- Alphabet unviels new Arm-based Axion TPU to compete with Nvidia, Intel, AMD.
- The v5p chip will only be accessibly via Google Cloud.
- GOOGL stock rises on the news despite market sell-off ahead of March CPI.
- Technicals show that Alphabet stock is likely to experience a near-term reversal.
Alphabet (GOOGL), the parent of Google, released its answer to Nvidia (NVDA) on Tuesday with an Arm-based type of central processing unit (CPU) meant to carry artificial intelligence (AI) workloads in data centers.
The announcement positions Alphabet as a competitor to Nvidia, whose various chips, mostly GPUs, currently dominate the AI training market for large language models (LLMs).
US markets disappointed after briefly rising on Tuesday at the start of the session. Traders are nervous about Wednesday’s Consumer Price Index (CPI) for March. Nvidia stock drops 3% on Tuesday, while Arm Holdings (ARM) stock has lost more than 2% at the time of writing.
Specifically, Alphabet unveiled its tensor processing unit (TPU) called the Axion via Google Cloud or TPU v5p. The name stems from the fact that customers will only have access to the chip via Google’s cloud platform.
Officials say that operating in clusters of 8,960 chips allows the technology to render AI workloads at twice the performance of earlier TPU versions.
“We’re making it easy for customers to bring their existing workloads to Arm,” said Mark Lohmeyer, Google Cloud’s vice president and general manager of compute & machine learning infrastructure. “Axion is built on open foundations, but customers using Arm anywhere can easily adopt Axion without re-architecting or re-writing their apps.”
The company did not mention whether Broadcom (AVGO) had helped them design the new chip as the latter company did with earlier TPU versions. Both Amazon (AMZN) and Microsoft (MSFT) have previously designed their own CPUs in order to differentiate each of their cloud platforms.
Additionally, Google said that the TPU v5p chips perform 30% better than standard Arm chips and 50% better than x86 architectures produced by Advanced Micro Devices (AMD) and Intel (INTC).
Social media company Snap (SNAP) is expected to be first in line to begin testing out the new technology in Google Cloud.
AI stocks FAQs
First and foremost, artificial intelligence is an academic discipline that seeks to recreate the cognitive functions, logical understanding, perceptions and pattern recognition of humans in machines. Often abbreviated as AI, artificial intelligence has a number of sub-fields including artificial neural networks, machine learning or predictive analytics, symbolic reasoning, deep learning, natural language processing, speech recognition, image recognition and expert systems. The end goal of the entire field is the creation of artificial general intelligence or AGI. This means producing a machine that can solve arbitrary problems that it has not been trained to solve.
There are a number of different use cases for artificial intelligence. The most well-known of them are generative AI platforms that use training on large language models (LLMs) to answer text-based queries. These include ChatGPT and Google’s Bard platform. Midjourney is a program that generates original images based on user-created text. Other forms of AI utilize probabilistic techniques to determine a quality or perception of an entity, like Upstart’s lending platform, which uses an AI-enhanced credit rating system to determine credit worthiness of applicants by scouring the internet for data related to their career, wealth profile and relationships. Other types of AI use large databases from scientific studies to generate new ideas for possible pharmaceuticals to be tested in laboratories. YouTube, Spotify, Facebook and other content aggregators use AI applications to suggest personalized content to users by collecting and organizing data on their viewing habits.
Nvidia (NVDA) is a semiconductor company that builds both the AI-focused computer chips and some of the platforms that AI engineers use to build their applications. Many proponents view Nvidia as the pick-and-shovel play for the AI revolution since it builds the tools needed to carry out further applications of artificial intelligence. Palantir Technologies (PLTR) is a “big data” analytics company. It has large contracts with the US intelligence community, which uses its Gotham platform to sift through data and determine intelligence leads and inform on pattern recognition. Its Foundry product is used by major corporations to track employee and customer data for use in predictive analytics and discovering anomalies. Microsoft (MSFT) has a large stake in ChatGPT creator OpenAI, the latter of which has not gone public. Microsoft has integrated OpenAI’s technology with its Bing search engine.
Following the introduction of ChatGPT to the general public in late 2022, many stocks associated with AI began to rally. Nvidia for instance advanced well over 200% in the six months following the release. Immediately, pundits on Wall Street began to wonder whether the market was being consumed by another tech bubble. Famous investor Stanley Druckenmiller, who has held major investments in both Palantir and Nvidia, said that bubbles never last just six months. He said that if the excitement over AI did become a bubble, then the extreme valuations would last at least two and a half years or long like the DotCom bubble in the late 1990s. At the midpoint of 2023, the best guess is that the market is not in a bubble, at least for now. Yes, Nvidia traded at 27 times forward sales at that time, but analysts were predicting extremely high revenue growth for years to come. At the height of the DotCom bubble, the NASDAQ 100 traded for 60 times earnings, but in mid-2023 the index traded at 25 times earnings.
GOOGL stock saw its light dim since the morning session on Tuesday when the stock shot up about 2.4% to $158.56, an all-time high. With prospects for the market dimming until the CPI data comes in for March on Wednesday, traders are standing pat. If the CPI data comes in hot, then expect GOOGL stock to return to support at $153, which was an area of resistance in January.
Somewhat more importantly, the technicals for Alphabet stock are bearish. First, Tuesday’s candlestick looks like a textbook Shooting Star. Rendered like an upside down hammer, the chart pattern at the top of a staircase rally often signals that a bearish reversal is in order.
Secondly, the ascending top trendline of the higher highs (orange line) on the daily chart is inverse to the descending trendline (also orange) on the Relative Strength Index (RSI) over the same period from late January to early April. This RSI Divergence pattern is also a sign of a bearish countertrend. The chart says to sell despite the day’s news on the chip front.
GOOGL daily stock chart