Alphabet (NASDAQ:GOOGL) stock is down 13.5% over one month. That’s not as bad as some of its peers, but the stock is down 25% from its highs. In other words, a £10,000 investment made one month ago would be worth £8,650 today. The pound is roughly flat against the dollar so there’s no need to account for currency fluctuations.
What’s behind the sell-off?
Alphabet, Google’s parent company, has faced a challenging year, with its stock plunging by 23.4%. The broader market sell-off has played a role. However, the decline has been exacerbated by Google’s struggles in keeping pace with competitors in AI search.
Since the rise of large language models (LLMs) in 2023, Google’s Gemini has lagged significantly behind OpenAI’s ChatGPT, missing out on the crucial first-mover advantage.
Gemini, formerly known as Bard, has encountered multiple setbacks since its launch. Its debut in March 2023 was marred by a factual error during a live demo. This caused Alphabet’s stock to nosedive before recovering.
Subsequent issues, such as challenges in image generation, further hindered Gemini’s progress. These missteps led Google to adopt a cautious approach to AI development—a strategy that has come at the expense of market share. Today, Gemini holds only 13.5% of the market.
While Gemini has some unique features — like being particularly helpful for novice coders — its overall capabilities have been underwhelming compared to competitors. Features like photo and document uploads were initially absent, and its coding skills and reasoning abilities lagged behind rivals like ChatGPT and Claude. Even though Google’s NotebookLM offered some functionalities, poor user awareness limited its impact.
AI developments
Since early 2025, CEO Sundar Pichai has pushed for faster rollouts of Gemini updates. The latest version, Gemini 2.5, introduces reasoning capabilities, marking a step forward in performance and accuracy.
According to Google, these “thinking models” can reason through their responses before delivering them. Gemini 2.5 is now better equipped for multimodal reasoning, advanced coding tasks, and complex math and science questions.
Benchmark data shows some improvement. Gemini 2.5 Pro Experimental scores higher than OpenAI’s o3-mini in areas like science (84% vs. 79.7%) and math (86.7% vs. 86.5%). However, it still trails GPT-4.5 in critical benchmarks like Factuality SimpleQA (FSQA), which measures accuracy in basic fact-based queries. This is a key metric for everyday interactions.
Looking beyond AI
While Gemini 2.5 may be a step forward, the company has some broader risks. Firstly, the US economy looks set fall into a recession this year, and that not going to be positive for Google’s advertising revenue. It currently controls more than 90% of the search market share and has a dominant position is digital advertising. Likewise, Google Cloud growth was slower than expected in recent quarters.
However, I’m tempted to look beyond these near-term challenges. And that’s why I’ve recently added a little to my portfolio. The company has catalysts in the form of Waymo — it’s robotaxi venture — and quantum computing where, I believe, big tech companies are likely to make the biggest commercial breakthroughs. It’s also trading at a 36% discount to its five-year average forward price-to-earnings. This may signal a near-term downturn in earnings. But I’m bullish in the long run.
