk

References 484% CONFIDENCE

The first entry is always the pin's source. Overall confidence is a weighted average of how firmly each reference supports the start and end times used above; a reference counts half as much for every 180 days older than the newest.

[1]90%
qualcomm.com/news/releases/2025/10/qualcomm-unveils-ai200-and-ai250-redefining-rack-scale-data-cent
qualcomm.com· Posted Sep 25, 2026· Starts Oct 27, 2025 ✓· 34% of score
CNBC^[4] reports 'Qualcomm^[2][3] announced Monday' (Mon, Oct 27 2025, published 10:00 AM EDT) that it will release the AI200 and AI250; Qualcomm's release carries the same announcement.
[2]82%
Qualcomm Dragonfly AI200 product page (Qualcomm)
qualcomm.com· Added Sep 25, 2026· 34% of score
Today's product page: 768 GB LPDDR per card, 43 TB per 140 kW liquid-cooled ORv3 rack, models up to 10 trillion parameters.^[1]
[3]75%
Qualcomm Unveils Comprehensive Data Center Roadmap ... Qualcomm Dragonfly Portfolio (Qualcomm)
qualcomm.com· Published Jun 24, 2026· 23% of score
Later release naming them the 'previously announced Qualcomm^[1][2] Dragonfly AI200 and AI250'.
[4]90%
Qualcomm announces AI chips to compete with AMD and Nvidia - stock soars 11% (CNBC)
cnbc.com· Published Oct 27, 2025· Starts Oct 27, 2025· 9% of score
Dated Mon, Oct 27 2025: AI200 'will go on sale in 2026' and AI250 is 'planned for 2027'; shares rose 11%; 768 GB of memory per card.^[1]

Suggest a correction

Something missing or wrong? Say it in your own words: a link that backs this pin up, a different start or end date and why, or a fact it lacks or gets wrong. The AI checks it against this pin's sources, searches for better ones, and adds any page that backs you up. The pin's own sources still count most.

Comments

Thread

Respond to this Pin

AI09/25/2026@TechDesk
Qualcomm and Amazon Agree a Multi-Generation AI Data-Centre Chip Deal
QualcommQualcomm N, 5775, Morehouse Drive, Barnes Canyon, Sorrento Mesa, San Diego, San Diego County, California, 92121, United States
Starts 09/08/2026 1:00PMCONFIRMED 92% CONFIDENCE17 days agoThe Business Wire copy of the release was published at 13:00 UTC on 8 September 2026; Qualcomm's^[1] 8-K says the warrant was issued 'On September 3, 2026, in connection with a strategic collaboration' with Amazon.
Qualcomm will build customised AI inference silicon and up-to-1.6T optical connectivity for Amazon's data centres over several generations, and gave Amazon a warrant for up to 25 million Qualcomm shares vesting against up to $60 billion of purchases.
$60B
0.79
AI09/25/2026@TechDesk
HUMAIN Picks Qualcomm AI200 and AI250 Racks for 200 MW in Saudi Arabia
QualcommRiyadh, Riyadh governorate, Riyadh Region, 12643, Saudi Arabia
Starts 10/27/2025CONFIRMED 87% CONFIDENCE333 days agoThe release dated 27 October 2025 says HUMAIN and Qualcomm^[1][2][3][4] 'today announced' the collaboration 'ahead of the 9th edition of the Future Investment Initiative (FII) conference'.
Saudi AI company HUMAIN became the first customer for Qualcomm's data-centre chips, targeting 200 megawatts of AI200 and AI250 racks from 2026 to sell AI inference services in the Kingdom and worldwide.
0.78
Semiconductors09/25/2026@TechDesk
Qualcomm's Dragonfly AI250 With High Bandwidth Compute Due in 2027
Qualcomm N, 5775, Morehouse Drive, Barnes Canyon, Sorrento Mesa, San Diego, San Diego County, California, 92121, United StatesQualcomm
Starts 12/31/2027ESTIMATED 66% CONFIDENCEin 1.3 yearsThe October 2025 launch release says AI250 is 'expected to be commercially available in ... 2027', and the June 2026 roadmap expects 'Commercial sampling of HBC Gen 1 with AI250 ... in mid-2027'; a bare year is dated to 31 December.^[1]
Qualcomm's second-generation data-center AI accelerator, the Dragonfly AI250, is due in 2027 with High Bandwidth Compute near-memory stacking that gives 133 TB/s of effective bandwidth per card, 18 times the AI200.
0.77
Business09/25/2026@TechDesk
Qualcomm Agrees to Buy AI Software Startup Modular
Qualcomm N, 5775, Morehouse Drive, Barnes Canyon, Sorrento Mesa, San Diego, San Diego County, California, 92121, United StatesQualcomm
Starts 06/24/2026 12:00PMCONFIRMED 86% CONFIDENCE93 days agoThe Business Wire copy of 'Qualcomm^[1] to Acquire Modular^[5]' was published at 12:00 UTC on 24 June 2026; Qualcomm's 8-K says the definitive agreement was entered into on June 21, 2026.
Qualcomm agreed to buy Modular, Chris Lattner's startup whose AI software runs models across CPUs, GPUs, NPUs and custom chips without rewrites, in an all-stock deal of up to 19.2 million shares valued at about $3.9 billion.
$3.9B
0.76
Robotics09/25/2026@TechDesk
Arduino Opens Pre-Orders for the $299 VENTUNO Q Edge-AI Board
QualcommTurin, Piedmont, Italy
Starts 08/25/2026 2:00PMCONFIRMED 88% CONFIDENCE31 days agoThe release is datelined 'Aug 25, 2026 | TURIN' and says Arduino^[2] 'today opened pre-orders for Arduino VENTUNO Q'; Qualcomm's^[1][3] page stamps it 14:00 UTC.
Qualcomm's Arduino opened worldwide pre-orders for the VENTUNO Q, a $299 board pairing a 40 TOPS Dragonwing IQ-8275 processor and 16 GB of memory with an STM32H5 microcontroller so robots and machines can run AI models and act on them offline.
$299.00
0.73
Semiconductors09/25/2026@TechDesk
Qualcomm Launches Dragonfly Data Center Line: C1000 CPU and AI300 at Investor Day
QualcommManhattan, New York County, New York, United States
Starts 06/24/2026 7:30PMCONFIRMED 84% CONFIDENCE93 days agoThe release, dated Jun 24, 2026 and datelined NEW YORK, says Qualcomm^[1][2][3] 'today announced at its Investor Day' the new data center solutions; Qualcomm's page stamps it 19:30 UTC.
At its 2026 Investor Day in New York, Qualcomm branded its data-center products Dragonfly and unveiled the 250-plus-core Dragonfly C1000 CPU, the AI300 accelerator and High Bandwidth Compute memory, with Meta signed up for the CPU; the stock jumped 15%.
0.72
Business09/25/2026@TechDesk
Qualcomm Adds Carnegie Mellon AI Professor Zico Kolter to Its Board
QualcommQualcomm N, 5775, Morehouse Drive, Barnes Canyon, Sorrento Mesa, San Diego, San Diego County, California, 92121, United States
Starts 09/02/2025CONFIRMED 84% CONFIDENCE1.1 years agoThe 8-K says "On September 2, 2025, the Board of Directors ... elected Dr. Jeremy (Zico) Kolter".^[1]
Qualcomm elected Zico Kolter, head of Carnegie Mellon's Machine Learning Department, to its board as it pushes into AI chips for devices and data centers.
0.72
Semiconductors09/24/2026@TechDesk
OpenAI and Broadcom Unveil Jalapeño, OpenAI's First AI Chip
3421 Hillview Avenue, Palo Alto, California 94306, United StatesBroadcom
Starts 06/24/2026 1:00PMCONFIRMED 83% CONFIDENCE93 days agoThe release is dated "June 24, 2026" and says OpenAI and Broadcom^[1] "today unveiled" Jalapeño; Nasdaq's^[4] GlobeNewswire copy is stamped "Jun 24, 2026 9:00am EDT" (13:00 UTC).
OpenAI's first Intelligence Processor, an inference accelerator it designed and Broadcom implemented, went from design to tape-out in nine months; Hock Tan and Charlie Kawwas handed it to Sam Altman and Greg Brockman.
0.72
AI09/25/2026@TechDesk
Qualcomm and HUMAIN Sign an MoU to Build AI Data Centres in Saudi Arabia
King Abdul Aziz International Conference Center, Shaban Bin Abi Sheha, Al Hada District, Shmeisi Municipality, Riyadh, Riyadh governorate, Riyadh Region, 12916, Saudi ArabiaQualcomm
Starts 05/13/2025CONFIRMED 87% CONFIDENCE1.4 years agoThe release dated 13 May 2025 says the companies 'today signed a Memorandum of Understanding' during the Saudi-US Investment Forum in Riyadh; the White House dates that forum Tuesday, May 13, 2025.^[1]
At the Saudi-US Investment Forum in Riyadh, Qualcomm and Saudi Arabia's national AI company HUMAIN agreed to develop AI data centres and supply Qualcomm data-centre CPU and AI chips, with a Qualcomm chip design centre for the Kingdom.
0.72
Semiconductors07/23/2017@ThePinGang
Intel Movidius Neural Compute Stick available for US$79
Intel
Starts 07/21/2017 CONFIRMED 90% CONFIDENCE9.2 years agoHappened as dated, per hexus.net^[1]: "Intel^[2] Movidius Neural Compute Stick available for US$79 Tags: Intel (NASDAQ:INTC) Quick Link: HE…"
Offers more than 100 GFLOPS of inference performance within a 1W power envelope.
$79.00
0.71
AI09/23/2026@TechDesk
NVIDIA Vera Rubin NVL72 Debuts in MLPerf Inference v6.1
NVIDIANVIDIA Endeavor, 2788 San Tomas Expressway, Santa Clara, California 95051, United States
Starts 09/16/2026 3:00PMCONFIRMED 86% CONFIDENCE9 days agoMLCommons^[4] published the MLPerf Inference v6.1 results on 16 September 2026 - its GlobeNewswire release is datelined 'SAN FRANCISCO, Sept. 16, 2026' and stamped '11:00 AM EDT' - and NVIDIA's^[1][2] post of the same day says the results were 'released today'.
In its first peer-reviewed benchmark, NVIDIA's Vera Rubin NVL72 rack posted up to 3.7x the throughput of GB300 NVL72 on Qwen3-VL and 2.5x on DeepSeek-R1, submitted in the preview category of MLCommons' MLPerf Inference v6.1 round.
0.71
Semiconductors09/20/2026@TechDesk
AMD Launches Helios Rack and Instinct MI450 GPUs
AMDSanta Clara, California
Starts 07/23/2026CONFIRMED 88% CONFIDENCE64 days agoYahoo^[1] Finance reported the launch on July 23, 2026.
AMD launched its Helios rack-scale AI system, built on MI455X GPUs and Venice EPYC CPUs, with shipments to Microsoft starting in the second half of 2026 and a 2 GW Anthropic deployment from 2027.
0.70

More like this