Welcome to "#SocialStocks," The Fly's weekly recap of Wall Street's reactions to social media stock news.

CARPE DIEM: The Diem Association, the cryptocurrency project Facebook (FB) founder Mark Zuckerburg once defended before Congress, is unraveling following regulatory pressure, Bloomberg's Liana Baker, Jesse Hamilton, and Olga Kharif reported, citing people familiar with the matter. The project, the initiative once known as Libra, backed by Meta Platforms (MVRS), is mulling a sale of its assets as a way to return capital to its investor members, according to the report.

NFT: Alphabet's (GOOG, GOOGL) YouTube is exploring adding nonfungible token features for its video creators, Bloomberg's Mark Bergen reported, citing a letter to the site's broadcasters from YouTube CEO Susan Wojcicki. Twitter (TWTR) began letting users post NFTs as profile photos and Instagram is working on a similar offering, according to reports. "We're always focused on expanding the YouTube ecosystem to help creators capitalize on emerging technologies, including things like NFTs, while continuing to strengthen and enhance the experiences creators and fans have on YouTube," Wojcicki wrote. A YouTube spokeswoman declined to share more about the site's NFT plans. Separately, Meta is working on plans to allow users to create and sell NFTs, Hannah Murphy and Cristina Criddle of The Financial Times reported. Teams at Facebook and Instagram are currently readying a feature that would allow users to display their NFTs on their social media profiles as well as working on plans to help users create the tokens, several people familiar with the matter told the Times. Two of The Financial Times' sources said Meta is discussing the creation of an NFT marketplace that would allow users to buy and sell NFTs. These plans are in the early stages and may be changed.

BIRDS OF A FEATHER: Twitter is working on a feature called Flock which will let users curate a list of people to send selected tweets out to, reminiscent of Instagram's Close Friends feature, Input's Matt Wille reported. Much like the Instagram counterpart, Flock will let people add followers to a list and later send out tweets exclusively to those followers, the author says, noting that Twitter is currently experimenting with a list cap of 150 people. The feature is reportedly still in the development pipeline. The feature was previously known as “Trusted Friends,” but the name has since shifted to “Flock,” according to code found by Alessandro Paluzzi. On the other hand, Twitter's interest in generating more personal timeline experiences is not limited to the new Flock feature. The Communities feature, which is still in beta testing, allows users to create a more close-knit Twitter experience. Both features address those who want to tweet something to a specific subset of followers.

ANTITRUST LATEST: Google has requested that a federal judge throw out the majority of an antitrust suit filed by the state of Texas and other states that alleged the search giant abused its dominance in the online advertising market, Reuters' Diane Bartz reported. The company said in its court filing that the states failed to show that it illegally worked with Facebook to counter a technology the publishers use to make more money from advertising placed on their sites, the author said, noting that Facebook has not been named as a defendant in the litigation. Allegedly,  Google used at least three programs to manipulate ad auctions to basically force advertisers and publishers to use Google's tools. In response, Google claims the states had a "collection of grievances" but no proof of wrongdoing. Google also argued the states waited too long to file its lawsuit on a number of charges. "They criticize Google for not designing its products to better suit its rivals' needs and for making improvements to those products that leave its competitors too far behind. They see the 'solution' to Google’s success as holding Google back," the company said in its filing.

SECURITY: Twitter terminated its head of security Peiter Zatko this week, and chief information security officer Rinki Sethi will depart in the coming weeks, The New York Times' Mike Isaac and Kate Conger reported. Both executives joined Twitter in 2020. The changes follow "an assessment of how the organization was being led and the impact on top priority work," according to a memo from new CEO Parag Agrawal. Agrawal was appointed Twitter’s chief executive in November and has reorganized the company’s executives since taking over from Jack Dorsey. In December, Agrawal shuffled the leadership team and dismissed Dantley Davis, the chief design officer, and Michael Montano, the head of engineering.

HEADSET COMPETITION: Google has recently begun ramping up work on an AR headset, internally codenamed "Project Iris," and hopes to ship the device in 2024, according to The Verge's Alex Heath, citing two people familiar with the project. Google's device, which is still early in development "without a clearly defined go-to-market strategy," would use outward-facing cameras to blend computer graphics with a video feed of the real world, similar to expected-to-be-upcoming headsets from Meta (FB) and Apple (AAPL), the report said. Project Iris marks a return to a hardware category that Google has dabbled in in the past. It started with Google Glass in 2012 and then a multi-year effort to sell VR headsets petered out in 2019. Google has since been silent about its hardware aspirations.

SUPERCOMPUTER: Meta said in a blog post that, "Today, Meta is announcing that we've designed and built the AI Research SuperCluster (RSC) - which we believe is among the fastest AI supercomputers running today and will be the fastest AI supercomputer in the world when it's fully built out in mid-2022. Our researchers have already started using RSC to train large models in natural language processing (NLP) and computer vision for research, with the aim of one day training models with trillions of parameters. RSC will help Meta's AI researchers build new and better AI models that can learn from trillions of examples; work across hundreds of different languages; seamlessly analyze text, images, and video together; develop new augmented reality tools; and much more. Our researchers will be able to train the largest models needed to develop advanced AI for computer vision, NLP, speech recognition, and more. We hope RSC will help us build entirely new AI systems that can, for example, power real-time voice translations to large groups of people, each speaking a different language, so they can seamlessly collaborate on a research project or play an AR game together. Ultimately, the work done with RSC will pave the way toward building technologies for the next major computing platform - the metaverse, where AI-driven applications and products will play an important role... AI supercomputers are built by combining multiple GPUs into compute nodes, which are then connected by a high-performance network fabric to allow fast communication between those GPUs. RSC today comprises a total of 760 NVIDIA DGX A100 systems as its compute nodes, for a total of 6,080 GPUs - with each A100 GPU being more powerful than the V100 used in our previous system. The GPUs communicate via an NVIDIA Quantum 200 Gb/s InfiniBand two-level Clos fabric that has no oversubscription. RSC's storage tier has 175 petabytes of Pure Storage FlashArray, 46 petabytes of cache storage in Penguin Computing Altus systems, and 10 petabytes of Pure Storage FlashBlade... RSC is up and running today, but its development is ongoing. Once we complete phase two of building out RSC, we believe it will be the fastest AI supercomputer in the world, performing at nearly 5 exaflops of mixed precision compute. Through 2022, we'll work to increase the number of GPUs from 6,080 to 16,000, which will increase AI training performance by more than 2.5x." Penguin Computing, Inc. a division of SGH (SGH), announced its role in providing AI-optimized architecture and managed services for the AI Research SuperCluster, or RSC, which it identifies as Meta's "cutting-edge AI supercomputer for AI research." Meta announced via a blog post the introduction of the RSC, which Meta believes is "among the fastest AI supercomputers running today and will be the fastest AI supercomputer in the world when it's fully built out in mid-2022," Penguin stated in a press release. "Penguin Computing, a long-time partner with Meta, provided the managed services and AI-optimized infrastructure for the RSC, which is comprised of 760 NVIDIA DGX A100 systems with more than 6,000 NVIDIA GPUs and 46 petabytes of cache storage-enabling what Meta believes is one of the fastest AI supercomputers currently in operation," the company added. Additionally, Pure Storage (PSTG) reported a role in Meta's Research SuperCluster. The company said, "RSC is helping Meta's AI researchers build new and better AI models that can learn from trillions of examples, work across hundreds of different languages, seamlessly analyze text, images, and video together, develop new augmented reality tools, and much more. TSC will pave the way toward building technologies for Meta's next major computing platform, the metaverse, where AI-driven applications and products will play an important role. Meta chose Pure as it needed a storage partner that can deliver robust and scalable storage capabilities to power RSC."

ANALYST COMMETARY: Truist analyst Youssef Squali lowered the firm's price target on Snap (SNAP) to $65 from $75 but keeps a Buy rating on the shares as part of a broader research note on Internet and Digital Media. The company should deliver 32% y/y growth on the top-line and 14% EBITDA margins thanks to further monetization gains and DAU growth, but revenue growth and margins should both be a step down from Q3's +57% and 16%, respectively, reflecting advertiser disruptions from iOS 14.5 and tough Y/Y comps, the analyst tells investors in a research note. Wedbush analyst Ygal Arounian downgraded Snap to Neutral from Outperform with a $36 price target as he sees risk to the company's revenue growth targets stemming from IDFA headwinds, difficult comps from stellar growth in 2020-2021, and increasing competition from TikTok in particular. The analyst has seen little evidence of progress against IDFA since Snap reported Q3 earnings, with his checks indicating continued headwinds across digital advertising. While that is the biggest near- and potentially mid-term impact to Snap's ability to get back to 50%-plus revenue growth, competitive factors are increasing as well, which leave him more cautious on a risk/reward basis, particularly in an environment where high growth risk multiples are under duress, Arounian argues.

Also, OTR Global downgraded its view on Pinterest (PINS) to Mixed from Positive after checks on Q4 with sources at full-service and digital advertising agencies in the U.S. and Europe.

RBC Capital analyst Brad Erickson lowered the firm's price target on Meta Platforms to $400 from $415 but keeps an Outperform rating on the shares. The analyst says the data with 26 agency and ad network channel contacts related to overall spending trends are "mixed". Erickson adds that while Meta is "winning the Apple (AAPL) signal loss battle" relative to other social peers, it may also be more exposed to Tiktok risk than investors may realize.

Squali lowered the firm's price target on Twitter to $50 from $60 but keeps a Buy rating on the shares as part of a broader research note on Internet and Digital Media. The company's Q4 revenue growth should decelerate sequentially to 22% from 37% in Q3, reflecting tough Y/Y comps, and normalization from peak engagement on the platform due to COVID, though he still expects over 20% Y/Y top-line growth, fueled by healthy ad budgets, new monetization tools, and mDAU growth, the analyst tells investors in a research note. Mizuho analyst James Lee lowered the firm's price target on Twitter to $56 from $70 and keeps a Neutral rating on the shares. While the management changes are necessary to enable Twitter to accelerate the development, introduction and update of its products, it could take time for the new strategy to take shape for users and revenues to reach the company's long term targets, Lee tells investors in a research note. With that in mind, he reduced his fiscal 2023 EBITDA estimate by nearly 20% to $2.4B. Lee believes the stock lacks a catalyst in the near term.