BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
BEGIN:DAYLIGHT
DTSTART:20250309T030000
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:PDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20251102T010000
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:PST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251003T040450Z
UID:3D3C4823-306B-4F83-B097-D3F0E5643C9C
DTSTART;TZID=America/Los_Angeles:20251002T183000
DTEND;TZID=America/Los_Angeles:20251002T200000
DESCRIPTION:The race toward Artificial General Intelligence (AGI) is accele
 rating\, powered by ever-larger and more capable LLMs. But as these models
  scale\, we are hitting the physical limits of general-purpose GPUs. Memor
 y access is now slower and more energy-hungry than computation itself. Pow
 er demands are straining datacenter infrastructure. And as models span mor
 e chips\, interconnect delays—not compute—become the bottleneck.\n\nTh
 is talk explores how custom silicon—such as AWS Trainium\, Google TPU\, 
 Cerebras WSE\, and Groq LPU—is designed to overcome these barriers.\n\nB
 y keeping memory closer to compute\, optimizing data movement\, and delive
 ring higher efficiency per watt\, domain-specific accelerators are reshapi
 ng the hardware stack around the needs of LLMs.\n\nAttendees will gain a c
 lear framework for understanding why this shift matters\, how it breaks pa
 st the Memory\, Power\, and Interconnect walls\, and what it means for bui
 lding a more scalable\, cost-effective\, and energy-efficient foundation o
 n the road to AGI.\n\nThis event is hosted by the [TEMS Silicon Valley Cha
 pter](https://r6.ieee.org/scv-tems/custom-silicon-beyond-gpus-scaling-inte
 lligence-into-the-agi-age/).\n\nSpeaker(s): Asheesh Goja\, \n\nMountain Mi
 ke's Pizza\, 2510 Augustine Drive\, Santa Clara\, California\, United Stat
 es\, 95054
LOCATION:Mountain Mike's Pizza\, 2510 Augustine Drive\, Santa Clara\, Calif
 ornia\, United States\, 95054
ORGANIZER:edcheng@ieee.org
SEQUENCE:79
SUMMARY:Custom Silicon Beyond GPUs—Scaling Intelligence into the AGI Age
URL;VALUE=URI:https://events.vtools.ieee.org/m/498701
X-ALT-DESC:Description: <br /><p>The race toward Artificial General Intelli
 gence (AGI) is accelerating\, powered by ever-larger and more capable LLMs
 . But as these models scale\, we are hitting the physical limits of genera
 l-purpose GPUs. Memory access is now slower and more energy-hungry than co
 mputation itself. Power demands are straining datacenter infrastructure. A
 nd as models span more chips\, interconnect delays&mdash\;not compute&mdas
 h\;become the bottleneck.</p>\n<p>This talk explores how custom silicon&md
 ash\;such as AWS Trainium\, Google TPU\, Cerebras WSE\, and Groq LPU&mdash
 \;is designed to overcome these barriers.</p>\n<p>By keeping memory closer
  to compute\, optimizing data movement\, and delivering higher efficiency 
 per watt\, domain-specific accelerators are reshaping the hardware stack a
 round the needs of LLMs.</p>\n<p>Attendees will gain a clear framework for
  understanding why this shift matters\, how it breaks past the Memory\, Po
 wer\, and Interconnect walls\, and what it means for building a more scala
 ble\, cost-effective\, and energy-efficient foundation on the road to AGI.
 </p>\n<p>This event is hosted by the <a href="https://r6.ieee.org/scv-tems
 /custom-silicon-beyond-gpus-scaling-intelligence-into-the-agi-age/" target
 ="_blank" rel="noopener">TEMS Silicon Valley Chapter</a>.</p>
END:VEVENT
END:VCALENDAR