Google is winning the AI race. The rest of big tech is trying to figure out what that even means.
That's the blunt verdict from MG Siegler, the tech blogger and former Google Ventures partner who has been mapping the AI strategies of the largest companies since ChatGPT landed in late 2022. His latest assessment — delivered on the Big Technology Podcast on January 12, 2026 — puts Alphabet alone at the top of what he calls the AI chaos ladder. Amazon.com, Microsoft, Apple, and Meta Platforms are all, in his words, "chaotic." None of them has a standalone AI product that people chose to use.
Alphabet's position at the top isn't theoretical. In January 2026, when Apple and Google announced a deal to run future versions of Siri and other Apple AI features on Gemini, Alphabet briefly touched a $4 trillion market cap for the first time, according to Fast Company. The deal itself is concrete validation: Apple is paying Google roughly $1 billion per year — as BreezyScroll reported — for access to a custom 1.2-trillion-parameter Gemini model that will power a redesigned Siri, as NerdLevelTech reported. The two companies made the joint announcement on January 12, 2026, according to 9to5Mac. Google isn't just participating in the AI era — it is the infrastructure.
The other four mega-cap companies are in a different position entirely, and the common thread is damning. "What standout AI product does Microsoft have outside of selling AI services from Azure?" Siegler asked on the podcast. "What standout AI product does Amazon have? What standout AI product does Apple have? What standout AI product does Meta have? They have none."
He's not wrong. Microsoft has leaned hard on Copilot, but the branding is fractured — the enterprise product, the consumer product, and the features embedded in Office all share a name without sharing an identity. "They can't figure out the brand," Siegler said. The underlying product has similarly failed to cohere into something users seek out. Amazon's Alexa, now in the Plus iteration, has been iteratively improved but remains an overlaid assistant on a voice platform from a decade ago — better, Siegler acknowledges, but not native. Apple, which observers have long suspected lacks a credible in-house AI strategy, is paying Google to solve that problem for it. The $1 billion check is an admission.
Meta has announced AI features embedded across WhatsApp, Instagram, and the Ray-Ban smart glasses. But none of them is a standalone AI product in the way ChatGPT is — something users open deliberately, repeatedly, by choice.
The pattern — bolting AI onto existing products rather than building AI-first — is exactly what OpenAI's Sam Altman got right when he told Siegler in a separate conversation that "you can't just bolt AI on existing products. You need to build ground up." The companies that built from scratch — ChatGPT, Claude, Gemini — became the AI products. The companies that tried to retrofit their existing platforms are watching.
The divergence runs deeper than strategy. OpenAI is moving toward Google's playbook, exploring ads and transaction revenue alongside its subscription model. Anthropic, the AI safety company behind Claude, is staying premium — doubling down on subscriptions and data tools as its core business. These are fundamentally different bets about where value accrues in the AI stack.
The IPO race is sharpening that divide. Anthropic is discussing a Q4 2026 initial public offering with bankers expecting the company to raise more than $60 billion, as WinBuzzer reported. OpenAI closed a $110 billion funding round in February 2026 at an $840 billion post-money valuation, according to Business Insider, and is targeting its own public listing. Siegler argued that if Anthropic goes public first, it would become existential for OpenAI — absorbing the pent-up investor demand that a public Anthropic listing would attract. OpenAI's $25 billion in annual revenue — reported by Techi and not independently confirmed — needs to compound significantly before the IPO window closes on terms that make sense for the company.
Samsung Electronics is a wildcard. The company has pre-installed Perplexity on the Galaxy S26, as Medianama reported. Siegler predicted Samsung will acquire Perplexity outright for approximately $20 billion within twelve months. The reasoning: AI-powered search is being absorbed into general chatbots, squeezing standalone AI search products. If true, that would remove one of the few independent consumer AI products from the field and raise questions about whether any non-lab AI company can build a durable consumer franchise.
On the question of who can sell AI the way Steve Jobs sold the iPhone, Siegler argued that Nvidia's Jensen Huang is the closest the industry has come — someone who commands a stage in a way that makes even hardware feel relevant. Among AI researchers themselves, Siegler argued that Google's Demis Hassabis comes closest, citing his scientific credibility. According to Siegler, the documentary "The Thinking Game," which follows Hassabis and DeepMind, has drawn approximately 260 million views on YouTube — a figure Siegler cited on the podcast without independent verification. The comment section tone on that documentary contrasts sharply with the personal vitriol that often greets AI company leaders.
Tim Cook, Apple's chief executive, denied on ABC's GMA in March 2026 that he is preparing to step aside, as CNBC reported. But Siegler predicts he will retire as CEO before the end of the year. Apple is expected to release its first foldable iPhone — with an unfolded display around 7.8 inches and a folded size of approximately 5.5 inches — in either September or December 2026, according to MacRumors. Barclays analyst Tim Long, as MacRumors reported, said the December launch targets the holiday season months after the iPhone 18 Pro. If the device lands with a rethought Siri powered by Gemini, it will be the most concrete AI product Apple has shipped — and it will run on Google's models.
The real story here isn't the Steve Jobs question. It's that four of the five most valuable technology companies in the world cannot point to a consumer AI product that people chose to use. Google can. Everyone else is either running someone else's AI inside their products, or trying to bolt AI onto platforms that were never designed for it. That is the structural problem. No amount of keynotes will fix it.