Daily Management Review

Qualcomm And Nvidia Compete For The Top Rank In AI Chip Efficiency Testing


Qualcomm And Nvidia Compete For The Top Rank In AI Chip Efficiency Testing
In a new set of test results released on Wednesday, artificial intelligence chips from Qualcomm Inc. beat Nvidia Corp. in two of three measures of power efficiency, while a Taiwanese firm outperformed both in one.
When it comes to using massive volumes of data to train AI models, Nvidia dominates the market. However, after these AI models have been trained, they are used more broadly in "inference" tasks like creating text responses to questions and determining whether an image contains a cat.
Analysts predict that as firms integrate AI technology into their products, the market for data center inference chips will expand significantly. However, corporations like Alphabet Inc.'s Google are already looking into ways to limit the additional expenses that this will entail.
Electricity is one of those significant costs, and Qualcomm has utilized its experience creating CPUs for battery-powered devices like smartphones to produce a chip called the Cloud AI 100 that strives for thrifty power consumption.
Based on how many data center server queries each chip can handle per watt, testing results released on Wednesday by MLCommons, an engineering consortium that maintains testing benchmarks widely used in the AI chip industry, showed that Qualcomm's AI 100 chip outperformed Nvidia's flagship H100 chip at classifying images.
In comparison to Nvidia, Qualcomm's processors achieved 197.6 server queries per watt. With 227 queries per watt, Neuchips, a business founded by renowned Taiwanese chip academic Youn-Long Lin, won the competition.
In the object detection test, Qualcomm outperformed Nvidia with a score of 3.2 queries per watt as opposed to Nvidia's 2.4 queries per watt. Applications for object detection include looking at video from stores to determine where customers frequent most.
Nvidia, however, won a test of natural language processing, the AI technology most frequently employed in systems like chatbots, in both absolute performance terms and power efficiency terms. Nvidia achieved a sampling rate of 10.8 samples per watt, followed by Neuchips at 8.9 samples per watt and Qualcomm at 7.5 samples per watt.
Indian Startups Request An Antitrust Investigation Into The Google In-App Payment Cost
According to filing papers, Alphabet's Google is being investigated for allegedly violating an antitrust guideline by charging a high service fee for in-app purchases, according to Tinder owner Match Group and Indian entrepreneurs.
The lawsuits filed by Match and the Alliance of Digital India Foundation (ADIF) are the latest exchange of words between Google and its competitors, who have long criticized the American giant for what they perceive to be unfair business constraints.
"Google's policy change of a charging service fee even on transactions processed by third-party payment processors ... has detrimental consequences for users and app developers," the 15-page confidential March complaint by ADIF said.
The Android mobile operating system and Google Play app store are both supported by the service fee, according to Google, which declined to comment. It also pays for developer tools and analytic analytics.
The ADIF and Match files' specifics, which Reuters examined on Thursday, have not previously been discussed. Requests for comment from ADIF, Match, and the CCI went unanswered.
In October, the Competition Commission of India (CCI) fined Google $113 million and ordered it to permit third-party billing and cease requiring developers to use its in-app payment system, which levies a commission of 15%–30%.
The ADIF claimed in its lawsuit that the new method charges a "service fee" when Google later opted to start offering User Choice Billing (UCB) for permitting other payments in addition to Google's when purchasing in-app digital content.
"The app developers will have to pay 1%-3% for alternate payment service providers and 11%-26% to Google, which makes the entire ecosystem unsustainable," ADIF said.
Match claimed that the method was "anti-competitive" in its March 21 complaint, asking the CCI to order Google not to collect or impose any commissions or service fees, including through user choice invoicing.
Other legal difficulties exist, forcing Google—which sees India as a significant growth market—to alter the way it promotes the Android operating system.
In an order issued in October, the CCI said that Google had abused its market dominance and that the mandated adoption of its proprietary payment system had constrained the possibility for technical advancement and innovation among payment processors and app developers.
This has been contested by Google in a court in India.
In its March lawsuit, ADIF, which represents Indian startups such as social media app ShareChat and digital payments company Paytm, said that Google was exploiting the new service charge scheme to evade the antitrust rule that prohibited it from imposing any "unfair and disproportionate" restrictions.
"The policy of UCB is unfair and the same would lead to unjust enrichment to Google," the ADIF filing said.