Fusion energy has been “20 years away” for decades, but has the science finally caught up? Private investment in fusion companies surged from $10 billion to $15 billion in just months, and the money is coming from places you wouldn’t expect. 

Watch as Rebecca Bellan and guest host Tim De Chant sit down with Rachel Slaybaugh, general partner at DCVC on this episode of TechCrunch’s Equity podcast. The trio breaks down why serious investors are finally treating fusion as a real asset class, and what the return thesis actually looks like when no one expects a power plant in their fund lifetime. 

Subscribe to Equity on YouTube, Apple Podcasts, Overcast, Spotify and all the casts. You also can follow Equity on X and Threads, at @EquityPod. 

As part of its slate of Google Cloud Next announcements on Wednesday, the company shared plans to bring “auto browse” agentic capabilities to Chrome users in the enterprise, along with enhanced security measures.

With auto browse, Chrome users can take advantage of Gemini to understand the live context in their open browser tabs, and then use the AI to handle various tasks like booking travel, inputting data, scheduling meetings, and others related to web-based work.

Google suggests the tool could be used for things like inputting information in the company’s preferred CRM system based on content in a Google Doc, comparing vendor pricing across tabs, summarizing a candidate’s portfolio before an interview, pulling key data from a competitor’s product page, and more.

The company notes that its workflows will still require a “human in the loop,” meaning that the user will have to manually review and confirm the AI’s input before any final action takes place.

However, the idea is to help speed up these types of more tedious tasks to free up people to focus on what Google refers to as more “strategic work.”

This is the larger promise from AI advocates: that you’ll get your time back by using this new technology. But in practice, studies have shown that AI isn’t reducing work — it’s intensifying it. It remains to be seen how this will play out at the enterprise level as AI becomes a standard part of the workflow. Presumably, that could mean managers will expect that people can get more tasks done in less time.

Google says the new feature will initially be available to Workspace users in the U.S., as a part of Google’s push to infuse its AI into one of its most-used apps in the workplace, the web browser nearly everyone uses. It can be enabled via a policy, and Google states that an organization’s prompts won’t be used to train its AI models. (A disclosure that is increasingly necessary these days, given that Meta is even using its own employees’ keystrokes to train its AI.)

Like the consumer-facing version of the feature, Workspace users will be able to save their most common workflows for later use. These “Skills,” as they’re called, can be pulled up by either typing a forward slash (” / “) or by clicking the plus sign to access the needed Skill.

In addition to the infusion of AI into Chrome, Google is touting its ability to detect unsanctioned AI tools in the workplace via Chrome Enterprise Premium. Now, it’s expanding those capabilities to help IT teams look for compromised browser extensions or other AI services — specifically “anomalous agent activity.”

Google is correct to position this as a security feature, but it has another advantage, too. The tech giant is essentially leveraging corporate IT to shut down any other AI agents that could be taking root in the enterprise world organically. Years ago, this was how many web services established themselves in the workplace, amid an employee-driven “Enterprise 2.0” rush to adopt new technology like cloud storage, collaborative docs, or file sharing.

This new feature, which Google somewhat ominously dubs “Shadow IT risk detection,” will give IT teams visibility into the usage of both sanctioned and unsanctioned GenAI and SaaS sites across their organization.

IT teams will also receive a “Gemini Summary” of the Chrome Enterprise release notes and other AI-powered suggestions. This will surface critical changes, new policies, and upcoming deprecations, along with recommendations about things like configuring new settings or reviewing managed browsers.

The company also announced an expanded partnership with Okta to secure the agentic workplace with added features to reduce session hijacking and other protections. It’s also upgrading its security controls for extensions and introducing Microsoft Information Protection (MIP) integration to help organizations enforce consistent security policies.

Google Cloud on Wednesday announced that its eighth generation of custom-built AI chips, or tensor processing units (TPUs), will be split in two. One chip, named the TPU 8t, will be geared for model training and another, the TPU 8i, is aimed at inference.

Inference is the ongoing usage of models, aka what happens after users submit prompts.

As you might expect, the company touts some impressive performance specs for these new TPUs compared to the previous generations: up to 3x faster AI model training, 80% better performance per dollar, and the ability to get 1 million+ TPUs to work together in a single cluster. The upshot should be a lot more compute for a lot less energy — and cost to customers — than previous versions. It calls these chips TPUs, not GPUs, because its custom low-power chips were originally named Tensor.

But Google’s chips are not a full frontal assault on Nvidia’s future, at least not yet. Like the other giant cloud providers, including Microsoft and Amazon, Google is using these chips to supplement the Nvidia-based systems it offers in its infrastructure. It is not flat-out replacing Nvidia. In fact, Google promises its cloud will have Nvidia’s latest chip, Vera Rubin, available later this year.

One day the hyperscalers building their own AI chips (which includes Amazon, Microsoft, and Google) may grow to need Nvidia less, as enterprises move their AI needs to their clouds and port their apps to these chips.

Still, as things stand today, it’s not profitable to bet against Nvidia. As notable chip market analyst Patrick Moorhead jokingly posted on X, he had predicted that Google’s TPU could be bad news for Nvidia (and Intel) back in 2016 when the search giant launched its first one. Nvidia is now a nearly $5 trillion market cap company, meaning that prediction didn’t exactly hold up to the test of time.

If all goes according to Nvidia’s plan, Google’s growth as an AI cloud provider would result in more business for the chip maker not less, even if many a workload runs on Google’s chips.

In fact, Google also says it has agreed to work with Nvidia to engineer computer networking that allows Nvidia-based systems to perform even more efficiently in its cloud. In particular, the two tech giants are working to beef up the software-based networking tech called Falcon, which Google created and open sourced in 2023 under the godfather of all open source data center hardware organizations, the Open Compute Project.