从电视修理工到电磁兼容专家：一位IEEE终身会员的职业逆袭

Opens in a new window Opens an external website Opens an external website in a new window This website utilizes technologies such as cookies to enable essential site functionality, as well as for analytics, personalization, and targeted advertising. [Privacy Policy](https://www.ieee.org/security-privacy.html) Accept Deny Non-Essential Manage Preferences Storage Preferences [IEEE.org](https://www.ieee.org/)[IEEE Xplore](https://ieeexplore.ieee.org/Xplore/home.jsp)[IEEE Standards](https://standards.ieee.org/)[IEEE Job Site](https://jobs.ieee.org/)[More Sites](https://www.ieee.org/sitemap.html) [Sign In](https://www.ieee.org/profile/public/createwebaccount/showCreateAccount.html?ShowMGAMarkeatbilityOptIn=true&sourceCode=spectrum&signinurl=https%3A%2F%2Fspectrum.ieee.org%2Fcore%2Fsaml%2Fmain%2Flogin%3Fnext_url%3Dhttps%3A%2F%2Fspectrum.ieee.org%2Fcore%2Fintegrations%2Fieee%2Fchanges%3Fredirect%3Dhttps%3A%2F%2Fspectrum.ieee.org%2Ffrom-tv-repairman-to-electromagnetic-compatibility-expert&url=https%3A%2F%2Fspectrum.ieee.org%2Fcore%2Fsaml%2Fmain%2Flogin%3Fnext_url%3Dhttps%3A%2F%2Fspectrum.ieee.org%2Fcore%2Fintegrations%2Fieee%2Fchanges%3Fredirect%3Dhttps%3A%2F%2Fspectrum.ieee.org%2Ffrom-tv-repairman-to-electromagnetic-compatibility-expert&autoSignin=Y&car=IEEE-Spectrum)[Join IEEE](https://spectrum.ieee.org/join) [](https://spectrum.ieee.org/ "Return to homepage") From TV Repairman to Electromagnetic Compatibility Expert Share FOR THE TECHNOLOGY INSIDER Search: Explore by topic [Aerospace](https://spectrum.ieee.org/topic/aerospace/)[AI](https://spectrum.ieee.org/topic/artificial-intelligence/)[Biomedical](https://spectrum.ieee.org/topic/biomedical/)[Climate Tech](https://spectrum.ieee.org/topic/climate-tech/)[Computing](https://spectrum.ieee.org/topic/computing/)[Consumer Electronics](https://spectrum.ieee.org/topic/consumer-electronics/)[Energy](https://spectrum.ieee.org/topic/energy/)[History of Technology](https://spectrum.ieee.org/topic/tech-history/)[Robotics](https://spectrum.ieee.org/topic/robotics/)[Semiconductors](https://spectrum.ieee.org/topic/semiconductors/)[Telecommunications](https://spectrum.ieee.org/topic/telecommunications/)[Transportation](https://spectrum.ieee.org/topic/transportation/) [IEEE Spectrum](https://spectrum.ieee.org/) FOR THE TECHNOLOGY INSIDER ### Topics [Aerospace](https://spectrum.ieee.org/topic/aerospace/)[AI](https://spectrum.ieee.org/topic/artificial-intelligence/)[Biomedical](https://spectrum.ieee.org/topic/biomedical/)[Climate Tech](https://spectrum.ieee.org/topic/climate-tech/)[Computing](https://spectrum.ieee.org/topic/computing/)[Consumer Electronics](https://spectrum.ieee.org/topic/consumer-electronics/)[Energy](https://spectrum.ieee.org/topic/energy/)[History of Technology](https://spectrum.ieee.org/topic/tech-history/)[Robotics](https://spectrum.ieee.org/topic/robotics/)[Semiconductors](https://spectrum.ieee.org/topic/semiconductors/)[Telecommunications](https://spectrum.ieee.org/topic/telecommunications/)[Transportation](https://spectrum.ieee.org/topic/transportation/) ### Sections [Features](https://spectrum.ieee.org/type/feature/)[News](https://spectrum.ieee.org/type/news/)[Opinion](https://spectrum.ieee.org/type/opinion/)[Careers](https://spectrum.ieee.org/topic/careers/)[DIY](https://spectrum.ieee.org/topic/diy/)[Engineering Resources](https://spectrum.ieee.org/engineering-resources/) ### More [Newsletters](https://spectrum.ieee.org/newsletters/)[Special Reports](https://spectrum.ieee.org/special-reports/)[Collections](https://spectrum.ieee.org/collections/)[Explainers](https://spectrum.ieee.org/type/explainer/)[Top Programming Languages](https://spectrum.ieee.org/top-programming-languages)[Robots Guide ↗](https://robotsguide.com/)[IEEE Job Site ↗](https://jobs.ieee.org/) ### For IEEE Members [Current Issue](https://spectrum.ieee.org/magazine/current-issue)[Magazine Archive](https://spectrum.ieee.org/magazine/)[The Institute](https://spectrum.ieee.org/the-institute/)[The Institute Archive](https://spectrum.ieee.org/the-institute/ti-archive/) ### For IEEE Members [Current Issue](https://spectrum.ieee.org/magazine/current-issue)[Magazine Archive](https://spectrum.ieee.org/magazine/)[The Institute](https://spectrum.ieee.org/the-institute/)[The Institute Archive](https://spectrum.ieee.org/the-institute/ti-archive/) ### IEEE Spectrum [About Us](https://spectrum.ieee.org/about)[Contact Us](https://spectrum.ieee.org/contact)[Reprints & Permissions ↗](https://www.parsintl.com/publications/ieee-media/)[Advertising ↗](https://advertise.ieee.org/) ### Follow IEEE Spectrum [](https://twitter.com/ieeespectrum)[](https://www.facebook.com/IEEE.Spectrum)[](https://www.instagram.com/ieeespectrum/)[](https://www.threads.net/@ieeespectrum)[](https://www.linkedin.com/company/ieee-spectrum/)[](https://www.youtube.com/c/ieeespectrum)[](https://www.tiktok.com/@ieeespectrum)[](https://spectrum.ieee.org/customfeeds/feed/all-topics/rss) ### Support IEEE Spectrum _IEEE Spectrum_ is the flagship publication of the IEEE — the world’s largest professional organization devoted to engineering and applied sciences. Our articles, videos, and infographics inform our readers about developments in technology, engineering, and science. [Subscribe](https://ieee.dragonforms.com/spectrum_subscribe?utm_source=ieee-spectrum&utm_medium=nav&utm_content=hugemenu) [About IEEE](https://www.ieee.org/about/)[Contact & Support](https://www.ieee.org/about/contact.html)[Accessibility](https://www.ieee.org/accessibility-statement.html)[Nondiscrimination Policy](https://www.ieee.org/about/corporate/governance/p9-26.html)[Terms](https://www.ieee.org/about/help/site-terms-conditions.html)[IEEE Privacy Policy](https://www.ieee.org/security-privacy.html)[Cookie Preferences](https://spectrum.ieee.org/from-tv-repairman-to-electromagnetic-compatibility-expert#)[Ad Privacy Options](https://spectrum.ieee.org/st/ppid-info) ? Copyright 2026 IEEE — All rights reserved. A public charity, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity. ## Enjoy more free content and benefits by creating an account ## Saving articles to read later requires an IEEE Spectrum account ## The Institute content is only available for members ## Downloading full PDF issues is exclusive for IEEE Members ## Downloading this e-book is exclusive for IEEE Members ## Access to _Spectrum_ 's Digital Edition is exclusive for IEEE Members ## Following topics is a feature exclusive for IEEE Members ## Adding your response to an article requires an IEEE Spectrum account ## Create an account to access more content and features on _IEEE Spectrum_ , including the ability to save articles to read later, download Spectrum Collections, and participate in conversations with readers and editors. For more exclusive content and features, consider [Joining IEEE](https://spectrum.ieee.org/join) . ## Join the world’s largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum’s articles, archives, PDF downloads, and other benefits. [Learn more about IEEE →](https://spectrum.ieee.org/join) ## Join the world’s largest professional organization devoted to engineering and applied sciences and get access to this e-book plus all of _IEEE Spectrum’s_ articles, archives, PDF downloads, and other benefits. [Learn more about IEEE →](https://spectrum.ieee.org/join) [CREATE AN ACCOUNT](https://www.ieee.org/profile/public/createwebaccount/showCreateAccount.html?ShowMGAMarkeatbilityOptIn=true&sourceCode=spectrum&signinurl=https%3A%2F%2Fspectrum.ieee.org%2Fcore%2Fsaml%2Fmain%2Flogin%3Fnext_url%3Dhttps%3A%2F%2Fspectrum.ieee.org%2Fcore%2Fintegrations%2Fieee%2Fchanges%0A&url=https://spectrum.ieee.org/&autoSignin=Y&car=IEEE-Spectrum)[SIGN IN](https://spectrum.ieee.org/core/saml/main/login?next_url=https://spectrum.ieee.org/core/integrations/ieee/changes) [JOIN IEEE](https://www.ieee.org/membership-application/public/join.html?promo=JOINLITE&style=SPECTRUM&joinlite=TRUE)[SIGN IN](https://spectrum.ieee.org/core/saml/main/login?next_url=https://spectrum.ieee.org/core/integrations/ieee/changes) [Close](javascript:;) ## Access Thousands of Articles — Completely Free ## Create an account and get exclusive content and features: **Save articles, download collections,** and **post comments** — all free! For full access and benefits, [subscribe](https://ieee.dragonforms.com/spectrum_subscribe) to _Spectrum_. [CREATE AN ACCOUNT](https://www.ieee.org/profile/public/createwebaccount/showCreateAccount.html?ShowMGAMarkeatbilityOptIn=true&sourceCode=spectrum3c&signinurl=https%3A%2F%2Fspectrum.ieee.org%2Fcore%2Fsaml%2Fmain%2Flogin%3Fnext_url%3Dhttps%3A%2F%2Fspectrum.ieee.org%2Fcore%2Fintegrations%2Fieee%2Fchanges%0A&url=https://spectrum.ieee.org/&autoSignin=Y&car=IEEE-Spectrum)[SIGN IN](https://spectrum.ieee.org/core/saml/main/login?next_url=https://spectrum.ieee.org/core/integrations/ieee/changes) [The Institute](https://spectrum.ieee.org/topic/the-institute/)[IEEE Member News](https://spectrum.ieee.org/the-institute/collections/ieee-member-news/)[Careers](https://spectrum.ieee.org/topic/careers/)[Profile](https://spectrum.ieee.org/type/profile/) # From TV Repairman to Electromagnetic Compatibility Expert ## David Weston achieved tech career heights with only a certificate [Kathy Pretz](https://spectrum.ieee.org/u/kathy-pretz) 05 Mar 2026 7 min read [](javascript:;) 4 Kathy Pretz is the editor in chief of The Institute, IEEE's member publication Electromagnetic compatibility expert IEEE Life Member David Weston, now retired, ran a successful consulting company for more than 40 years. Michael Venables [](mailto:?subject=From%20TV%20Repairman%20to%20Electromagnetic%20Compatibility%20Expert&body=https://spectrum.ieee.org/from-tv-repairman-to-electromagnetic-compatibility-expert) [IEEE member news](https://spectrum.ieee.org/tag/ieee-member-news)[type:ti](https://spectrum.ieee.org/tag/type-ti)[careers](https://spectrum.ieee.org/tag/careers)[emi](https://spectrum.ieee.org/tag/emi)[emc](https://spectrum.ieee.org/tag/emc)[electromagnetic compatibility](https://spectrum.ieee.org/tag/electromagnetic-compatibility) No one had very high career aspirations for teenager [David A. Weston](https://ieeexplore.ieee.org/author/38227767000)—except for Weston himself. Growing up in London, he scored low on the U.K. national assessment test given to students finishing primary school. The result meant that his next path was either to become a laborer or attend a vocational school to learn a trade. What Weston really wanted to do was to work as a radio and TV repairman. He was fascinated by how the devices worked. He had taught himself to build an AM radio when he was 15. Even after showing it to his parents and teachers, though, they still didn’t think he was smart enough to pursue his chosen career, he says. ### David A.Weston **Employer** EMC Consulting, in in Arnprior, Ont., Canada **Job title** Retired consultant **Member grade** Life member **Alma mater** Croydon Technical College, London So, later that year, the underweight teen got a job on a construction site carrying heavy loads of building materials in a hod, a three-sided wooden trough. The experience convinced him he wasn’t cut out for manual labor. He eventually earned a certificate in radio and television, the only credential he holds. The lack of academic degrees did not hold him back, though. He went on to become an expert in [electromagnetic interference](https://spectrum.ieee.org/tag/electromagnetic-interference) (EMI) and electromagnetic compatibility (EMC). An EMI field has unwanted energy that causes interference. EMC is the capacity for electronic devices to work correctly in a shared electromagnetic environment without causing interference or suffering from it in nearby devices or signals. After working for a number of companies, he launched his own business more than 40 years ago: [EMC Consulting](https://www.emcconsultinginc.com/), in Arnprior, Ont., Canada. The company has helped clients meet EMI and EMC regulatory requirements. Now 83 years old and retired, the IEEE life member recently self-published his memoir, [_From a Hod to an Odd EM Wave_](https://www.barnesandnoble.com/w/from-a-hod-to-an-odd-em-wave-david-weston/1148995654). “My memoir is about engineering persistence and human and technical discoveries,” he says. “I wanted to interest a young person, or perhaps a person later in life, in a career in engineering. If I can show that engineering is a personal, human endeavor with exciting opportunities in different fields such as medical, scientific, and the arts, maybe more women would be attracted to it.” ## From repairing radios to designing underwater devices In 1960 Weston enrolled in the radio and electronics program at London’s Croydon Technical College (now [Croydon College](https://croydon.ac.uk/)). The school covered topics from the [City and Guilds of London Institute](https://www.cityandguilds.com/)’s radio and television certificate program. He attended classes one day a week for five years while working to put himself through school. Although his parents and his teachers might not have recognized Weston’s potential, employers did. He got his first job in 1960, fixing televisions in a small repair shop. Then he helped repair tape recorders. In his spare time, he studied [transistors](https://spectrum.ieee.org/tag/transistors) and [semiconductors](https://spectrum.ieee.org/topic/semiconductors/). Everything he knows, he says, he learned by reading books and research papers, and from on-the-job training. Later in 1960, he worked as a mechanical examiner for the U.K. [Ministry of Aviation](https://en.wikipedia.org/wiki/Ministry_of_Aviation), where he calibrated precision meters and potentiometers, which are variable resistors that monitor, control, and measure industrial equipment. “Engineering is creative. To have a new idea or design accepted is rewarding, satisfying, pleasurable, and even exciting.” He left the ministry in 1963 because he found the work boring, he says, and he was hired as a technician with the [Medical Research Council](https://www.gov.uk/government/organisations/medical-research-council/about)’s neuropsychiatric research unit in Carshalton. The institution researches the biological causes of mental illness. His manager was interested in learning about advances in medical electronics and eagerly shared his knowledge with Weston. One of Weston’s tasks was to build an [electroencephalography](https://spectrum.ieee.org/tag/electroencephalography) ([EEG](https://spectrum.ieee.org/tag/eeg)) calibrator to measure responses from a patient’s brain activity. The methods used at the time to detect a brain tumor—before [MRI machines](https://spectrum.ieee.org/mri-pioneer-to-receive-ieee-medal-for-innovations-in-healthcare-technology) were developed—involved monitoring the patient’s speech and coordination, followed by taking a biopsy, which was not without danger, he says. He used an ultrasonic transmitter and receiver to measure the time of transmission to the midline in the brain to determine whether the person had a tumor. If the midline had shifted, it would indicate the presence of a tumor, and a biopsy would be performed to confirm it. The measure of the evoked response in the brain was the only reliable indicator. Weston earned his radio and TV certificate in 1965, leaving the research facility a year later to join Divcon (now part of [Oceaneering International](https://www.oceaneering.com/)), a commercial diving company based in London that developed deep-sea helium diving helmets. Weston helped design a waterproof handheld communication device for divers that could withstand the high pressure in diving bells, the open-bottom pressurized chambers that transported them underwater. Weston then moved to Hamburg, [Germany](https://spectrum.ieee.org/tag/germany), in 1969 to work for [Plath](https://plath-corporation.de/en/), an electronics manufacturer. He was tasked, along with other engineers from England, to design a servo control loop. “Unfortunately it oscillated so badly when first being turned on that it shook itself to bits,” he says. He left to work as a senior engineer at [Dr. Staiger Mohilo and Co.](https://www.kistler.com/INT/en/kistler-acquisitions-win-win-for-all-concerned/C00000494) (now part of [Kistler](https://www.kistler.com/INT/en/)), in Schorndorf, Germany. It manufactured torque sensors, force transducers, and specialized test stand systems. Weston designed a process control computer. He says his boss told him that the controller had to work in close proximity to—and from the same power source as—a nearby machine without interfering with it or being interfered by it. “I was thus introduced to the idea of electromagnetic compatibility,” he says. After three years, he left to join the [Siemens Mobility](https://www.mobility.siemens.com/global/en.html) train group in Braunschweig, Germany, where he helped develop an electronic train-crossing light controller. The original warning lights on crossing gates used a [mercury](https://spectrum.ieee.org/tag/mercury) tube as a switch. “The concern was the danger to personnel if the tube broke,” he says. “The simple and inexpensive solution was to put the tube in a metal container.” Weston and his wife decided to leave Germany for Canada in 1975, after their young son began forgetting how to speak English. ## Working on the [space shuttle](https://spectrum.ieee.org/tag/space-shuttle) and a [particle accelerator](https://spectrum.ieee.org/tag/particle-accelerator) His first job in the country was as an engineer for [Canadian Aviation Electronics](https://www.cae.com/) in Montreal. CAE helped design the remote manipulator system in robotic hand controllers and simulation systems used to train astronauts for the space shuttle. The [robotic arm](https://spectrum.ieee.org/tag/robotic-arm), known as [Canadarm](https://en.wikipedia.org/wiki/Canadarm), was used to deploy, maneuver, and capture payloads for the astronauts. Weston’s engineering team designed the display and control panel as well as the hand controllers located in the shuttle’s flight deck. “I was attracted to the EMC aspects of the project and avidly studied everything I could on the topic,” he says. He also helped develop a system that would protect an aircraft’s deployable black box from lightning strikes. “I used a computer program to analyze the EMI field at close proximity to the black box to predict the lightning current flowing into the aircraft structure,” he says. While enjoying the warm winter weather during a 1975 visit to a supplier on Long Island, N.Y., he decided he wanted to move his family there and asked whether any companies in the area were hiring. He was told that [Brookhaven National Laboratory](https://www.bnl.gov/world/), in Upton, was, so he applied for a position working on the ring system for the [Isabelle proton colliding-beam particle accelerator](https://en.wikipedia.org/wiki/ISABELLE). The project, later known as the [colliding beam accelerator](https://spectrum.ieee.org/supercolliders), was a collaboration between the lab and the [U.S. Department of Energy](https://www.energy.gov/). The 200+200 giga-electron volt proton-proton collider was designed to use advanced superconducting magnets cooled by a massive helium [refrigeration](https://spectrum.ieee.org/tag/refrigeration) system to produce high-energy collisions. The GeV refers to the collision energy in a particle accelerator. ### Weston’s Advice for Budding Engineers * Follow the field in which you are most interested. * Don’t be afraid to work in other countries; it can be a rewarding, enriching experience. * Question the results of measurements or analyses. If it doesn’t seem right, it probably isn’t. Look at a similar publication on the same topic for a good correlation. * Don’t be too shy to ask simple questions. That’s how we learn and grow. * Keep an open mind. The lab hired him in 1978, and the family moved to Long Island. After a few weeks of meeting with different departments, his boss asked him what kind of work he wanted to do. Weston told him about his idea for designing a device to detect a helium leak, should there ever be one. His machine would cover the entire 3,834-meter circumference area of the ring. “The danger with increased helium-enriched air is that the oxygen level reduces until the person breathing becomes adversely affected,” he wrote in his memoir. “I found that the speed of the sound of helium increased enough to be detected, but not sufficient enough to cause a person trouble if they were in the tunnel. “Brookhaven was considering machines that only covered a small area of the ring, but these would be unrealistic because too many machines would be needed, and the cost would have been astronomical.” Weston’s system included an [ultrasonic](https://spectrum.ieee.org/tag/ultrasonic) transmitter, a receiver, a power amplifier, and a preamplifier. It would sound an alarm if the helium content went above a certain level. People in the tunnel would be directed to go to the nearest oxygen-breathing equipment, put on a mask, and immediately evacuate. It was successfully tested. Weston wrote a report detailing the ultrasonic helium leak detector, but shortly after, he and his wife had to return to [Canada](https://spectrum.ieee.org/tag/canada) in 1978 because they were unable to get additional work permits in the [United States](https://spectrum.ieee.org/tag/united-states). When he returned to Brookhaven for a visit, his former boss told him the report was well-received. And he shared some news that upset Weston. “My boss told me he took my report, changed the name on the report to his, did not mention me, and published the report as his,” Weston wrote in his memoir. But the system was never built. The Isabelle project was canceled in July 1983 due to technical problems with fabricating the superconducting magnets. Weston got a job working for [CAL Corp.](https://satistar.com/portfolio/cal-corporation/), an aerospace telecommunications company in Montreal. For the next 14 years, he fixed EMI problems for the company’s products, including its charge-coupled device-based space-qualified cameras, which were designed to be carried aboard a satellite. In 1992 he realized that nearly all his work involved consulting for the company’s customers, so he decided to start his own agency. CAL generously let him take the clients he worked with, he says. Weston then conducted EMI analysis and testing and designed EMC systems for companies around the world. “I always had enough customers and have never had to look for work,” he says. “For me, having my own business was more secure than working for a company.” He retired in 2022. ## IEEE as an educator To broaden his education, he joined IEEE in 1976 to get access to its research papers and attend its conferences, he says. He is a member of the [IEEE Electromagnetic Compatibility Society](https://www.emcs.org/). Because he is self-educated, he was “keen to learn as much as possible by reading practical papers published by IEEE,” he says. “I met people at IEEE symposiums and listened to the authors presenting their papers.” Those included EMC experts such as Life Fellows [Lothar O. “Bud” Hoeft](https://ieeexplore.ieee.org/search/searchresult.jsp?newsearch=true&queryText=L.O.%20Hoeft), [Richard J. Mohr](https://ieeexplore.ieee.org/search/searchresult.jsp?newsearch=true&searchWithin=%22First%20Name%22:Richard&searchWithin=%22Last%20Name%22:Mohr), and [Clayton R. Paul](https://ieeexplore.ieee.org/search/searchresult.jsp?newsearch=true&queryText=Clayton%20R.%20Paul), whose papers are published in the[IEEE Xplore Digital Library](https://ieeexplore.ieee.org/Xplore/home.jsp). [Several of Weston’s papers](https://ieeexplore.ieee.org/search/searchresult.jsp?queryText=david%20weston&highlight=true&returnFacets=ALL&returnType=SEARCH&matchPubs=true&refinements=Author:David%20Weston) are in the library as well. His book [_\_Electromagnetic Compatibility: Methods, Analysis, Circuits, and Measurement\__](https://www.emcconsultinginc.com/publications/) references many IEEE papers on data and analysis methods. “Engineering is creative,” he says. “To have a new idea or design accepted is rewarding, satisfying, pleasurable, and even exciting.” From Your Site Articles * [Rediscovering the Lost Legacy of Chemist Jan Czochralski ?](https://spectrum.ieee.org/legacy-chemist-jan-czochralski) * [This Physics Professor Credits Collaboration for Her Success ?](https://spectrum.ieee.org/cinzia-davia-physics) * [Xiangyi Cheng Is Bringing AR to Classrooms and Hospitals ?](https://spectrum.ieee.org/xiangyi-cheng-ar-classrooms-hospitals) Related Articles Around the Web * [New eBook: From a Hod to an Odd EM Wave | Interference Technology ?](https://interferencetechnology.com/ebook-from-a-hod-to-an-odd-em-wave/) [IEEE member news](https://spectrum.ieee.org/tag/ieee-member-news)[type:ti](https://spectrum.ieee.org/tag/type-ti)[careers](https://spectrum.ieee.org/tag/careers)[emi](https://spectrum.ieee.org/tag/emi)[emc](https://spectrum.ieee.org/tag/emc)[electromagnetic compatibility](https://spectrum.ieee.org/tag/electromagnetic-compatibility) [Kathy Pretz](https://spectrum.ieee.org/u/kathy-pretz) is editor in chief for _The Institute_, which covers all aspects of IEEE, its members, and the technology they're involved in. [See full bio→](https://spectrum.ieee.org/u/kathy-pretz) The Conversation (3) Publish Sort by Bruce MacDougall 17 Mar, 2026 M A wonderful story witnessing an amazing career. Thank-you and best wishes to Mr. Weston. [](javascript:; "reply")[](javascript:; "share") [](https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fspectrum.ieee.org%2Ffrom-tv-repairman-to-electromagnetic-compatibility-expert%3Fmessageid%3Dpost%3A2675549346%3Acomments%3A69b95693bd772869a6ec6f11%23comment-69b95693bd772869a6ec6f11%3Fxrs%3DRebelMouse_fb%26ts%3D "Share to Facebook")[](https://twitter.com/intent/tweet?url=https%3A%2F%2Fspectrum.ieee.org%2Ffrom-tv-repairman-to-electromagnetic-compatibility-expert%3Fmessageid%3Dpost%3A2675549346%3Acomments%3A69b95693bd772869a6ec6f11%23comment-69b95693bd772869a6ec6f11 "Share to Twitter")[](https://spectrum.ieee.org/from-tv-repairman-to-electromagnetic-compatibility-expert?messageid=post:2675549346:comments:69b95693bd772869a6ec6f11#comment-69b95693bd772869a6ec6f11 "Open permalink")[](javascript: void(0) "Copy permalink") copied to clipboard [](javascript:; "like")[](javascript:; "dislike") [0 Replies](javascript:;)[Hide replies](javascript:;) Show More Replies Purnendu Kumar 20 Mar, 2026 M It's very inspiring to read the journey of Mr. Weston. Just sad that such paths are now almost closed, every job comes with a fixed degree requirements. [](javascript:; "reply")[](javascript:; "share") [](https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fspectrum.ieee.org%2Ffrom-tv-repairman-to-electromagnetic-compatibility-expert%3Fmessageid%3Dpost%3A2675549346%3Acomments%3A69bc9dc65ae81980c94501f8%23comment-69bc9dc65ae81980c94501f8%3Fxrs%3DRebelMouse_fb%26ts%3D "Share to Facebook")[](https://twitter.com/intent/tweet?url=https%3A%2F%2Fspectrum.ieee.org%2Ffrom-tv-repairman-to-electromagnetic-compatibility-expert%3Fmessageid%3Dpost%3A2675549346%3Acomments%3A69bc9dc65ae81980c94501f8%23comment-69bc9dc65ae81980c94501f8 "Share to Twitter")[](https://spectrum.ieee.org/from-tv-repairman-to-electromagnetic-compatibility-expert?messageid=post:2675549346:comments:69bc9dc65ae81980c94501f8#comment-69bc9dc65ae81980c94501f8 "Open permalink")[](javascript: void(0) "Copy permalink") copied to clipboard [](javascript:; "like")[](javascript:; "dislike") [0 Replies](javascript:;)[Hide replies](javascript:;) Show More Replies Douglas Manchester 18 Mar, 2026 LS Mr. Weston’s career is a wonderful demonstration that concurrent development of practical and theoretical knowledge can make for an engaging and successful life as an engineer. While the “degree first” route is typical, and in some instances mandatory, Mr. Weston’s path tells us there are exceptions to the rule. Well done, Mr. Weston. [](javascript:; "reply")[](javascript:; "share") [](https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fspectrum.ieee.org%2Ffrom-tv-repairman-to-electromagnetic-compatibility-expert%3Fmessageid%3Dpost%3A2675549346%3Acomments%3A69ba3b17e843f475ed9e38f0%23comment-69ba3b17e843f475ed9e38f0%3Fxrs%3DRebelMouse_fb%26ts%3D "Share to Facebook")[](https://twitter.com/intent/tweet?url=https%3A%2F%2Fspectrum.ieee.org%2Ffrom-tv-repairman-to-electromagnetic-compatibility-expert%3Fmessageid%3Dpost%3A2675549346%3Acomments%3A69ba3b17e843f475ed9e38f0%23comment-69ba3b17e843f475ed9e38f0 "Share to Twitter")[](https://spectrum.ieee.org/from-tv-repairman-to-electromagnetic-compatibility-expert?messageid=post:2675549346:comments:69ba3b17e843f475ed9e38f0#comment-69ba3b17e843f475ed9e38f0 "Open permalink")[](javascript: void(0) "Copy permalink") copied to clipboard [](javascript:; "like")[](javascript:; "dislike") [0 Replies](javascript:;)[Hide replies](javascript:;) Show More Replies [](https://spectrum.ieee.org/social-media-bans) [Telecommunications](https://spectrum.ieee.org/topic/telecommunications/)[Guest Article](https://spectrum.ieee.org/type/guest-article/) ## [Social Media Bans Aren’t Enough to Make Children Safe](https://spectrum.ieee.org/social-media-bans) 13 hours ago 5 min read [](javascript:;) [](https://spectrum.ieee.org/fanless-liquid-cooled-ai-servers-coolit) [AI](https://spectrum.ieee.org/topic/artificial-intelligence/)[Energy](https://spectrum.ieee.org/topic/energy/)[Sponsored Article](https://spectrum.ieee.org/type/sponsored/)[Computing](https://spectrum.ieee.org/topic/computing/) ## [The Future Is Fanless: 100% Heat Capture for Liquid Cooled AI Servers](https://spectrum.ieee.org/fanless-liquid-cooled-ai-servers-coolit) 22 Sep 2026 3 min read [](javascript:;) 1 [](https://spectrum.ieee.org/rivian-self-driving) [Transportation](https://spectrum.ieee.org/topic/transportation/)[Magazine](https://spectrum.ieee.org/magazine/)[Feature](https://spectrum.ieee.org/type/feature/) ## [Rivian’s Gambit for Full Autonomy](https://spectrum.ieee.org/rivian-self-driving) 08 Sep 2026 13 min read [](javascript:;) 4 ## More from The Institute [The Institute](https://spectrum.ieee.org/topic/the-institute/)[Robotics](https://spectrum.ieee.org/topic/robotics/)[Article](https://spectrum.ieee.org/type/article/) ## [Mexican EPICS in IEEE Team Builds Portable Educational Platform](https://spectrum.ieee.org/epics-in-ieee-portable-educational) [The Institute](https://spectrum.ieee.org/topic/the-institute/)[Aerospace](https://spectrum.ieee.org/topic/aerospace/)[IEEE Member News](https://spectrum.ieee.org/the-institute/collections/ieee-member-news/)[Careers](https://spectrum.ieee.org/topic/careers/)[Profile](https://spectrum.ieee.org/type/profile/) ## [Spain’s First Astronaut, Pedro Duque, Named IEEE Honorary Member](https://spectrum.ieee.org/spain-astronaut-ieee-honorary-member) [The Institute](https://spectrum.ieee.org/topic/the-institute/)[Careers](https://spectrum.ieee.org/topic/careers/)[Article](https://spectrum.ieee.org/type/article/) ## [Turning Tech Talent Into Leadership Legacy](https://spectrum.ieee.org/tech-talent-into-leadership-legacy) [The Institute](https://spectrum.ieee.org/topic/the-institute/)[IEEE Member News](https://spectrum.ieee.org/the-institute/collections/ieee-member-news/)[World Standards Day](https://spectrum.ieee.org/collections/world-standards-day/)[Spectrum Collections](https://spectrum.ieee.org/collections/)[Opinion](https://spectrum.ieee.org/type/opinion/) ## [Countries Seek to Curb Social Media Addiction for Kids](https://spectrum.ieee.org/countries-seek-to-curb-social-media-addiction-for-kids) [The Institute](https://spectrum.ieee.org/topic/the-institute/)[News](https://spectrum.ieee.org/type/news/) ## [IEEE to Reward Sections for High Voter Turnout in Annual Election](https://spectrum.ieee.org/ieee-to-reward-sections) [Computing](https://spectrum.ieee.org/topic/computing/)[Magazine](https://spectrum.ieee.org/magazine/)[Feature](https://spectrum.ieee.org/type/feature/) ## [The AI Inference Revolution Is Here](https://spectrum.ieee.org/inference-hardware-revolution) ### Today’s tidal wave of queries is forcing hardware makers to pivot [Matthew S. Smith](https://spectrum.ieee.org/u/matthew_s_smith) 15 Sep 2026 12 min read [](javascript:;) 14 [](mailto:?subject=The%20AI%20Inference%20Revolution%20Is%20Here&body=https://spectrum.ieee.org/inference-hardware-revolution) Vertical [](https://spectrum.ieee.org/inference-hardware-revolution) DarkBlue1 **Since about 2020,**AI has largely focused on training bigger and better models. Large language models (LLMs) ballooned from millions of parameters to trillions. This proved effective: The largest version of OpenAI’s GPT-3, released in 2020, correctly [answered](https://arxiv.org/pdf/2009.03300) just 43.9 percent of questions on a popular knowledge-and-reasoning benchmark. Just four years later, GPT-4o [reached](https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/) a score of 88.7 percent on the same exam, effectively matching those of human experts. Advanced AI labs are still training ever larger models, but that training has somewhat receded to the background of the AI conversation. In 2026, inference—the use of trained models to produce code, write essays, or make images of ourselves as elves—has come to the forefront. “It’s like training is yesterday’s news,” says [Matt Kimball](https://moorinsightsstrategy.com/team/matt-kimball/), principal data-center analyst at Moor Insights & Strategy. “All that any chief information officer wants to talk about is inference.” Nvidia CEO Jensen Huang, speaking at the company’s GTC 2026 conference, touted this change as the “[inflection point of inference](https://www.youtube.com/watch?v=jw_o0xr8MWU).” Part of what’s caused the shift is very simple: LLMs are becoming useful, so people are using them. On top of that, many models on the market today are reasoning models. In response to a user’s query, they run inference not just once but multiple times, reprompting themselves in a process called[_\_chain of thought\__](https://arxiv.org/pdf/2201.11903). Reasoning models generate longer outputs, and models with high reasoning effort can produce up to [20 times](https://www.linkedin.com/posts/artificial-analysis_how-many-tokens-do-reasoning-models-use-vs-activity-7318302119206289408-3mt1/) as much text as those with low or no effort. Adding even more to the world’s inference workload, the rise of [agentic AI](https://spectrum.ieee.org/ai-agents) has resulted in inference running not just as a real-time response to a user’s query but also around the clock, working autonomously toward a user-defined goal. Amazon’s Trainium chip was originally designed for AI training. However, Amazon Web Services chose to break up AI inference into two parts, with Trainium running the more computationally complex portion and Cerebras’s wafer-scale engine taking on the more memory-intensive portion.Amazon The resulting explosion in inference demand has led to unexpected alliances among tech giants. [OpenAI](https://openai.com/index/cerebras-partnership/) and [Amazon](https://www.reuters.com/business/retail-consumer/cerebras-systems-amazon-strike-deal-offer-cerebras-ai-chips-amazons-cloud-2026-03-13/) have deployed chips the size of a [dinner plate](https://spectrum.ieee.org/cerebrass-giant-chip-will-smash-deep-learnings-speed-barrier) designed by [Cerebras](https://www.cerebras.ai/), despite Amazon having its own [Trainium](https://aws.amazon.com/ai/machine-learning/trainium/) chips. Nvidia [bought](https://www.cnbc.com/2025/12/24/nvidia-buying-ai-chip-startup-groq-for-about-20-billion-biggest-deal.html) key talent and intellectual property from AI-inference startup [Groq](https://groq.com/) in a controversial deal worth US $20 billion. And [Anthropic](https://www.anthropic.com/) is [paying](https://x.ai/news/anthropic-compute-partnership) LLM competitor[SpaceXAI](https://x.ai/) over a billion dollars per month to lease spare compute. Although they might seem similar, AI training and AI inference are computationally different. These big moves from tech giants signal that in order to support the inference demand, we’re going to need a very different mix of hardware than experts may have expected even a couple of years ago. ## How does AI inference differ from AI training? An untrained LLM is like a jumble of Scrabble tiles on a table. Instead of single letters, though, the tiles show fragments of words, called tokens. Everything you’d need to write almost anything is present, but nothing makes sense. Training a model organizes this jumble using a guessing game played at scale. The model is shown real text with the next token hidden and asked to predict what comes next. After each guess, the correct token is revealed and then compared to the prediction, and the difference is used to calculate the model’s accuracy. The game is played not with a single sentence but over billions of passages. While a real game of Scrabble can be played over a bag of chips and a few drinks, AI training is computationally intense. The model updates its parameters through [backpropagation](https://spectrum.ieee.org/what-is-deep-learning/backpropagation), a process that repeatedly calculates how each of a model’s billions or trillions of parameters should shift to make the next prediction better. This is why tech giants are [building](https://spectrum.ieee.org/5gw-data-center) larger data centers than ever before. Eventually the model’s creator decides further training isn’t worth the cost, and the guessing game stops. Backpropagation ends, the parameters are frozen, and the LLM becomes a pretrained model. Fine-tuning—a short training run on smaller, more specialized data—adds final tweaks, and the model is deployed. Nvidia’s Groq 3 language-processing unit minimizes data movement by placing on-chip SRAM memory and computational blocks in the order they are needed on-chip. Nvidia Next comes inference. This is the process of using the deployed model, which, now that it’s been trained, has learned to spit out Scrabble tiles—tokens—in a sensible order. You might think that AI inference is less computationally demanding because the backpropagation calculations used to update parameters are eliminated. But [Sudeep Bhoja](https://www.linkedin.com/in/sudeep-bhoja-070a111/), founder and CTO of the inference-hardware company [d-Matrix](https://www.d-matrix.ai/), explains that inference adds new challenges. The models are “autoregressive” in nature. That is, the next output depends on the previous one. “So to generate the next token, you have to read all of the weights and all of the [context] from the previous token,” explains Bhoja. The context includes all of your prompts, all of the LLM’s replies, and all of the files you upload. It’s a lot of data and a lot of processing. An LLM generates its reply in two phases: prefill and decode. Prefill is the model reading a prompt. It processes every token at once, computing how each token relates to all the others. This operation is called [attention](https://en.wikipedia.org/wiki/Attention_(machine_learning)), and it’s a defining characteristic of the transformer architecture behind modern LLMs. It allows them to respond to a word in its sentence, paragraph, and larger context rather than on its own. Think of it like arranging Scrabble tiles before you place them in a game. Many players move tiles around to imagine how they connect. Self-attention plays a similar role, though instead of moving physical tiles, each token sends a query to the others and receives a score indicating the token’s relevance. These queries result in two types of vectors: the keys and values. They are typically placed in a store called the KV cache. This isn’t strictly required, as a model could instead recompute these vectors with each new token it generates. But nearly all LLMs use a KV cache to reduce how much computing they do. The KV cache is stored in memory and becomes a scratchpad to which the LLM can return to understand a conversation, and though it starts small, it can swell to dozens of gigabytes. Prefill is a problem that can be easily divided up and worked on in parallel. This is why GPUs became the dominant AI accelerator as LLMs surged in popularity. Graphics rasterization (computing the color of every pixel on a screen) is also massively parallel, so GPU architectures were a natural fit. Cerebras’s wafer-scale engine chips maximize memory bandwidth by keeping everything—both memory and computational units—side by side on the dinner-plate-size chips. Cerebras Next comes decode. Here, the model generates its reply one token at a time. At each step it takes the most recent token, weighs it against everything in the KV cache, uses that information to predict the next token, and adds the new token’s key and value to the cache. Then it repeats in sequence, token by token. This is where the autoregressive nature of the model works against inference speed. Predicting each token requires reading the entire model from memory, and that model consists of possibly tens to hundreds of gigabytes of parameters (the numbers representing what the model learned in training). Crucially, this is in addition to the memory required to store the KV cache. As a result, the movement of all this data through memory often requires more bandwidth than inference hardware has available. So at least some of the computing parts of a GPU sit idle as it waits for data. Researchers [found](https://arxiv.org/pdf/2503.08311) that Nvidia H100 GPUs running open-source LLMs sit idle 50 to 80 percent of the time. ## Memory’s role in inferencing [Shahriar “Sha” Rabii](https://www.linkedin.com/in/rabii/), former head of silicon engineering at Meta and cofounder of the AI startup [Majestic Labs](https://majestic-labs.ai/), says idled processors are why many companies that are trying to improve AI-inference performance are laser-focused on memory. “With the GPU-based approach, you end up greatly over-provisioning compute and starved on memory. That’s driving the big [memory] scale out,” he says. Bhoja’s d-Matrix and Rabii’s Majestic Labs both focus on this memory bottleneck. However, their companies imagine different solutions. d-Matrix’s second-generation AI accelerator, [Raptor](https://www.d-matrix.ai/announcements/d-matrix-and-alchip-announce-collaboration-on-worlds-first-3d-dram-solution-to-supercharge-ai-inference/), aims to improve inference performance by minimizing the distance between compute and memory. The GPUs in most current AI-inference deployments do this by placing high-bandwidth memory (HBM) around the perimeter of the GPU. Each HBM is a stack of DRAM dies linked together and connected to a superfast interface to the GPU. This is great for training, but for inference, the amount of memory you can stack this way and the bandwidth it can provide leave something to be desired. ### d-Matrix’s stacked-die architecture Memory bandwidth—how quickly data can be read from memory to logic—is a major bottleneck in AI inference. The startup d-Matrix is increasing that bandwidth by stacking the logic die directly on top of the memory, in this case DRAM. This allows for lots of extremely short interconnects. Chris Philpot d-Matrix’s Raptor removes that bottleneck by stacking an AI accelerator on a DRAM die. Instead of stacking memory, d-Matrix stacks memory and compute. Bhoja says this reduces the distance that data must travel to “micrometers instead of millimeters.” Like building a skyscraper, going vertical makes it possible to do more inside the same physical footprint. Majestic takes the opposite approach. Instead of trying to minimize the length that data must travel between compute and memory, the company is focused on improving the memory interface to accommodate longer wire traces while keeping bandwidth high. Longer wires allow Majestic to connect memory stacks that aren’t directly next to the GPU, removing the space limitation of HBM. “A memory interface has a very short physical distance it can operate over. In the case of HBM, it’s up to 2 or 3 millimeters. You have this shoreline around the periphery, which is the only place where you can put HBM,” says Rabii. Majestic [claims](https://www.techradar.com/pro/startup-swaps-costly-ai-gpus-for-arm-cores-and-up-to-128tb-of-cheap-lpddr6-ram-instead-of-expensive-hbm-to-smash-through-the-memory-wall) its memory interface can transmit bits as far as about a meter. That’s achieved with a proprietary copper link and a memory-aggregator chip that coordinates data. “The aggregator is the endpoint for the high-speed interface and a way to fan out to many, many commodity DRAM chips,” says Rabii. Because of this, Majestic can support up to 128 terabytes of DRAM memory in a single server rack—a significant increase over Nvidia’s [GB300 NVL72 rack](https://www.nvidia.com/en-us/data-center/gb300-nvl72/), which has about[20 TB of HBM3E](https://resources.nvidia.com/en-us-blackwell-architecture/blackwell-ultra-datasheet?ncid=no-ncid). ### Majestic Labs’ memory-aggregation architecture Majestic Labs plans to satisfy AI’s memory appetite via a proprietary interconnect and a memory-aggregator chip, allowing a single rack access to up to 128 terabytes of cheap DRAM memory. Chris Philpot d-Matrix and Majestic have one thing in common: Instead of HBM, they both use off-the-shelf DRAM. This is the most common type of computer memory in the world; it’s in everything from smartphones to cars. Memory analyst [Jim Handy](https://thememoryguy.com/) says HBM costs two to three times as much as DRAM. d-Matrix and Majestic chose DRAM in part because of this price advantage. However, the proponents of HBM, which include memory giants like [Samsung](https://www.samsung.com/us/) and [SK Hynix](https://www.skhynix.com/), aren’t sitting idle. HBM4, the latest version of HBM memory, is now in production and will be used by [Nvidia’s Vera Rubin GPU](https://spectrum.ieee.org/nvidia-rubin-networking), which is expected to ship in the second half of 2026. [Hoshik Kim](https://www.linkedin.com/in/hoshikk/), head of memory-systems research at [SK Hynix](https://www.skhynix.com/), says HBM4 “will decisively break the memory bottlenecks constraining AI inference today” by doubling HBM’s maximum memory bandwidth and increasing the amount of HBM memory per stack. ## Combining chips for faster inference The big players—Nvidia and Amazon—are going for an all-chips-on-deck approach. Nvidia’s GPUs and Amazon’s Trainium training accelerators are still great for part of the inference workload: the prefill stage, where all the context keys and values are calculated. But to accelerate decode, the part where new tokens are generated, they are looking to new, memory-centric architectures from smaller players. In Nvidia’s case, the smaller player was Groq (not to be confused with Grok, the family of LLMs trained by SpaceXAI). Nvidia purchased intellectual property and hired talent from Groq at the end of 2025, and just three months later at the Nvidia’s GTC 2026 conference, Jensen Huang [unveiled](https://spectrum.ieee.org/nvidia-groq-3) the Nvidia Groq 3 language-processing unit ([LPU](https://www.nvidia.com/en-us/data-center/lpx/)). Groq’s architecture relies on memory—in its case, SRAM—built directly into the chip’s architecture. Unless you’re a chip architect, or a [hardcore PC gamer,](https://www.pcworld.com/article/2634140/why-i-care-about-cpu-cache-as-a-pc-gamer-the-obscure-spec-explained.html) you probably never give SRAM a thought. SRAM has the benefit of being tightly integrated into a compute chip’s architecture—it’s on the same piece of silicon as the processor—and has the drawback of being less dense and more expensive than DRAM. Most chips include only a few dozen megabytes of SRAM. AI inference, however, has ignited new interest in SRAM as a means of bringing the model weights stored in memory closer to compute. [Ian Buck](https://www.linkedin.com/in/ian-buck-19201315/), vice-president and general manager of hyperscale and high-performance computing at [Nvidia](https://blogs.nvidia.com/blog/author/ian-buck/), says the LPU has a much different set of priorities than the company’s GPUs. The LPU has far less raw computing power than a standard GPU, but it gains 500 megabytes of on-die SRAM connected directly to its floating-point math units. “The benefit is the memory bandwidth. The LPU has seven times the memory bandwidth of the GPU,” he says. Between the Rubin GPU and the Groq LPU, prefill and decode can both be accelerated to get the best of both worlds, the theory goes. “We do all the attention math and context processing on the Vera Rubin [GPU] rack,” explains Buck. “For all the expert calculations…the matrix multiplications, we do that part on the LPU.” The company packs 256 LPUs into the Groq 3 LPX, a system the size of a data-center rack. ### Nvidia’s two-chip approach to inference Nvidia also plans to split the inference workload across two chips. The company’s newest Rubin GPUs will tackle the compute-intensive prefill phase, while the Groq 3 language-processing unit (LPU), with lots of on-chip SRAM, will handle the memory-intensive decode phase. Chris Philpot Amazon Web Services (AWS), for its part, struck a [deal](https://www.aboutamazon.com/news/aws/aws-cerebras-ai-inference) with [Cerebras](https://spectrum.ieee.org/tag/cerebras), to pair the Trainium accelerator with [Cerebras’s Wafer-Scale Engine 3 (WSE-3)](https://www.cerebras.ai/chip). Cerebras takes a similar approach to Groq, though at a much larger scale. WSE-3 turns an entire silicon wafer into a single chip that contains over 4 trillion transistors. The design doesn’t connect to external memory but instead etches 44 gigabytes of SRAM into each wafer. “We store the [model] weights on the SRAM,” says [James Wang](https://www.linkedin.com/in/james-wang-5166575/), formerly director of product marketing at Cerebras who has since moved to SpaceXAI. “So that’s easily 40 to up to 80 billion parameters that we can support on one chip.” Amazon plans to use AWS Trainium chips for prefill, and Cerebras for decode. But Cerebras’s chips can also go it alone in inference. WSE-3 was [deployed by OpenAI to power GPT-5.3-Codex-Spark](https://openai.com/index/introducing-gpt-5-3-codex-spark/), a variant of the company’s coding mode, outputting over 1,000 tokens per second. For comparison, OpenAI’s standard GPT-5.4 deployment outputs 50 to 125 tokens per second. ### Amazon Web Services’ two-chip inference strategy Amazon Web Services combined their Trainium chips with Cerebras’s dinner-plate-sized wafer-scale engine (WSE) to tackle different parts of AI inference. Trainium chips handle the computationally intensive prefill phase, while the WSE, with interleaved on-chip SRAM memory, handles the memory-bandwidth-limited decode phase. Chris Philpot Cerebras can also tackle prefill without moving the workload to different specialized chips. For this, it networks together multiple WSE-3 chips to form a single pool of memory. Cerebras has demonstrated it can serve models with up to 1T parameters, such as [Moonshot](https://spectrum.ieee.org/tag/moonshot) AI’s Kimi 2.6, though Wang says “the architecture has no innate limitation in terms of how many parameters it will do.” Despite these differences in strategy, Nvidia and AWS seem to agree that the future of AI inference will be solved by a systems approach that pools different kinds of chips together to tackle the largest LLMs. Or, as Buck says: “To do modern AI inference, you need all the chips.” ## Learning to do more with less (bits) Nvidia became the world’s most valuable tech company because it designed the world’s most desired GPUs. But not all of the attention is focused on improving AI-inference hardware. AI researchers are also learning how to optimize LLM software and hardware in tandem to make the best use of the memory and compute components. Most computers store numbers in a 32-bit or 64-bit format. These determine how many bits are available to represent a single number. If too few bits are available, the number can’t be stored without losing information. The quality of an LLM benefits from more-precise number formats, but this creates a problem for inference performance. More-precise numbers aren’t free. The bits that describe them take up more space in memory and require more silicon and energy to compute. [Gilles Backhus](https://www.linkedin.com/in/gillesbackhus/?originalSubdomain=de), cofounder of the AI-accelerator company [Tensordyne](https://www.tensordyne.ai/), says this creates a tension between model size and number precision. “Would you prefer a model that is size _x_ but runs in 8-bit, or would you prefer a model that is twice the size but runs in 4-bit?” The size of each model will be roughly the same in terms of memory and compute, “but the 4-bit approach gives you twice as many synapses, if you will. And people are figuring out that [the 4-bit approach] is worth it.” The process of converting an LLM from a more-precise number format to a less-precise format is called [quantization](https://spectrum.ieee.org/1-bit-llm), and it’s been in use for several years. However, researchers are finding new ways to quantize models down while retaining a large majority of the model’s quality. Nvidia recently created a new 4-bit number format, [NVFP4](https://developer.nvidia.com/blog/introducing-nvfp4-for-efficient-and-accurate-low-precision-inference/), for this purpose. [AMD](https://www.amd.com/en.html), [Intel](https://www.intel.com/content/www/us/en/homepage.html), and [Qualcomm](https://www.qualcomm.com/) have instead rallied around a competing 4-bit number format called [MXFP4](https://huggingface.co/blog/RakshitAralimatti/learn-ai-with-me) that Nvidia also contributed to developing. “It’s the black art of AI,” says Buck, of Nvidia. When Nvidia quantized DeepSeek-R1 from FP8 to NVFP4, scores on seven major benchmarks degraded by less than one percent while [performance improved by three times](https://developer.nvidia.com/blog/3-ways-nvfp4-accelerates-ai-training-and-inference/), the company says. Quantization is likely just the tip of the spear, as AI researchers and startups are investigating a diversity of opportunities for optimization, some of which could dramatically change the silicon found in AI-inference hardware. Tensordyne’s unique approach to AI inference combines a logarithmic number format with bespoke hardware in the company’s Napier chip. Tensordyne Tensordyne is expected to [accelerate](https://spectrum.ieee.org/tensordyne-inference-claim) AI inference with a logarithmic number system that leans on a property of logarithms: The log of A times B equals the log of A plus the log of B. So, storing numbers as their exponents lets the chip add where it would otherwise multiply. That matters in silicon because multiplier circuits draw more power and use more die area than adders do. Tensordyne says its rack-scale hardware, called Napier, can produce up to 1,300 tokens per second per user, and can do so while using less than a [tenth](https://www.tensordyne.ai/stories/tensordyne-announces-breakthrough-inference-system-to-end-ais-speed-vs-cost-trade-off) as much power as comparable Nvidia hardware. [Etched](https://www.etched.com/), a startup based in San Jose, Calif., is even designing AI accelerators that translate the transformer architecture used by LLMs directly into silicon. Rather than building general-purpose GPUs, the company is wiring up the connections needed for efficient transformer calculations into its chip, making the chip much less flexible but more efficient for the tasks most performed by current LLMs. The company says its first AI accelerator, [Sohu](https://www.spheron.network/blog/etched-ai-sohu-vs-nvidia-transformer-asic-inference/), can run Meta’s [Llama](https://spectrum.ieee.org/tag/llama) 70B model at a stunning 500,000 tokens per second, though this approach also means it won’t be able to run LLMs that move away from a typical transformer architecture. Whether these ideas will prove fruitful remains to be seen. Etched just [shipped](https://www.etched.com/progress/from-zero-to-one) their first rack in August. Tensordyne believes its first hardware will be available in 2027. Even so, these startups show how the demand for inference performance is fueling unconventional ideas. ## Inference is everyone’s game The sheer variety of approaches to AI-inference acceleration—stacking compute on memory, extending interfaces from millimeters to meters, using an entire silicon wafer for SRAM, squeezing models into 4 bits—raises a question: Which is going to win, and which is going to lose? But that’s likely not the right question, experts say. The demand for AI is currently insatiable, and while fears of an AI bubble stalk the industry, it has yet to hamper growth. On the contrary, Kimball of Moor Insights & Strategy thinks inference could drive intense demand for AI hardware in the long term, because it’s not obvious where that demand will end. “You could add a million agents into your organization,” he says. “These things work 24 hours a day; they don’t go home at five at night like we do.” If AI inference remains as desirable as Kimball expects, the evolution is likely to follow the same trajectory as the CPU. The CPU didn’t improve along a single axis but instead across [multiple fronts](https://spectrum.ieee.org/intel-i860) simultaneously. Once transistor scaling slowed, chip and system architecture innovations of all kinds proliferated. The list of individual innovations that led to today’s ubiquitous, powerful personal compute could fill dozens of books. A few decades from now, the history of AI inference innovation will show similar depth. From Your Site Articles * [With Nvidia Groq 3, the Era of AI Inference Is (Probably) Here ?](https://spectrum.ieee.org/nvidia-groq-3) * [AI Inference Competition Heats Up ?](https://spectrum.ieee.org/new-inference-chips) Related Articles Around the Web * [Ultimate Guide – The Best and Fastest AI Inference Engines of 2026 - SiliconFlow ?](https://www.siliconflow.com/articles/the-fastest-ai-inference-engine) * [LLM Inference Hardware: Emerging from Nvidia's Shadow ?](https://gradientflow.substack.com/p/llm-inference-hardware-emerging-from) Keep Reading ↓ {"imageShortcodeIds":[]} [Robotics](https://spectrum.ieee.org/topic/robotics/)[AI](https://spectrum.ieee.org/topic/artificial-intelligence/)[Computing](https://spectrum.ieee.org/topic/computing/)[Humanoid Robots](https://spectrum.ieee.org/topic/robotics/humanoid-robots/)[Sponsored Article](https://spectrum.ieee.org/type/sponsored/) ## [Rethinking Robot Safety in the Age of AI](https://spectrum.ieee.org/physical-ai-robot-cybersecurity-vicone) ### How to secure today’s AI robots against stealth cyber threats [VicOne](https://spectrum.ieee.org/u/vicone-) [VicOne](https://vicone.com/) is a Physical AI cybersecurity company protecting vehicles and robots from cyber threats that can alter behavior and compromise safety. Headquartered in Tokyo, Japan, the company provides automotive and robotics organizations with software and services to identify cyber risks early, assess their potential safety impact, and protect systems in operation, helping maintain safe, reliable behavior across real-world deployments. 16 Sep 2026 4 min read [](javascript:;) 3 [](https://spectrum.ieee.org/physical-ai-robot-cybersecurity-vicone) Cyberattacks can change how robots act without obvious signs of failure, requiring a new approach to safety that combines AI model and vulnerability scanning, simulation-based validation, and continuous monitoring to help secure robots from development through operation. VicOne [](mailto:?subject=Rethinking%20Robot%20Safety%20in%20the%20Age%20of%20AI&body=https://spectrum.ieee.org/physical-ai-robot-cybersecurity-vicone) _This article is brought to you by [VicOne](https://vicone.com/?utm\_source=ieee-spectrum&utm\_medium=sponsored-content&utm\_campaign=2026-09)._ Robot safety has traditionally asked: Can a machine remain safe when something goes wrong? Physical AI raises a harder question: Can a machine remain safe when an attacker changes what it sees, decides, or does even when nothing appears to have failed? As AI and robotics continue to advance at an unprecedented pace, modern robots perceive through multimodal sensors, interpret context using [AI models](https://spectrum.ieee.org/tag/ai-models), and translate those interpretations into physical action. As they move into dynamic environments, their safety increasingly depends on the integrity of the data guiding their decisions. That dependence creates risks that conventional safety assessments may not fully capture. Recent research has demonstrated that manipulating what a robot sees, hears, or interprets can influence its behavior without requiring direct control. Such manipulation can occur anywhere across its complex sensing and decision-making system — a layered attack surface encompassing training pipelines, system infrastructure, and runtime perception. ## Layer One: Corrupting intelligence at its source In 2017, [BadNets](https://arxiv.org/abs/1708.06733) demonstrated that a model could behave normally under most conditions, yet fail in the presence of a specific hidden trigger. In one example, a subtle pattern caused a stop sign to be misclassified as a speed limit sign without affecting the model’s behavior on other inputs. What began as a classification vulnerability has since evolved into action manipulation. At [NeurIPS](https://spectrum.ieee.org/tag/neurips) 2025, researchers introduced [BadVLA](https://arxiv.org/abs/2505.16640)a backdoor attack targeting Vision-Language-Action (VLA) models that allow robots to see, interpret instructions, and produce coordinated physical movement. Rather than altering a single label, the attack caused conditional