Vahdat, a vice president at Google Cloud, presented slides showing the company needs to scale "the next 1000x in 4-5 years." While a thousandfold increase in compute capacity sounds ambitious by itself, Vahdat noted some key constraints: Google needs to be able to deliver this increase in capability, compute, and storage networking "for essentially the same cost and increasingly, the same power, the same energy level," he told employees during the meeting.
Amazon recently issued $15 billion in debt, including a rare 40-year bond, and saw demand approach $80 billion - more than five times oversubscribed . Investors accepted yields only about 80 basis points above comparable U.S. Treasuries, effectively treating Amazon as quasi-sovereign credit despite the long maturity. This level of demand is extraordinary for a private company raising funds for aggressive capital expenditures.
But it's a laboratory, hidden away at the back of the campus behind some trees, that he is most excited for me to see. This is where the invention that Google believes is its secret weapon is being developed. Known as a Tensor Processing Unit (or TPU), it looks like an unassuming little chip but, says Mr Pichai, it will one day power every AI query that goes through Google. This makes it potentially one of the most important objects in the world economy
Even as the investing world continues to debate whether we are or are not in an "AI bubble," there shouldn't be any question whether AI-driven dividend stocks are still enjoying a moment. As artificial intelligence dominates much of the daily conversation in the tech world, these stocks and their shareholders are all enjoying outstanding returns. The good news is that not every AI-driven dividend stock is under the microscope as part of the bubble.
The forecasts are eye-popping: utilities saying they'll need two or three times more electricity within a few years to power massive new data centers that are feeding a fast-growing AI economy. But the challenges - some say the impossibility - of building new power plants to meet that demand so quickly has set off alarm bells for lawmakers, policymakers and regulators who wonder if those utility forecasts can be trusted.
More recently, Lumen partnered with IBM to unlock scalable AI for businesses, as well as with Google Cloud to provide advanced cloud and network solutions to meet the growing demands of AI workloads. Lumen strengthened its financial position and freed up capital for long-term growth by refinancing its term loans, as well as selling off its fiber-to-the-home business to AT&T.
DDR5 isn't normally associated with AI accelerators. Those typically employ GDDR or HBM, which trade capacity for the higher bandwidth required for those workloads. However, the systems into which these accelerators are deployed do use DRR5 for the system memory to feed the CPU and to store intermediary data, like the key value caches used to prolong a model's short-term memory over extended periods.
"While others plan, we're building the infrastructure that makes AI possible," said Bill Long, chief product and strategy officer at Zayo. "Without connectivity, datacentres and AI factories are just expensive refrigerators: cold boxes of compute with no way for data to get in or out. We're delivering the capacity and reach where it's needed to ensure AI can work, scale and innovate without limits."
Bitfarms, one of North America's largest Bitcoin miners, announced it will gradually wind down its mining operations over the next two years. The company plans to shift its focus to high-performance computing (HPC) and artificial intelligence (AI) infrastructure. The move reflects a broader trend among crypto miners. Falling Bitcoin prices and shrinking profit margins are pushing operators to explore more stable revenue streams.
The European colocation market is plugging the demand gaps for datacentre capacity caused by the waning appetites of hyperscale cloud firms by cozying up to neocloud providers, research from CBRE shows. The real estate consultancy's third quarter look at the state of the European datacentre market reveals that, despite a drop-off in hyperscaler demand for colocation capacity, the amount being acquired for hosting artificial intelligence (AI) workloads has soared.
More than sixty years after the HDD was invented, this fundamentally outdated technology still proves to be sufficiently suitable for modern applications. Let there be no doubt: the use of hard drives for AI is anything but obvious. In a domain where the lowest latency is required, HDD speeds are often ten times slower than those of SSDs. But to run AI models, HBM memory is needed that is placed directly next to the chip.
The company exceeded expectations last quarter with revenue of $14.67 billion and EPS of $0.99, as AI infrastructure orders more than doubled its prior-year target. CEO Chuck Robbins said the company has "established a solid foundation" for its strongest year yet, driven by record webscale orders, early enterprise AI deployments, and refreshed product lines built around its Silicon One platform.
What's lifting growth now isn't consumer demand: it's infrastructure. From data centers and chip foundries to electric utilities and fiber cabling, billions are being poured into the AI ecosystem. Companies like Nvidia (NASDAQ: NVDA) are selling record volumes of GPUs, while utilities expand capacity to power the next wave of computing. That spending circulates across the economy: construction, logistics, real estate, and materials, creating a multiplier effect reminiscent of the early internet years.
Alphabet's revenue grew 16% to $102.4 billion, ahead of analyst estimates of $99.9 billion. EPS of $3.10 was also ahead of analyst estimates of $2.33. Google's search business revenues grew 15% to $56.6 billion. YouTube advertising revenues grew from $8.9 billion a year ago to $10.3 billion, ahead of the market's estimated $10.1 billion. Alphabet reported overall advertising revenues of $74.2 billion, compared with $65.9 billion a year ago. Cloud revenues of $15.2 billion grew 35%, ahead of the market's estimated $14.7 billion.
Verizon AI Connect comprises a network infrastructure and suite of products designed to enable global enterprises to deploy AI workloads at scale. It is said to bring together and "reimagine" existing assets such as power, space and cooling across telco facilities; Verizon's One Fiber infrastructure; long-haul, metro and local fibre and optical networks; and the Verizon 5G Network, offering functionality such as network slicing.
