WP Pro Online

The growing demand for AI, particularly generative AI (i.e., AI that generates images, text and more), is supercharging the AI inferencing chip market. Inferencing chips accelerate the AI inferencing process, which is where AI systems generate outputs (e.g., text, images, audio) based on what they learned while “training” on a specific set of data. AI inferencing chips can be — and have been — used to yield faster generations from systems such as Stable Diffusion, which translates text prompts into artwork, and OpenAI’s GPT-3, which extends a few lines of prose into full-length poems, essays and more.

A number of vendors — both startups and well-established players — are actively developing and selling access to AI inferencing chips. There’s Hailo, Mythic and Flex Logix, to name a few upstarts. And on the incumbent side, Google’s competing for dominance with its tensor processing units (TPUs) while Amazon’s betting on Inferentia. But the competition, while fierce, hasn’t scared away firms like NeuReality, which occupy the AI chip inferencing market but aim to differentiate themselves by offering a suite of software and services to support their hardware.

On the subject, NeuReality today announced that it raised $35 million in a Series A funding round led by Samsung Ventures, Cardumen Capital, Varana Capital, OurCrowd and XT Hi-Tech with participation from SK Hynix, Cleveland Avenue, Korean Investment Partners, StoneBridge, and Glory Ventures. Co-founder and CEO Moshe Tanach tells TechCrunch that the tranche will be put toward finalizing the design of NeuReality’s flagship AI inferencing chip in early 2023 and shipping it to customers.

“NeuReality was founded with the vision to build a new generation of AI inferencing solutions that are unleashed from traditional CPU-centric architectures and deliver high performance and low latency, with the best possible efficiency in cost and power consumption,” Tanach told TechCrunch via email. “Most companies that can leverage AI don’t have the funds nor the huge R&D that Amazon, Meta and other huge companies investing in AI have. NeuReality will bring AI tech to anyone who wants to deploy easily and affordably.”

NeuReality was co-founded in 2019 by Tzvika Shmueli, Yossi Kasus and Tanach, who previously served as a director of engineering at Marvell and Intel. Shmueli was formerly the VP of back-end infrastructure at Mellanox Technologies and the VP of engineering at Habana Labs. As for Kasus, he held a senior director of engineering role at Mellanox and was the head of integrations at semiconductor company EZchip.

From the start, NeuReality focused on bringing to market AI hardware for cloud data centers and “edge” computers, or machines that run on-premises and do most of their data processing offline. Tanach says that the startup’s current-generation product lineup, the Network Attached Processing Unit (NAPU), is optimized for AI inference applications, including computer vision (think algorithms that recognize objects in photos), natural language processing (text-generating and classifying systems) and recommendation engines (like the type that suggest products on e-commerce sites).

NeuReality’s NAPU is essentially a hybrid of multiple types of processors. It can perform functions like AI inferencing load balancing, job scheduling and queue management, which have traditionally been done in software but not necessarily very efficiently.

NeuReality’s NR1, an FPGA-based SKU within the NAPU family, is a network-attached “server on a chip” with an embedded AI inferencing accelerator along with networking and virtualization capabilities. NeuReality also offers the NR1-M module, a PCIe card containing an NR1 and a network-attached inference server, and a separate module — the NR1-S — that pairs several NR1-Ms with the NR1.

On the software side, NeuReality delivers a set of tools, including a software development kit for cloud and local workloads, a deployment manager to help with runtime issues and a monitoring dashboard.

“The software for AI inference [and] the tools for heterogeneous compute and automated flow of compilation and deployment … is the magic that supports our innovative hardware approach,” Tanach said. “The first beneficiaries of the NAPU technology are enterprises and cloud solution providers that need infrastructure to support their chatbots, voice bots, automatic transcriptions and sentiment analysis as well as computer vision use cases for document scans, defect detection, etc. … While the world was focusing on the deep learning processor improvements, NeuReality focused on optimizing the system around it and the software layers above it to provide higher efficiency and a much easier flow to deploy inference.”

NeuReality, it must be noted, has yet to back up some of its performance claims with empirical evidence. It told ZDNet in a recent article that it estimates its hardware will deliver a 15x improvement in performance per dollar compared to the available GPUs and ASICs offered by deep learning accelerator vendors, but NeuReality hasn’t released validating benchmarking data. The startup also hasn’t detailed its proprietary networking protocol, a protocol that it has previously claimed is more performant than existing solutions.

Those items aside, delivering hardware at massive scale isn’t easy — particularly where it involves custom AI inferencing chips. But Tanach argues that NeuReality has laid the necessary groundwork, partnering with AMD-owned semiconductor manufacturer Xilinx for production and inking a partnership with IBM to work on hardware requirements for the NR1. (IBM, which is also a NeuReality design partner, previously said it’s “evaluating” the startup’s products for use in the IBM cloud.) NeuReality has been shipping prototypes to partners since May 2021, Tanach says.

According to Tanach, beyond IBM, NeuReality is working with Lenovo, AMD and unnamed cloud solution providers, system integrators, deep learning accelerator vendors and “inference-consuming” enterprises on deployments. Tanach declined, however, to reveal how many customers the startup currently has or what roughly it’s projecting in terms of revenue.

“We see that the pandemic is slowing companies down and pushing for consolidation between the many deep learning vendors. However, for us it doesn’t change anything, since late next year or sometime through 2024 inference deployment is expected to explode — and our technology is exactly the enabler and driver of that growth,” Tanach said. “The NAPU will bring AI for a broader set of less technical companies. It is also set to allow large-scale users such as ‘hyperscalers’ and next-wave data center customers to support their growing scale of AI usage.”

Ori Kirshner, the head of Samsung Ventures in Israel, added in an emailed statement: “We see substantial and immediate need for higher efficiency and easy-to-deploy inference solutions for data centers and on-premises use cases, and this is why we are investing in NeuReality. The company’s innovative disaggregation, data movement and processing technologies improve computation flows, compute-storage flows, and in-storage compute — all of which are critical for the ability to adopt and grow AI solutions.”

NeuReality, which currently has 40 employees, plans to hire 20 more over the next two fiscal quarters. To date, it’s raised $38 million in venture capital.

NeuReality lands $35M to bring AI accelerator chips to market by Kyle Wiggers originally published on TechCrunch

Rezonate, a Boston- and Tel Aviv-based startup that offers an agent-less cloud identity protection platform that aims to help DevOps teams minimize attackers’ opportunities to breach cloud identity and access, is coming out of stealth today and announcing an $8.7 million seed funding round, led by State of Mind Ventures and Flybridge, with participation from toDay Ventures, Merlin Ventures and a number of angel investors.

Founded in January 2022, Rezonate is part of a group of modern identity and access management (IAM) startups that aim to modernize the current state of affairs in this space, which is struggling to meet the demands of modern cloud infrastructure systems. This shift is creating new attack surfaces, especially as enterprises move to the cloud — and more dynamic infrastructure systems — at an ever-increasing rate. The number of security breaches stemming from issues with identity and access management is already on the rise. Indeed, Gartner expects that by 2023, “75% of security failures will result from inadequate management of identities, access, and privileges.”

Co-founder and CEO Roy Akerman was previously the head of the Israeli Cyberdefense Operations, while Rezonate co-founder and CTO Ori Amiga previously led R&D for this unit. Both received the Medal of Honor for their contributions to Israel’s National Security.

“The rapidly-changing cloudscape together with the proliferation of human and machine identities requires a different approach,” said Akerman. “Modern infrastructures require a precise and nimble way to outsmart attackers. One that prioritizes cloud identities and access at its core and is constantly adapting to current dynamics over yesterday’s snapshots and, for the first time, gives defenders and builders the means to act confidently.”

Rezonate promises to discover all of a company’s cloud and identity providers and the corresponding access privileges of its employees. The platform automatically detects security gaps and abnormal access attempts in real time. Rezonate promises that within minutes of deploying its solution, its platform can identify cloud identity and access risks and provide guidance for remediating them, or even automatically remove access and terminate sessions.

At the core of all of this is what Rezonate calls its ‘Identity Storyline,’ which aims to provide DevOps and security teams with a context-rich dashboard that helps them understand the security risk across a company’s cloud estate. With this, users get an easy-to-read dashboard that clearly lays out what kind of access every user has — and where there are potential issues.

“The fact that in just ten months from our first line of code we already have active customers, solving key gaps daily, affirms the criticality of the cloud identity and access issue. In a cloud world where everything is changing all of the time, DevOps teams need a solution as dynamic and automated as the infrastructure they need to protect is,” said Amiga.

Rezonate raises $8.7M and launches its cloud identity protection platform out of stealth by Frederic Lardinois originally published on TechCrunch

A year ago, The Wall Street Journal revealed that Facebook operated a two-tiered content moderation system. Normal users were subject to the platforms stated rules while VIP users were secretly flagged into special in a program internally called “cross check.”

That list included everyone from Brazilian soccer star Neymar and former President Donald Trump to conservative commentator Candace Owens and the company’s founder, Mark Zuckerberg. Per the WSJ, that system was designed to minimize instances in which Facebook might moderate content from a VIP in the normal course of moderation and kick off a firestorm of bad press in the process.

“If Facebook’s systems conclude that one of those accounts might have broken its rules, they don’t remove the content—at least not right away, the documents indicate,” the WSJ reported. “They route the complaint into a separate system, staffed by better-trained, full-time employees, for additional layers of review.”

Cross-check came to light in mid September of last year and by the end of the month the company was asking the Oversight Board, Meta’s semi-independent policy-making council, to review the system and suggest ways to fix it. “Specifically, we will ask the board for guidance on the criteria we use to determine what is prioritized for a secondary review via cross-check, as well as how we manage the program,” Meta VP of Global Affairs Nick Clegg wrote.

The Oversight Board is now back with their recommendations, calling for “significant improvements” to the cross-check program.

“For years, cross-check allowed content from a select group of politicians, business partners, celebrities, and others to remain on Facebook and Instagram for several days when it would have otherwise been removed quickly,” the group wrote in a blog post, noting that some content that fell under cross-check has remained up for 7 months before the company made a decision about whether to remove it.

The Oversight Board offered 32 recommended changes to that process, including a few steps that would make a previously secret program much more transparent. The board called on the company to publish “clear criteria” describing what accounts are eligible for cross-check’s extra review process, to visibly mark accounts that are in the program and to allow people who might meet the requirements to apply for the special account status.

The board also requested that Meta prioritize “users who are likely to produce expression that is important for human rights” like journalists and civil rights groups in the cross-check system rather than making those calls based on its business interests. “While the number of followers can indicate public interest in a user’s expression, a user’s celebrity or follower count should not be the sole criterion for receiving additional protection,” the board wrote. “If users included due to their commercial importance frequently post violating content, they should no longer benefit from special protection.”

The full set of recommendations, which is published on the Oversight Board’s blog, calls on Meta to dramatically realign its content moderation priorities for high profile users. How much of this the company will actually implement remains to be seen, but this whole process certainly looks like a well-oiled machine compared to the policy setting going on over at Twitter these days.

Meta’s Oversight Board wants Facebook to be more transparent about VIP accounts by Taylor Hatmaker originally published on TechCrunch

Apple has announced that its Self Service Repair store for iPhones and MacBooks is now open for business in Europe.

First announced last November, the repair program essentially enables anyone to purchase genuine Apple components to repair their damaged devices, while the Cupertino company also provides online manuals to guide consumers through the self-service repair process.

It’s worth noting that while the program is open to anyone where the repair store is available, repairing iPhones and MacBooks probably isn’t for the average consumer, as just getting into the devices to being the repairs is a complex process. But for any have-a-go-heroes out there willing to invest a bit of time and money learning, Apple is also selling the tools necessary to carry out fault-specific repairs, with an option to rent a repair kit for $49 if they only have a one-off repair they wish to carry out.

Apple first opened its Service Service Repair store to iPhone users in the U.S. back in April, with support for the full range of iPhone 12 and 13 models, as well as the third edition SE model. The company then extended the program to include a selection of MacBook Air and Pro models in August.

From today, Apple is opening Self Service Repair to eight European countries, including the U.K., France, Germany, Belgium, Italy, Poland, Spain, and Sweden.

Apple expands Self Service Repair to iPhone and MacBook users in Europe by Paul Sawers originally published on TechCrunch

Hackers backed by the Iranian government targeted human rights activists, journalists, diplomats and politicians working in the Middle East during an ongoing social engineering and credential phishing campaign, according to Human Rights Watch.

In an analysis published on Monday, Human Rights Watch said it had attributed the espionage campaign to APT42, an Iran-backed hacking group first identified by cybersecurity firm Mandiant in September. Mandiant said APT42 – also referred to as TA453, Phosphorus and Charming Kitten – supports Iran’s Islamic Revolutionary Guard Corps intelligence collection efforts and has launched over 30 confirmed operations against various non-profit, education and government targets globally since 2015.

Human Rights Watch said it first became aware of APT42’s latest espionage campaign after one of its employees received suspicious messages on WhatsApp from someone pretending to work for a think tank based in Lebanon. The advocacy group found that a link included in the message directed the target to a fake login page that captured their email password and multi-factor authentication code.

In its analysis, conducted alongside Amnesty International’s Security Lab, Human Rights Watch identified 18 additional victims who had been targeted as part of the same campaign, and 15 of these targets confirmed that they had received the same WhatsApp messages between September 15 and November 25. On November 23, a second Human Rights Watch staff member received the same WhatsApp messages from the same number that contacted other targets.

For the three people whose accounts were known to be compromised — a correspondent for a major U.S. newspaper, a women’s rights defender based in the Gulf region, and an advocacy consultant for Refugees International based in Lebanon — the attackers gained access to emails, cloud storage drives, contacts and calendars. In at least one case, the attackers also performed a Google Takeout, a service that exports all of an account’s activity and information, including web searches, payments, travel and locations, ads clicked on, YouTube activity, and additional account information.

“Iran’s state-backed hackers are aggressively using sophisticated social engineering and credential harvesting tactics to access sensitive information and contacts held by Middle East-focused researchers and civil society groups,” said Abir Ghattas, information security director at Human Rights Watch. “This significantly increases the risks that journalists and human rights defenders face in Iran and elsewhere in the region.”

In light of its investigation, Human Rights Watch is calling on Google to strengthen its Gmail account security warnings to protect better its most at-risk users, including journalists and human rights defenders, after it uncovered “inadequacies” in Google’s security protections.

“Individuals successfully targeted by the phishing attack told Human Rights Watch that they did not realize their Gmail accounts had been compromised or a Google Takeout had been initiated, in part because the security warnings under Google’s account activity do not push or display any permanent notification in a user’s inbox or send a push message to the Gmail app on their phone,” Human Rights Watch said in its analysis.

“Google’s security activity revealed that the attackers accessed the targets’ accounts almost immediately after the compromise, and they maintained access to the accounts until the Human Rights Watch and Amnesty International research team informed them and assisted them in removing the attacker’s connected device.”

Google spokesperson Kimberly Samra told TechCrunch that Google implements protections for high-risk users so their Google accounts are “protected against threats against Google services, or on other platforms as seen in this case.”

“Some of these protections include our Advanced Protection Program (APP) and 2-Step Verification (2SV) auto enrollments,” Samra said. “Google also remains committed to threat collaboration and sharing our ongoing research to raise awareness on bad actors across the industry, as it helps to more quickly respond to attacks and protect online users.”

Iran-backed hackers linked to espionage campaign targeting journalists and activists by Carly Page originally published on TechCrunch

The Discovery+ app is getting one of the most anticipated feature — and frankly, a must-have feature for today’s streaming apps — offline viewing. The company announced Monday that users of the ad-free plan in the U.S. will be able to download content for offline viewing on the service’s iOS and Android apps.

This move comes after the company tested this feature with users in Brazil. Discovery+ mentioned that the service has more than 58,000 episodes — including popular shows “House of Hammer,” “Fixer Upper,” and “90 Day Fiancé” — that are eligible for downloads.

Users with the ad-free plan will see a “Download” button next to the content title. They can also choose the quality of the download and whether to download it over Wi-FI or cellular data.

Discovery+ noted that the downloaded title will live on the device for 30 days in the unplayed state or 48 hours after you hit the play button. Plus, you can renew the title download once you’re online after it expires.

The company is not putting any geo-restriction on the downloaded content. So if you are visiting another country, you can watch it if you are offline. The app will default to the downloads section when a user is offline.

Warner Bros. Discovery is set to merge HBO Max and Discovery+ services into one app next year. But until then, Discovery+ users will get to enjoy this feature that was missing for the longest time. Competitors like YouTube and Netflix have had an offline viewing option for years. While Discovery+ launched just last year, it’s a feature that streaming service users have come to expect.

Discovery+ brings the offline download feature to the US by Ivan Mehta originally published on TechCrunch

There’s a growing market for private cellular networks, or dedicated cell networks configured to support a company’s specific requirements within a confined area (think a warehouse or wind farm). For example, some mining and energy companies have adopted private cellular networks to help facilitate operations that span over significant distances, where the increased range of cellular offers advantages compared to the Wi-Fi alternatives.

A number of major vendors provide private cellular network services, including AT&T and T-Mobile, as do some startups, including Celona, Anterix and Airspan Networks. (Recently, asset management giant BlackRock announced that it’ll deploy a private network at its new headquarters in partnership with Verizon.) But that hasn’t stopped new ventures from cropping up to challenge the incumbents. See Expeto, which is developing a platform that allows corporate customers to extend their private networks via virtually any type of cellular connectivity.

Demonstrating that there’s investor appetite for upstarts in the market, Expeto this week closed a $13 million Series B round led by Sorenson Capital with participation from 5G Open Innovation Lab, Samsung Next and Mistral Venture Partners. Bringing Expeto’s total raised to over $25 million (inclusive of $5 million in debt financing), the proceeds will be put toward product R&D as well as customer and partner acquisition, according to CEO Michael Anderson.

“Large, mission-critical enterprises in sectors like energy, mining and manufacturing need mobile connectivity to support next-gen use cases. But today, mobile operator connectivity is hard to buy, integrate and manage,” Anderson told TechCrunch in an email interview. “If they want connectivity, enterprises have to deal with this mobile network operator (MNO) over here and that MNO over there … Expeto lets enterprises manage mobile connectivity as if it were part of their own network.”

Co-founded in 2015 by Terje Strand, Ryley MacKenzie and Brian Baird, with Anderson coming onboard in 2020, Expeto doesn’t sell cell service. Instead, it offers tools to enable companies to manage multiple private cellular networks, including 3G, 4G and 5G networks, from a single pane of glass. Expeto supports managing network policies in addition to monitoring network usage and performance, and it is designed to work from behind a corporate firewall, in a private or public cloud, at the edge or from a geographically remote site depending on the use case requirements.

“With Expeto, enterprises see just one network — no matter how many mobile network operators or deployment sites they’re managing,” Anderson said. “[They can deliver] mobility for any type of device using just a single SIM [card], maintain network device privileges across different connectivity types and physical sites [and] make real-time, self-service changes to the mobile networks they manage.”

Expeto currently has “more than” 20 customers supporting over 50 private wireless sites, according to Anderson, who wouldn’t reveal revenue figures when asked. He claims that the pandemic has driven “significant demand” for Expeto’s products as companies embrace automation and remote operations, even as some organizations look to reduce costs in the face of economic uncertainty.

To his point, there appears to be high interest among the enterprise in deploying private cell networks. A recent by TECHnalysis Research survey of over 600 IT decision-makers found that nearly three-quarters believe their company will eventually use a private 5G network. A separate reportfrom ABI Research estimates that the market for 5G private networks will reach $47.5 billion in 2030, up from $3.7 billion in 2021.

“Because Expeto’s use cases are associated with meaningful and relevant advances in operational efficiencies and growth initiatives, we have not experienced significant headwinds and we continue to sell on the basis of the savings and efficiencies Expeto’s platform uniquely enables,” Anderson said. “We plan to raise our next round of capital in the second half of 2024. We determined that the best approach to continue our high-growth momentum is a financial structure comprising equity and debt with focus and accountability to deliver differentiated growth, innovation and references from recognized leaders in multiple vertical markets.”

Samsung Next’s Andy Duong argues that Expeto is “uniquely positioned” to scale by leveraging its customer base as a distribution channel. The long-term play could be collaborating with cell providers to sell additional services; Expeto already works with Nokia and Ericsson for their radio access networks, the parts of mobile networks that connect devices like smartphones to the cloud.

“We invested in Expeto because it simplifies enterprise networking over private and public mobile networks,” he added. “While 5G’s network slices can be customized to meet the service-level agreement for emerging connected applications, they also bring additional complexities in network management. This is made easier thanks to Expeto’s cloud-based platform that enables hybrid private mobile networks for enterprise customers.”

Expeto, a startup selling tools to manage private cellular networks, raises $13M by Kyle Wiggers originally published on TechCrunch

Osome is a corporate “super app” that helps business owners with administrative tasks like payroll, accounting and tax reporting. The company announced today it has raised a $25 million Series B from Illuminate Financial, AFG Partners and Winter Capital. This brings Osome’s total raised since it was founded in 2017 to $51 million.

The company says revenues have doubled since its Series A announced in June 2021. It plans to become cash flow positive within the next 12 months, and recently announced a digital banking partnership with Singapore financial service corporation OCBC.

Osome currently serves more than 11,000 business in Singapore (where it is is headquartered), Hong Kong and the United Kingdom. It also offers business incorporation services in Singapore, Hong Kong and the United Kingdom, and integrates with e-commerce platforms like Amazon, eBay, Shopify, Lazada, Etsy and Shopee.

Part of Osome’s new funding will be used to expand its operations in Asia by targeting side hustles and micro-entrepreneurs, in addition to its current customer base of SMEs.

Over the last year, Osome has launched an accounting platform to provide tax and financial reports, expenses and invoice management. It also runs a hybrid accounting service, called the Accounting Factory, that combines machine learning with human accountants and is meant to replace accounting software like Xero and Quickbooks. Machine learning is used to collect, extract and categorize financial data and reconcile it with bank transactions. Then Osome’s accountants look at that information and advise customers. Osome currently has more than 100 accountants and bookkeepers, who are full-time employees.

Other startups that offer corporate services include Sleek, Lanturn and BlueMeg. Osome founder Victor Lysenko said its building a competitive moats by providing a “set-up to scale-up service for businesses.”

“What business owners tell us is that they didn’t startup to do their own bookkeeping,” he said. “We take care of the bookkeeping so they can focus on their business. And we grow with them—our pricing model is based around revenue, not transactions, unlike our competitors.”

Singapore-based corporate services super app Osome raises $25M Series B by Catherine Shu originally published on TechCrunch