Google (GOOGL) is synonymous with searching for information. Most people interact with its product on a daily basis. Alphabet is the holding company that owns Google. Google is the technology company behind Google Search, Google Chrome, Chrome OS, YouTube, GoogleAds, Google Cloud, Pixel smart phones and tablets. Google has for years invested in AI and is leveraging AI to improve its search, Ads, YouTube recommendations, Pixel phone, Google Docs and more. This giant’s revenue is growing and its income is also growing year over year as of Q3 2023 and it has a strong cash position. On top of that it has an AI opportunity to increase its revenue and strengthen its products, making them more useful.
Recent News:
Gemini 1.5 Pro has 1 million token context window: https://investinwhatyouknow.com/google-introduces-gemini-1-5-pro-with-1-million-token-context-window/
Google Services:
| Android |
| Chrome |
| hardware |
| Google Maps |
| Google Play |
| Search |
| YouTube |
| Cloud Infrastructure |
| Google Workspace communication and collaboration tools |
AI Opportunity
Google has invested heavily on AI over the last decade. They have a world class AI team. They can use AI to make search and other product better. Its Google Cloud division can offer compute and AI services to other companies. PaLM2 is Google’s latest LLM model.
https://ai.google/discover/palm2
Google Bard chat bot is built on top of it. Google Bard is rebranded Gemini. Gemini is free but Gemini Advanced uses the capable Ultra 1.0 model costs $19.99/month. It comes bundled with 2 TB storage and Google One benefits like Google VPN.
Google Workspace starts at $6/user/month for Business Starter w/ 30 GB pooled storage per user, $12 /user /month for Business Standard and provides 2 TB pooled storage per user, $18 /user/month for Business Plus w/ 5TB pooled storage per user, and Enterprise where price is customer specific. https://workspace.google.com/pricing.html
Duet AI, Google’s AI enhancements to productivity products and cloud expands on its product offering and increase its revenue. This Google blog post introducing Duet AI list some of its key features. The article says Google Workspace is used by over 3 billion users (most of that free users) and more than 10 million paying customers.
https://workspace.google.com/blog/product-announcements/duet-ai-in-workspace-now-available
In a CNBC article published August 29, 2023 and on other tech news sites, Google plans to charge enterprise customers $30/user/month for Duet on top of what customers already pay for Google Workspace. https://www.cnbc.com/2023/08/29/google-will-charge-enterprises-30-a-month-for-duet-ai-in-workspace.html
This is a similar strategy and opportunity as Microsoft’s CoPilot.
Google have built its own AI chip called the TPU. According to its research paper: https://arxiv.org/abs/2304.01433, its TPU v4 is 1.2x-1.7x faster and 1.3x-1.9x less power than Nvidia A100. As of February 9th 2024, their current version is TPU v5p. TPU v5p is has 2x more FLOPS and 3x more HBM memory than TPU v4. https://cloud.google.com/blog/products/ai-machine-learning/introducing-cloud-tpu-v5p-and-ai-hypercomputer. It can train GPT3-175B param 2.8x faster than TPU v4. TPU v5p is used by Google AI and other services and is also available to Google Cloud customers by request. Not having to build Nvidia enables Google access to compute without having to fight for H100 supply and reduces cost as they do not have to pay Nvidia. Nvidia blog: https://developer.nvidia.com/blog/nvidia-hopper-architecture-in-depth/ shows H100 to be 4x faster than A100 for GPT3 175B param. Since TPU v4 is 1.3x -1.9x faster than A100, TPU v5p is 2.8x faster than v4, using simple bath, TPU v5p is 2.8x(1.3x to 1.9x)A100) = 3.6xA100 to 5.3xA100. TPU v5p is on par to superior to H100.
AI Challenge and New Competition
The AI also creates some challenges for Google. LLM are becoming very good. When a chatbot can answer a question, the user will not have to use Google to search. This will impact Google’s search revenue. While Google have a competitive LLM in Gemini, OpenAI GPT4, Mistral Large, Meta Llama offer good competition.
Perplexity is a new LLM powered search engine. https://www.perplexity.ai. It can answer search query using LLM with grounding from web pages to provide the source of its answer. The user can use the free LLM models or subscribe to use a strong or private LLM from OpenAI. A unique feature is Perplexity shows the user the source links that its answer is based on. Perplexity gets the job done and web site is clean and uncluttered. If user gets answer from Perplexity, they will not need to hit Google. Google will need to match Perplexity’s ease of use before users form new habits and use Perplexity as its goto source for answers.
Health Metrics
| Health Metric | Q1 2024 | Q4 2023 Millions | Q3 2023 | Q1 2023 | Q4 2022 | Q3 2022 |
| Total Revenue | 80,539 | 86,310 | 76,693 | 69,787 | 76,048 | 69,092 |
| Revenue change yoy | 15% | 13% | 11% | 3% | 1% | 6% |
| Net Income | 23,662 | 20,687 | 19,689 | 15,051 | 13,624 | 13,910 |
| Diluted EPS | 1.89 | 1.64 | 1.55 | 1.17 | 1.05 | 1.06 |
| Google Services Revenue | 46,156 | 76,311 | 67,986 | 40,359 | 67,838 | 61,377 |
| Google Cloud Revenue | 9,574 | 9,192 | 8,411 | 7,454 | 7,315 | 6,868 |
| Google Services Income | 27,897 | 26,730 | 23,937 | 21,737 | 20,222 | 18,883 |
| Google Cloud Income | 900 | 864 | 266 | 191 | (186) | -440 |
| Cash & cash equivalents and marketable securities | 108,090 | 110,916 | 119,935 | 110,916 | 113,762 | 113,762 |
| Price | 168.65 (5/10/2024) | 147.71 (1/22/2024) | ||||
| Earnings Per Share (TTM) | 6.52 | 5.21 (1/22/2024) | ||||
| PE (TTM) | 25.87 | 28.35 |
Source:
Q1 2024 Earnings Release: https://www.abc.xyz/assets/91/b3/3f9213d14ce3ae27e1038e01a0e0/2024q1-alphabet-earnings-release-pdf.pdf
Q1 2024 Earnings call and transcript: https://abc.xyz/2024-q1-earnings-call/
Q4 2023 Earnings Release: https://www.abc.xyz/assets/95/eb/9cef90184e09bac553796896c633/2023q4-alphabet-earnings-release.pdf
Alphabet Q3 2023 Earnings Release: https://www.abc.xyz/assets/4a/3e/3e08902c4a45b5cf530e267cf818/2023q3-alphabet-earnings-release.pdf