Don't Count Out Alphabet
Plus, quantum computing sees funding uptick as Google touts breakthrough
The Main Item
A Year-End Flex for Alphabet with Quantum Chip, Gemini 2.0 & Waymo Wins
Alphabet was too fat and happy to lead the generative AI revolution, or so said the conventional tech wisdom a year ago.
The company looked like it had been caught off-guard by ChatGPT, and didn’t have the competitive culture and mindset needed to win in the “age of efficiency.”
Google researchers might have come up with the “transformers” architecture that paved the way for the current AI mania, but most left the company years ago. Its early AI deployments drew mockery for overly woke and inaccurate results. Its search monopoly suddenly looked vulnerable. A government antitrust suit threatened its advertising money-machine.
Microsoft, meanwhile, had built a tight partnership with the new tech darling, OpenAI. A revived Meta was drawing plaudits for its operational rigor. Nvidia, not long ago a middle-tier chip-maker, was now touted as the standard-bearer for the new AI era, with a market cap to match.
But Alphabet this week showed emphatically that it is ceding technology leadership to no one, and indeed is poised to reap some substantial rewards for its steady commitment to costly basic research.
The first announcement came Monday, when Google detailed a breakthrough in quantum computing: its new Willow chip overcomes a long-standing problem with error rates, bringing a technology that promises exponential increases in computing power much closer to reality.
On Tuesday, Alphabet’s Waymo unit saw longtime rival Cruise fold its tent, underscoring its wide lead in autonomous vehicles. Waymo continues to expand its real-live robotaxi service to new cities while GM’s Cruise backpedals and Tesla’s much-hyped effort remains on the drawing board.
On Wednesday came a major update of Google’s flagship Gemini AI model — twice as fast as the previous one—alongside a host of initiatives around agents and related AI services. The new offerings will boost the “AI Overviews” that Google now delivers alongside many search results, a feature that looked at first like a scraped-together catch-up play but has proven quite successful.
Alphabet now looks remarkably well-positioned for what promises to be a pivotal 2025. The stock is up more than 11% since Monday and nearly 45% for the year — a cool $600 billion increase in market cap.
The quantum chip, though still far from commercial applications, shows what pure R&D can bring at a time when more product-oriented approaches are in favor across the industry.
Dubbed Willow, the chip can quickly complete a computing problem that would take the world’s fastest super computer 10 septillion years, a data point that’s hard to comprehend but gives a sense of the potential. It’s the breakthrough on error rates, though, that really caught the industry’s attention.
Pete Shadbolt, the co-founder and chief scientist at startup PsiQuantum, described the advance as one he’s been waiting on for more than a decade.
“There were skeptics in the world that really doubted if this was practically possible — they have been soundly put to rest,” said Shadbolt.
The Gemini announcements were more iterative, though certainly more important for the near-term bottom line. Google touted Gemini 2.0 as an LLM for the “agentics era,” and early reviewers were wowed by features such as a “streaming API,” where developers can feed back into the model in real time.
Paying Gemini customers will get immediate access to an AI agent that acts as a research assistant for “exploring complex topics and compiling reports on your behalf.” We’ll definitely be playing with that one here at Newcomer.
Waymo’s win, meanwhile, has been so long in the making that it’s easy to miss its significance: it now has a fully operational robotaxi service for paying customers, with hundreds of vehicles across multiple cities and many more on the way.
Even a year ago, Waymo’s driverless cars maneuvered cautiously and awkwardly around San Francisco, but not anymore. The fleets are surely the most dramatic real-world expression of how AI will change society.
The Waymo vehicles deploy costly lidar and radar, and Elon Musk insists this approach is a mistake. He believes cameras alone can do the job with enough training data, and will eventually enable a far, far cheaper robotaxi. It’s a debate with big implications in the long run, but in the meantime Waymo is collecting fares and setting a high bar for performance and safety, while Tesla struggles to validate its approach.