While there’s a much bigger title coming to the service later in the month, Xbox has revealed the first five Game Pass additions for October. Among them are some newcomers to the new Game Pass Standard tier.

Baseball sim and the enjoyable narrative game were already on Game Pass Ultimate, PC Game Pass and Game Pass for Console (which is no longer available unless you were grandfathered in). They’ll join the Standard library on consoles on October 2 alongside . That captivating brawler is also coming to Game Pass Ultimate and PC Game Pass.

Two other games will hit all versions of the service save for Xbox Game Pass Core in the coming days (have we talked about how needlessly convoluted the Game Pass setup is lately?). Physics-based party brawler Mad Streets will join the lineup on October 7, followed by Inscryption on October 10. That creepy roguelike deck-builder is one of our picks for the .

Inevitably, Xbox will be removing some games from the library in the coming days to make way for the newcomers (and also because various licensing deals will be coming to an end). On October 15, it will yank Dyson Sphere Program, Everspace 2, From Space, F1 Manager 2023 and Scorn.

As ever, Xbox will add more titles to Game Pass in the back half of the month, including a lil’ under-the-radar one a few people might have heard of called . That one will any time soon, however.

Equinix has raised $15 billion in funding to expand its xScale data centers for AI, particularly for investments in the U.S.

Redwood City, California-based Equinix has built one of the backbones of the internet with data centers around the world. I visited a secret site once and was amazed at how big the places were that house tons of servers and cabling and cooling — and they’re about to get bigger and more plentiful.

Krupal Raval, managing director of xScale data centers at Equinix, said in an interview that the digital infrastructure company has completed the signing of a joint venture agreement, raising over $15 billion in capital with its partners. The exact mix of equity and debt is to be determined.

The limited liability partnership — subject to close in the fourth quarter — include GIC and the Canada Pension Plan Investment Board (CPP Investments).

“The $15 billion announcement is associated with, frankly, just the scale of the opportunity and real projects that we’re targeting,” Raval said. “That’s one of the points behind the $15 billion and then the second element is that the partnerships are very key to this equation.”

Raval noted that xScale data centers and the plan behind them were hatched five years ago, and there have already been $8 billion in financial commitments prior to today’s announcement. GIC has supported the expansion in the past, and now CPP is joining to further invest in North America.

“We feel like it’s a great testament to the health of our partnership and great working relationship. So we’re beyond thrilled over the fact that GIC is continuing to double down. I guess it’s tripling down into this project. But in addition to GIC, we also have CPP as a new investor. And the reason for that is because a the scale of the opportunity is so large, we thought it prudent to bring in multiple investment investor parties.”

Driven by increasing artificial intelligence (AI) and cloud growth, the joint venture is intended to accelerate the Equinix xScale data center portfolio, which enables hyperscale companies to add core deployments to their existing access point footprints at Equinix International Business Exchange (IBX) data centers. At full buildout, this new JV will nearly triple the investment capital of the Equinix xScale program.

Raval noted that the xScale program already represents an $8 billion commitment of capital, and this additional $15 billion will be invested in the U.S. to build out data centers to handle AI demand primarily in the U.S.

“It will change everything,” said Raval. “We are just in the early innings of AI. Everyone is talking about this as the single most important technological shift in generations.”

With the capital raised through the JV, Equinix expects the joint venture to purchase land to build new state-of-the-art xScale facilities on multiple greater-than-100-megawatt (MW) campuses in the U.S., eventually adding more than 1.5 gigawatts of new capacity for hyperscale customers.

Equinix has a longstanding relationship with GIC, having previously partnered on xScale projects in Asia, the Americas and Europe (see links below for details on other joint ventures). This agreement represents the first joint venture between Equinix and CPP Investments, which manages the assets of the Canada Pension Plan for more than 22 million contributors and beneficiaries.

Under the terms of the agreement, CPP Investments and GIC will each control a 37.5% equity interest in the joint venture, and Equinix will own a 25% equity interest. Each party has made equity commitments, and the joint venture also expects to take on debt to raise the total pool of investable capital to more than $15 billion over time.

Equinix’s existing hyperscale joint venture portfolio in Europe, Asia-Pacific and the Americas has a committed investment of over $8 billion, which is expected to result in greater than 725 megawatts of power capacity across more than 35 facilities at full buildout.

Platform Equinix features nearly 40% of the private on-ramps to the top global cloud service providers, which is more than any other provider. As hyperscale companies scale their operations at Equinix, the ecosystem of over 10,000 enterprises and other companies currently operating at Equinix can benefit from increased opportunities to directly connect and operate in proximity to the largest global cloud operators.

xScale data centers serve the unique core workload deployment needs of the world’s largest cloud service providers, including hyperscalers, which are key players in the AI ecosystem. These companies can add core deployments to their existing access point footprints at Equinix IBX data centers, enabling their growth on a single platform that can immediately span 72 global metros and offer direct interconnection to an ecosystem of more than 10,000 customers.

Equinix said is committed to delivering sustainable digital infrastructure and engaging our suppliers and partners in supply chain responsibility. Equinix has continued to make advancements in the way it designs, builds and operates its data centers with high energy-efficiency standards, and all xScale data centers will be LEED certified (or certified in the regional equivalent).

Raval said that the company maintains the highest standards in its sustainable approach to building its data centers.

“It’s an industry gold standard in terms of where we stand and in terms of our commitment to sustainability,” he said. “For many years, we’ve had a commitment towards being 100% based on clean energy. By 2030 we have science based targets, and we’re the trailblazer in many of these things.”

The closing of the joint venture is subject to the receipt of required regulatory approvals, which are expected to be received in the fourth quarter of 2024. Morgan Stanley served as exclusive financial advisor to Equinix in connection with this transaction.

“As the world’s leading companies build out their infrastructure to support key workloads such as artificial intelligence, they require the combination of large-scale data center footprints optimized for AI training and interconnection nodes for the most efficient inferencing,” said Adaire Fox-Martin, CEO of Equinix, in a statement. “Our xScale and IBX offerings are uniquely positioned to address this business need, enabling companies to realize the powerful potential of AI.”

Goh Chin Kiong, chief investment officer for real estate at GIC, said in a statement, “We are proud to expand our years-long partnership with Equinix, addressing a massive and growing demand for digital infrastructure, driven by the rapid advancement of technology, including AI. GIC’s capital and scale, paired with Equinix’s operational expertise, has driven meaningful value across our investments together. Through this joint venture, we look forward to providing the funding needed to develop state-of-the-art digital infrastructure across the U.S. alongside our likeminded partner, CPP Investments.”

Max Biagosch, senior managing director at CPP Investments, said in a statement, “CPP Investments has invested in data centers for several years and we have developed strong expertise in this space. This investment will help meet the increasing demand for data centers driven by rapid technological advancements and marks a significant step forward in our broader data center strategy. We are pleased to partner with Equinix and GIC to deliver strong long-term risk-adjusted returns for the CPP Fund.”

Raval noted Equinix invests multiple billions of dollars in capital expansion in normal years.

“I don’t think that we should necessarily limit ourselves to whatever history is. I think we have the capability to do more,” Raval said.

Full told, the xScale commitment is now about $23 billion. Raval thinks of a data center as a “product.” This amounts to about 20-plus gigawatts of power needed to run these data centers.

“We’ve designed a product that’s flexible, that can accommodate liquid cooling, that can run the gamut,” he said.

I asked about whether Nvidia’s belief in the onset of sovereign AI, where countries re-create their data infrastructure to make sure they own their own data, is a reason for this. And he said, in brief, yes.

“I think AI is going to grow everywhere,” he said. “This particular announcement is focused on the United States because we believe that the biggest growth market in AI is going to be the United States.”

Equinix expects an unspecified amount of hiring related to this project, which will include construction jobs. The company has already acquired its first U.S. xScale data center site, which will support 240 megawatts of power in the Atlanta area.

Sometimes, a demo is all you need to understand a product. And that’s the case with Runware. If you head over to Runware’s website, enter a prompt and hit enter to generate an image, you’ll be surprised by how quickly Runware generates the image for you — it takes less than a second.

Runware is a newcomer in the AI inference, or generative AI, startup landscape. The company is building its own servers and optimizing the software layer on those servers to remove bottlenecks and improve inference speeds for image generation models. The startup has already secured $3 million in funding from Andreessen Horowitz’s Speedrun, LakeStar’s Halo II and Lunar Ventures.

The company doesn’t want to reinvent the wheel. It just wants to make it spin faster. Behind the scenes, Runware manufactures its own servers with as many GPUs as possible on the same motherboard. It has its own custom-made cooling system and manages its own data centers.

When it comes to running AI models on its servers, Runware has optimized the orchestration layer with BIOS and operating system optimizations to improve cold start times. It has developed its own algorithms that allocate interference workloads.

The demo is impressive by itself. Now, the company wants to use all this work in research and development and turn it into a business.

Unlike many GPU hosting companies, Runware isn’t going to rent its GPUs based on GPU time. Instead, it believes companies should be encouraged to speed up workloads. That’s why Runware is offering an image generation API with a traditional cost-per-API-call fee structure. It’s based on popular AI models from Flux and Stable Diffusion.

“If you look at Together AI, Replicate, Hugging Face — all of them — they are selling compute based on GPU time,” co-founder and CEO Flaviu Radulescu told TechCrunch. “If you compare the amount of time it takes for us to make an image versus them. And then you compare the pricing, you will see that we are so much cheaper, so much faster.”

“It’s going to be impossible for them to match this performance,” he added. “Especially in a cloud provider, you have to run on a virtualized environment, which adds additional delays.”

As Runware is looking at the entire inference pipeline, and optimizing hardware and software, the company hopes that it will be able to use GPUs from multiple vendors in the near future. This has been an important endeavor for several startups as Nvidia is the clear leader in the GPU space, which means that Nvidia GPUs tend to be quite expensive.

“Right now, we use just Nvidia GPUs. But this should be an abstraction of the software layer,” Radulescu said. “We can switch a model from GPU memory in and out very, very fast, which allow us to put multiple customers on the same GPUs.

“So we are not like our competitors. They just load a model into the GPU and then the GPU does a very specific type of task. In our case, we’ve developed this software solution, which allow us to switch a model in the GPU memory as we do inference.“

If AMD and other GPU vendors can create compatibility layers that work with typical AI workloads, Runware is well positioned to build a hybrid cloud that would rely on GPUs from multiple vendors. And that will certainly help if it wants to remain cheaper than competitors at AI inference.