The overriding narrative from Wall Street to Silicon Valley recently has been about the winners and laggards in the new generative artificial intelligence, or AI, world.
Lumped among “winners” were firms like software supremo Microsoft and its partner OpenAI, the creator of AI chatbot, ChatGPT. Another was Nvidia, the chip behemoth which provides the picks and shovels, or AI chips, to players who need it for training or inference of data. Dumped as a “laggard” is search giant Google’s parent Alphabet Inc because it has lost its mojo.
Now Alphabet is trying to prove it can still go toe-to-toe with its biggest tech peers, and win.
For its part, the market has punished the laggards and rewarded winners. On one bad trading day in February, Alphabet lost over US$90 billion ($121.8 billion) in market value as controversy over Google’s generative chatbot became the focus of Wall Street traders.
Yet, stock market traders are a fickle lot with the attention span of gnats. On April 9, Alphabet’s shares without much fanfare touched a new all-time high of US$159.89, giving the search engine owner a market valuation of over US$2 trillion. The stock is up 20% from its lows in early March, 90% over the past year and up 500% over the past 10 years.
The world’s most unloved tech stock is not only at an all-time high, it is actually the cheapest of all large-cap stocks. Did I mention that Google’s owner is an extraordinary cash-generating machine? Alphabet had free cash flow of US$70 billion last year, or US$11 billion more than the annual free cash flow of Microsoft, currently the world’s most valuable firm. Alphabet also has US$99 billion net cash on its balance sheet — the largest corporate cash pile in the world.
See also: Alibaba anoints new chief in revamp of stalling commerce arm
Shares of the search giant’s holdings firm were trading this week at 23.5 times current fiscal year’s earnings. iPhone maker Apple shares trade at 25.2 times this year’s earnings, social media powerhouse Meta Platform 26.3 times, Microsoft at 35.2 times, Nvidia at 35.1 times and e-commerce pioneer Amazon.com at 43.1 times this year’s earnings.
According to consensus estimates, Alphabet is expected to grow its revenues 11% this year, compared to 8.7% last year. In contrast, Apple is expected to grow its revenues about 5% this fiscal year, while Meta’s revenues are forecast to grow 14% in 2024 and Amazon’s expected sales growth is 11% in the current year.
So, what’s really ailing Google’s parent? Brent Thill, Internet analyst for Jefferies & Co in San Francisco, cites three big concerns: long-term threat to search from AI, questions on management leadership, and the fact that Google is still considered a laggard in cloud infrastructure.
See also: Break up Google? What’s at stake in antitrust action
Despite its efforts to diversify, Alphabet is still hugely dependent on advertising which made up 75% of total US$307 billion revenues last year, with search advertising accounting for 60% of the total revenues. YouTube Ads is the other big advertising engine.
The remaining 25% of revenues comes from Google Cloud, as well as subscriptions from YouTube TV, music streaming, NFL Sunday American football games, and device sales like Pixel smartphones, Google Nest hub, etc.
Ad revenues could be pressured if users shift from search engines to chatbots when looking for information, while margins could be under pressure if Google’s traffic shifts from traditional search to its own AI tools which carry higher training and inferencing costs, the Jefferies analyst notes.
In the short run, the market is a voting machine on narratives like who is ahead in the ongoing AI war. In the long run, however, the market is a weighing machine that takes into account the ability of companies like Google and others to generate profits over time and remain on the cutting edge of innovation.
Taking on AI chatbots
AI is shaping up to be a huge once-in-a-generation transformational disrupter akin to the major technological inventions of the past several hundred years. “Think the printing press, the steam engine, electricity, computing and the Internet,” Jamie Dimon, CEO of the world’s largest financial services group, JPMorgan Chase, wrote in his annual shareholders letter last week.
Over the past three decades, two sweeping transformations have propelled American tech giants — the advent of the Internet in the late 1990s and the switch to mobile platform 15 years ago with the advent of smartphones. The key to new transformation is access to capital and technology, and for now only a handful of US tech giants have the resources to invest in new infrastructure and platforms to harness the power of AI.
Is there a danger of Google search being usurped by, say, its closest competitor Microsoft’s Bing, or a combination of Bing and pay-to-use OpenAI-Microsoft chatbots like its Copilots? Or could new AI-focused players like Perplexity and You.com replace Google as the next generation search engine?
Sink your teeth into in-depth insights from our contributors, and dive into financial and economic trends
Yes and no. For one thing, Google still has a huge lead in search and while AI chatbots are likely to make dents, it will be a long time before a rival ousts a free search engine like Google from its perch. For another, search is not just about answers to simple queries. A lot of search is about navigating to another website.
Google is toying with the idea of charging for new premium features powered by generative AI. That is a huge leap for a company that has so far relied mostly only on its free-to-use search business. Google generates most of its revenues from search-based advertising. A shift to fee-based AI-powered search would be a paradigm shift for Google, which has long dominated how we search for everything online from news to restaurants to holiday destinations and flights.
The Alphabet subsidiary will offer generative AI solutions on its search engine so when you type in a search query on your smartphone, tablet or PC, an AI chatbot will create answers for you rather than generating a bunch of blue links, many of them ad-sponsored, that you need to click on which may or may not give you a quick, straightforward answer to your query.
The idea is that people would be willing to pay for an AI chatbot spitting out concise, prompt answers to your queries. What you get as a paying search customer is a more useful, and possibly more thoughtful, answer to your query rather than an ad-based blue link boosted by search engine optimisation.
Why would anyone pay for search through AI tools where they can use Google traditional search for free? While Netflix now has a cheaper ad-version of movies and TV programming, over 90% of Netflix users currently chose the no-ads version.
Ironically, Google was one of the earliest movers in the AI space investing tens of billions over the past decade or so. Among other things, it bought DeepMind Technologies, UK-based AI research firm, in 2014. DeepMind’s co-founder was Mustafa Suleyman, UK-born son of a Syrian taxi driver who is now head of AI products at Microsoft. It assembled one of the biggest deep-learning AI research teams under Google Brain umbrella which employed Geoffrey Hinton, known as the godfather of AI, as well as AI expert Andrew Ng, ex-chief scientist at Chinese Internet giant Baidu.
Much to its chagrin, however, rivals like OpenAI as well as its partner, Microsoft, which has a 49% capped economic interest in it, have turned AI research projects into products more quickly.
From Bard to Gemini
Until now, Google has dragged its feet for the fear of disrupting its own free search cash cow. Not surprisingly, it is now being seen more as a follower rather than a leader in AI. After OpenAI unveiled its pioneering ChatGPT nearly 18 months ago, Google launched its own Bard chatbot to mixed reviews.
The search giant has since launched Gemini, its answer to OpenAI’s advanced models. More recently, Google has abandoned Bard, amalgamating it with Gemini, a foundational model, similar in concept to GPT-4 and focused on advanced reasoning.
The rebranding of Bard follows a series of missteps. Last December, Google released a marketing video showcasing Gemini’s real-time capabilities. Unfortunately, the video was heavily edited and a tad misleading. Responses were sped up, and the model could not handle live video as shown. This sparked a big backlash against Alphabet for overhyping and misrepresenting the interaction. Management apologised, but the incident has raised concerns that Google was willing to go to any lengths to prove it was not an AI laggard.
Yet with over 5,000 AI engineers working at Google and the AI boom still in its infancy, the search giant still has plenty of time to catch up with Microsoft and OpenAI which employ a combined 7,200 AI engineers. Its parent, Alphabet, also has a substantial stake in OpenAI’s main rival, an AI start-up called Anthropic, which also counts Amazon as one of its shareholders.
Alphabet has by far the biggest research budget among tech giants. It spent over US$45 billion on research and development (R&D) last year compared to Meta which spent US$39 billion, Apple which spent US$30 billion, and Microsoft which spent US$27 billion on R&D.
Google is also much better-positioned for the AI boom than many of its peers. With its huge market share in search and its ownership of gmail, it certainly has a lot more data than Microsoft or Meta for training and inference. While its rivals may need to fork out a lot of money to pay for data they need for training, Google has enough data in-house that it can use, giving it an edge over many of its rivals.
Its core search engine aside, Google is continuing to diversify. Despite being a latecomer to the party, Google has quietly built its cloud services business, helping to provide the computing power that companies need to store data. Synergy Research Group estimates that the cloud infrastructure market is growing over 20% annually.
Four players dominate the global cloud infrastructure market with over 70% of the market. Amazon Web Service or AWS has 31% market share, Microsoft’s Azure 24%, Google 11% and Alibaba 4%. Because a big chunk of AI large language models is run on the cloud platforms, infrastructure providers such as AWS, Azure and Google Cloud are seeing burgeoning demand.
Google is also ahead of its cloud infrastructure rivals in making its own silicon, or semiconductor chips, to power its AI-focused data centres. Last week, it unveiled Axion, a new custom-built, ARM-based server chip that can handle everything from YouTube advertising to reducing its reliance on Nvidia’s top-of-line H-100 AI chips. The new chip is expected to be in production later this year.
Though Amazon and Microsoft are pouring money into AI chips, and both Intel and Meta have unveiled their own in-house designed AI chips in recent weeks, Google has a longer history designing its own silicon. Google has used its proprietary tensor processing units or TPUs for years to help expand its core search business.
Google needs to reinvent itself if it wants to be a leader in the new AI era. There have been rumblings about CEO Sundar Pichai, who has led the firm for nine years, and whether he is the right person to helm Alphabet in the age of AI.
But Pichai is a product guy in the mould of Meta’s Mark Zuckerberg. He is not a salesman like Microsoft’s former CEO Steve Balmer, who was blamed for the software giant’s lost decade before Satya Nadella took the reins 14 years ago and turned it around.
If Alphabet stock remains a laggard, its board could turn up the heat on Pichai or an activist investor might emerge.
Assif Shameen is a technology and business writer based in North America
